idnits 2.17.1 draft-ietf-detnet-use-cases-18.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- No issues found here. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year -- The document date (September 17, 2018) is 2041 days in the past. Is this intentional? Checking references for intended status: Informational ---------------------------------------------------------------------------- == Outdated reference: A later version (-30) exists of draft-ietf-6tisch-architecture-14 == Outdated reference: A later version (-13) exists of draft-ietf-detnet-architecture-07 == Outdated reference: A later version (-09) exists of draft-ietf-detnet-problem-statement-06 Summary: 0 errors (**), 0 flaws (~~), 4 warnings (==), 1 comment (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Internet Engineering Task Force E. Grossman, Ed. 3 Internet-Draft DOLBY 4 Intended status: Informational September 17, 2018 5 Expires: March 21, 2019 7 Deterministic Networking Use Cases 8 draft-ietf-detnet-use-cases-18 10 Abstract 12 This draft documents use cases in several diverse industries to 13 establish multi-hop paths for characterized flows with deterministic 14 properties. In this context deterministic implies that flows can be 15 established which provide guaranteed bandwidth and latency which can 16 be established from either a Layer 2 or Layer 3 (IP) interface, and 17 which can co-exist on an IP network with best-effort traffic. 19 Additional use case properties include optional redundant paths, very 20 high reliability paths, time synchronization, and clock distribution. 21 Industries considered include professional audio, electrical 22 utilities, building automation systems, wireless for industrial, 23 cellular radio, industrial machine-to-machine, mining, private 24 blockchain, and network slicing. 26 For each case, this document will identify the application, identify 27 representative solutions used today, and the improvements IETF DetNet 28 solutions may enable. 30 Status of This Memo 32 This Internet-Draft is submitted in full conformance with the 33 provisions of BCP 78 and BCP 79. 35 Internet-Drafts are working documents of the Internet Engineering 36 Task Force (IETF). Note that other groups may also distribute 37 working documents as Internet-Drafts. The list of current Internet- 38 Drafts is at https://datatracker.ietf.org/drafts/current/. 40 Internet-Drafts are draft documents valid for a maximum of six months 41 and may be updated, replaced, or obsoleted by other documents at any 42 time. It is inappropriate to use Internet-Drafts as reference 43 material or to cite them other than as "work in progress." 45 This Internet-Draft will expire on March 21, 2019. 47 Copyright Notice 49 Copyright (c) 2018 IETF Trust and the persons identified as the 50 document authors. All rights reserved. 52 This document is subject to BCP 78 and the IETF Trust's Legal 53 Provisions Relating to IETF Documents 54 (https://trustee.ietf.org/license-info) in effect on the date of 55 publication of this document. Please review these documents 56 carefully, as they describe your rights and restrictions with respect 57 to this document. Code Components extracted from this document must 58 include Simplified BSD License text as described in Section 4.e of 59 the Trust Legal Provisions and are provided without warranty as 60 described in the Simplified BSD License. 62 Table of Contents 64 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 5 65 2. Pro Audio and Video . . . . . . . . . . . . . . . . . . . . . 6 66 2.1. Use Case Description . . . . . . . . . . . . . . . . . . 6 67 2.1.1. Uninterrupted Stream Playback . . . . . . . . . . . . 7 68 2.1.2. Synchronized Stream Playback . . . . . . . . . . . . 7 69 2.1.3. Sound Reinforcement . . . . . . . . . . . . . . . . . 8 70 2.1.4. Deterministic Time to Establish Streaming . . . . . . 8 71 2.1.5. Secure Transmission . . . . . . . . . . . . . . . . . 8 72 2.1.5.1. Safety . . . . . . . . . . . . . . . . . . . . . 8 73 2.2. Pro Audio Today . . . . . . . . . . . . . . . . . . . . . 9 74 2.3. Pro Audio Future . . . . . . . . . . . . . . . . . . . . 9 75 2.3.1. Layer 3 Interconnecting Layer 2 Islands . . . . . . . 9 76 2.3.2. High Reliability Stream Paths . . . . . . . . . . . . 9 77 2.3.3. Integration of Reserved Streams into IT Networks . . 10 78 2.3.4. Use of Unused Reservations by Best-Effort Traffic . . 10 79 2.3.5. Traffic Segregation . . . . . . . . . . . . . . . . . 10 80 2.3.5.1. Packet Forwarding Rules, VLANs and Subnets . . . 11 81 2.3.5.2. Multicast Addressing (IPv4 and IPv6) . . . . . . 11 82 2.3.6. Latency Optimization by a Central Controller . . . . 11 83 2.3.7. Reduced Device Cost Due To Reduced Buffer Memory . . 11 84 2.4. Pro Audio Asks . . . . . . . . . . . . . . . . . . . . . 12 85 3. Electrical Utilities . . . . . . . . . . . . . . . . . . . . 12 86 3.1. Use Case Description . . . . . . . . . . . . . . . . . . 12 87 3.1.1. Transmission Use Cases . . . . . . . . . . . . . . . 12 88 3.1.1.1. Protection . . . . . . . . . . . . . . . . . . . 13 89 3.1.1.2. Intra-Substation Process Bus Communications . . . 18 90 3.1.1.3. Wide Area Monitoring and Control Systems . . . . 19 91 3.1.1.4. IEC 61850 WAN engineering guidelines requirement 92 classification . . . . . . . . . . . . . . . . . 20 93 3.1.2. Generation Use Case . . . . . . . . . . . . . . . . . 21 94 3.1.2.1. Control of the Generated Power . . . . . . . . . 21 95 3.1.2.2. Control of the Generation Infrastructure . . . . 22 96 3.1.3. Distribution use case . . . . . . . . . . . . . . . . 27 97 3.1.3.1. Fault Location Isolation and Service Restoration 98 (FLISR) . . . . . . . . . . . . . . . . . . . . . 27 99 3.2. Electrical Utilities Today . . . . . . . . . . . . . . . 28 100 3.2.1. Security Current Practices and Limitations . . . . . 28 101 3.3. Electrical Utilities Future . . . . . . . . . . . . . . . 30 102 3.3.1. Migration to Packet-Switched Network . . . . . . . . 31 103 3.3.2. Telecommunications Trends . . . . . . . . . . . . . . 31 104 3.3.2.1. General Telecommunications Requirements . . . . . 31 105 3.3.2.2. Specific Network topologies of Smart Grid 106 Applications . . . . . . . . . . . . . . . . . . 32 107 3.3.2.3. Precision Time Protocol . . . . . . . . . . . . . 33 108 3.3.3. Security Trends in Utility Networks . . . . . . . . . 34 109 3.4. Electrical Utilities Asks . . . . . . . . . . . . . . . . 36 110 4. Building Automation Systems . . . . . . . . . . . . . . . . . 36 111 4.1. Use Case Description . . . . . . . . . . . . . . . . . . 36 112 4.2. Building Automation Systems Today . . . . . . . . . . . . 37 113 4.2.1. BAS Architecture . . . . . . . . . . . . . . . . . . 37 114 4.2.2. BAS Deployment Model . . . . . . . . . . . . . . . . 38 115 4.2.3. Use Cases for Field Networks . . . . . . . . . . . . 40 116 4.2.3.1. Environmental Monitoring . . . . . . . . . . . . 40 117 4.2.3.2. Fire Detection . . . . . . . . . . . . . . . . . 40 118 4.2.3.3. Feedback Control . . . . . . . . . . . . . . . . 41 119 4.2.4. Security Considerations . . . . . . . . . . . . . . . 41 120 4.3. BAS Future . . . . . . . . . . . . . . . . . . . . . . . 41 121 4.4. BAS Asks . . . . . . . . . . . . . . . . . . . . . . . . 42 122 5. Wireless for Industrial . . . . . . . . . . . . . . . . . . . 42 123 5.1. Use Case Description . . . . . . . . . . . . . . . . . . 42 124 5.1.1. Network Convergence using 6TiSCH . . . . . . . . . . 43 125 5.1.2. Common Protocol Development for 6TiSCH . . . . . . . 43 126 5.2. Wireless Industrial Today . . . . . . . . . . . . . . . . 44 127 5.3. Wireless Industrial Future . . . . . . . . . . . . . . . 44 128 5.3.1. Unified Wireless Network and Management . . . . . . . 44 129 5.3.1.1. PCE and 6TiSCH ARQ Retries . . . . . . . . . . . 46 130 5.3.2. Schedule Management by a PCE . . . . . . . . . . . . 47 131 5.3.2.1. PCE Commands and 6TiSCH CoAP Requests . . . . . . 47 132 5.3.2.2. 6TiSCH IP Interface . . . . . . . . . . . . . . . 48 133 5.3.3. 6TiSCH Security Considerations . . . . . . . . . . . 49 134 5.4. Wireless Industrial Asks . . . . . . . . . . . . . . . . 49 135 6. Cellular Radio . . . . . . . . . . . . . . . . . . . . . . . 49 136 6.1. Use Case Description . . . . . . . . . . . . . . . . . . 49 137 6.1.1. Network Architecture . . . . . . . . . . . . . . . . 49 138 6.1.2. Delay Constraints . . . . . . . . . . . . . . . . . . 50 139 6.1.3. Time Synchronization Constraints . . . . . . . . . . 52 140 6.1.4. Transport Loss Constraints . . . . . . . . . . . . . 54 141 6.1.5. Security Considerations . . . . . . . . . . . . . . . 54 142 6.2. Cellular Radio Networks Today . . . . . . . . . . . . . . 55 143 6.2.1. Fronthaul . . . . . . . . . . . . . . . . . . . . . . 55 144 6.2.2. Midhaul and Backhaul . . . . . . . . . . . . . . . . 55 145 6.3. Cellular Radio Networks Future . . . . . . . . . . . . . 56 146 6.4. Cellular Radio Networks Asks . . . . . . . . . . . . . . 58 147 7. Industrial M2M . . . . . . . . . . . . . . . . . . . . . . . 59 148 7.1. Use Case Description . . . . . . . . . . . . . . . . . . 59 149 7.2. Industrial M2M Communication Today . . . . . . . . . . . 60 150 7.2.1. Transport Parameters . . . . . . . . . . . . . . . . 60 151 7.2.2. Stream Creation and Destruction . . . . . . . . . . . 61 152 7.3. Industrial M2M Future . . . . . . . . . . . . . . . . . . 61 153 7.4. Industrial M2M Asks . . . . . . . . . . . . . . . . . . . 62 154 8. Mining Industry . . . . . . . . . . . . . . . . . . . . . . . 62 155 8.1. Use Case Description . . . . . . . . . . . . . . . . . . 62 156 8.2. Mining Industry Today . . . . . . . . . . . . . . . . . . 63 157 8.3. Mining Industry Future . . . . . . . . . . . . . . . . . 63 158 8.4. Mining Industry Asks . . . . . . . . . . . . . . . . . . 64 159 9. Private Blockchain . . . . . . . . . . . . . . . . . . . . . 64 160 9.1. Use Case Description . . . . . . . . . . . . . . . . . . 64 161 9.1.1. Blockchain Operation . . . . . . . . . . . . . . . . 65 162 9.1.2. Blockchain Network Architecture . . . . . . . . . . . 65 163 9.1.3. Security Considerations . . . . . . . . . . . . . . . 66 164 9.2. Private Blockchain Today . . . . . . . . . . . . . . . . 66 165 9.3. Private Blockchain Future . . . . . . . . . . . . . . . . 66 166 9.4. Private Blockchain Asks . . . . . . . . . . . . . . . . . 66 167 10. Network Slicing . . . . . . . . . . . . . . . . . . . . . . . 67 168 10.1. Use Case Description . . . . . . . . . . . . . . . . . . 67 169 10.2. DetNet Applied to Network Slicing . . . . . . . . . . . 67 170 10.2.1. Resource Isolation Across Slices . . . . . . . . . . 67 171 10.2.2. Deterministic Services Within Slices . . . . . . . . 67 172 10.3. A Network Slicing Use Case Example - 5G Bearer Network . 68 173 10.4. Non-5G Applications of Network Slicing . . . . . . . . . 68 174 10.5. Limitations of DetNet in Network Slicing . . . . . . . . 69 175 10.6. Network Slicing Today and Future . . . . . . . . . . . . 69 176 10.7. Network Slicing Asks . . . . . . . . . . . . . . . . . . 69 177 11. Use Case Common Themes . . . . . . . . . . . . . . . . . . . 69 178 11.1. Unified, standards-based network . . . . . . . . . . . . 69 179 11.1.1. Extensions to Ethernet . . . . . . . . . . . . . . . 69 180 11.1.2. Centrally Administered . . . . . . . . . . . . . . . 69 181 11.1.3. Standardized Data Flow Information Models . . . . . 70 182 11.1.4. L2 and L3 Integration . . . . . . . . . . . . . . . 70 183 11.1.5. Consideration for IPv4 . . . . . . . . . . . . . . . 70 184 11.1.6. Guaranteed End-to-End Delivery . . . . . . . . . . . 70 185 11.1.7. Replacement for Multiple Proprietary Deterministic 186 Networks . . . . . . . . . . . . . . . . . . . . . . 70 187 11.1.8. Mix of Deterministic and Best-Effort Traffic . . . . 71 188 11.1.9. Unused Reserved BW to be Available to Best Effort 189 Traffic . . . . . . . . . . . . . . . . . . . . . . 71 190 11.1.10. Lower Cost, Multi-Vendor Solutions . . . . . . . . . 71 192 11.2. Scalable Size . . . . . . . . . . . . . . . . . . . . . 71 193 11.3. Scalable Timing Parameters and Accuracy . . . . . . . . 71 194 11.3.1. Bounded Latency . . . . . . . . . . . . . . . . . . 71 195 11.3.2. Low Latency . . . . . . . . . . . . . . . . . . . . 72 196 11.3.3. Symmetrical Path Delays . . . . . . . . . . . . . . 72 197 11.4. High Reliability and Availability . . . . . . . . . . . 72 198 11.5. Security . . . . . . . . . . . . . . . . . . . . . . . . 72 199 11.6. Deterministic Flows . . . . . . . . . . . . . . . . . . 73 200 12. Use Cases Explicitly Out of Scope for DetNet . . . . . . . . 73 201 12.1. DetNet Scope Limitations . . . . . . . . . . . . . . . . 73 202 12.2. Internet-based Applications . . . . . . . . . . . . . . 74 203 12.2.1. Use Case Description . . . . . . . . . . . . . . . . 74 204 12.2.1.1. Media Content Delivery . . . . . . . . . . . . . 74 205 12.2.1.2. Online Gaming . . . . . . . . . . . . . . . . . 74 206 12.2.1.3. Virtual Reality . . . . . . . . . . . . . . . . 74 207 12.2.2. Internet-Based Applications Today . . . . . . . . . 74 208 12.2.3. Internet-Based Applications Future . . . . . . . . . 74 209 12.2.4. Internet-Based Applications Asks . . . . . . . . . . 75 210 12.3. Pro Audio and Video - Digital Rights Management (DRM) . 75 211 12.4. Pro Audio and Video - Link Aggregation . . . . . . . . . 75 212 13. Contributors . . . . . . . . . . . . . . . . . . . . . . . . 76 213 14. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 77 214 14.1. Pro Audio . . . . . . . . . . . . . . . . . . . . . . . 77 215 14.2. Utility Telecom . . . . . . . . . . . . . . . . . . . . 78 216 14.3. Building Automation Systems . . . . . . . . . . . . . . 78 217 14.4. Wireless for Industrial . . . . . . . . . . . . . . . . 78 218 14.5. Cellular Radio . . . . . . . . . . . . . . . . . . . . . 78 219 14.6. Industrial M2M . . . . . . . . . . . . . . . . . . . . . 79 220 14.7. Internet Applications and CoMP . . . . . . . . . . . . . 79 221 14.8. Electrical Utilities . . . . . . . . . . . . . . . . . . 79 222 14.9. Network Slicing . . . . . . . . . . . . . . . . . . . . 79 223 14.10. Mining . . . . . . . . . . . . . . . . . . . . . . . . . 79 224 14.11. Private Blockchain . . . . . . . . . . . . . . . . . . . 79 225 15. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 79 226 16. Informative References . . . . . . . . . . . . . . . . . . . 79 227 Author's Address . . . . . . . . . . . . . . . . . . . . . . . . 86 229 1. Introduction 231 This draft presents use cases from diverse industries which have in 232 common a need for deterministic flows, but which also differ notably 233 in their network topologies and specific desired behavior. Together, 234 they provide broad industry context for DetNet and a yardstick 235 against which proposed DetNet designs can be measured (to what extent 236 does a proposed design satisfy these various use cases?) 238 For DetNet, use cases explicitly do not define requirements; The 239 DetNet WG will consider the use cases, decide which elements are in 240 scope for DetNet, and the results will be incorporated into future 241 drafts. Similarly, the DetNet use case draft explicitly does not 242 suggest any specific design, architecture or protocols, which will be 243 topics of future drafts. 245 We present for each use case the answers to the following questions: 247 o What is the use case? 249 o How is it addressed today? 251 o How would you like it to be addressed in the future? 253 o What do you want the IETF to deliver? 255 The level of detail in each use case should be sufficient to express 256 the relevant elements of the use case, but not more. 258 At the end we consider the use cases collectively, and examine the 259 most significant goals they have in common. 261 2. Pro Audio and Video 263 2.1. Use Case Description 265 The professional audio and video industry ("ProAV") includes: 267 o Music and film content creation 269 o Broadcast 271 o Cinema 273 o Live sound 275 o Public address, media and emergency systems at large venues 276 (airports, stadiums, churches, theme parks). 278 These industries have already transitioned audio and video signals 279 from analog to digital. However, the digital interconnect systems 280 remain primarily point-to-point with a single (or small number of) 281 signals per link, interconnected with purpose-built hardware. 283 These industries are now transitioning to packet-based infrastructure 284 to reduce cost, increase routing flexibility, and integrate with 285 existing IT infrastructure. 287 Today ProAV applications have no way to establish deterministic flows 288 from a standards-based Layer 3 (IP) interface, which is a fundamental 289 limitation to the use cases described here. Today deterministic 290 flows can be created within standards-based layer 2 LANs (e.g. using 291 IEEE 802.1 AVB) however these are not routable via IP and thus are 292 not effective for distribution over wider areas (for example 293 broadcast events that span wide geographical areas). 295 It would be highly desirable if such flows could be routed over the 296 open Internet, however solutions with more limited scope (e.g. 297 enterprise networks) would still provide a substantial improvement. 299 The following sections describe specific ProAV use cases. 301 2.1.1. Uninterrupted Stream Playback 303 Transmitting audio and video streams for live playback is unlike 304 common file transfer because uninterrupted stream playback in the 305 presence of network errors cannot be achieved by re-trying the 306 transmission; by the time the missing or corrupt packet has been 307 identified it is too late to execute a re-try operation. Buffering 308 can be used to provide enough delay to allow time for one or more 309 retries, however this is not an effective solution in applications 310 where large delays (latencies) are not acceptable (as discussed 311 below). 313 Streams with guaranteed bandwidth can eliminate congestion on the 314 network as a cause of transmission errors that would lead to playback 315 interruption. Use of redundant paths can further mitigate 316 transmission errors to provide greater stream reliability. 318 2.1.2. Synchronized Stream Playback 320 Latency in this context is the time between when a signal is 321 initially sent over a stream and when it is received. A common 322 example in ProAV is time-synchronizing audio and video when they take 323 separate paths through the playback system. In this case the latency 324 of both the audio and video streams must be bounded and consistent if 325 the sound is to remain matched to the movement in the video. A 326 common tolerance for audio/video sync is one NTSC video frame (about 327 33ms) and to maintain the audience perception of correct lip sync the 328 latency needs to be consistent within some reasonable tolerance, for 329 example 10%. 331 A common architecture for synchronizing multiple streams that have 332 different paths through the network (and thus potentially different 333 latencies) is to enable measurement of the latency of each path, and 334 have the data sinks (for example speakers) delay (buffer) all packets 335 on all but the slowest path. Each packet of each stream is assigned 336 a presentation time which is based on the longest required delay. 337 This implies that all sinks must maintain a common time reference of 338 sufficient accuracy, which can be achieved by any of various 339 techniques. 341 This type of architecture is commonly implemented using a central 342 controller that determines path delays and arbitrates buffering 343 delays. 345 2.1.3. Sound Reinforcement 347 Consider the latency (delay) from when a person speaks into a 348 microphone to when their voice emerges from the speaker. If this 349 delay is longer than about 10-15 milliseconds it is noticeable and 350 can make a sound reinforcement system unusable (see slide 6 of 351 [SRP_LATENCY]). (If you have ever tried to speak in the presence of 352 a delayed echo of your voice you may know this experience). 354 Note that the 15ms latency bound includes all parts of the signal 355 path, not just the network, so the network latency must be 356 significantly less than 15ms. 358 In some cases local performers must perform in synchrony with a 359 remote broadcast. In such cases the latencies of the broadcast 360 stream and the local performer must be adjusted to match each other, 361 with a worst case of one video frame (33ms for NTSC video). 363 In cases where audio phase is a consideration, for example beam- 364 forming using multiple speakers, latency can be in the 10 microsecond 365 range (1 audio sample at 96kHz). 367 2.1.4. Deterministic Time to Establish Streaming 369 Note: The WG has decided that guidelines for deterministic time to 370 establish stream startup is not within scope of DetNet. If bounded 371 timing of establishing or re-establish streams is required in a given 372 use case, it is up to the application/system to achieve this. (The 373 supporting text from this section has been removed as of draft 12). 375 2.1.5. Secure Transmission 377 2.1.5.1. Safety 379 Professional audio systems can include amplifiers that are capable of 380 generating hundreds or thousands of watts of audio power which if 381 used incorrectly can cause hearing damage to those in the vicinity. 382 Apart from the usual care required by the systems operators to 383 prevent such incidents, the network traffic that controls these 384 devices must be secured (as with any sensitive application traffic). 386 2.2. Pro Audio Today 388 Some proprietary systems have been created which enable deterministic 389 streams at Layer 3 however they are "engineered networks" which 390 require careful configuration to operate, often require that the 391 system be over-provisioned, and it is implied that all devices on the 392 network voluntarily play by the rules of that network. To enable 393 these industries to successfully transition to an interoperable 394 multi-vendor packet-based infrastructure requires effective open 395 standards, and we believe that establishing relevant IETF standards 396 is a crucial factor. 398 2.3. Pro Audio Future 400 2.3.1. Layer 3 Interconnecting Layer 2 Islands 402 It would be valuable to enable IP to connect multiple Layer 2 LANs. 404 As an example, ESPN recently constructed a state-of-the-art 194,000 405 sq ft, $125 million broadcast studio called DC2. The DC2 network is 406 capable of handling 46 Tbps of throughput with 60,000 simultaneous 407 signals. Inside the facility are 1,100 miles of fiber feeding four 408 audio control rooms (see [ESPN_DC2] ). 410 In designing DC2 they replaced as much point-to-point technology as 411 they could with packet-based technology. They constructed seven 412 individual studios using layer 2 LANS (using IEEE 802.1 AVB) that 413 were entirely effective at routing audio within the LANs. However to 414 interconnect these layer 2 LAN islands together they ended up using 415 dedicated paths in a custom SDN (Software Defined Networking) router 416 because there is no standards-based routing solution available. 418 2.3.2. High Reliability Stream Paths 420 On-air and other live media streams are often backed up with 421 redundant links that seamlessly act to deliver the content when the 422 primary link fails for any reason. In point-to-point systems this is 423 provided by an additional point-to-point link; the analogous 424 requirement in a packet-based system is to provide an alternate path 425 through the network such that no individual link can bring down the 426 system. 428 2.3.3. Integration of Reserved Streams into IT Networks 430 A commonly cited goal of moving to a packet based media 431 infrastructure is that costs can be reduced by using off the shelf, 432 commodity network hardware. In addition, economy of scale can be 433 realized by combining media infrastructure with IT infrastructure. 434 In keeping with these goals, stream reservation technology should be 435 compatible with existing protocols, and not compromise use of the 436 network for best effort (non-time-sensitive) traffic. 438 2.3.4. Use of Unused Reservations by Best-Effort Traffic 440 In cases where stream bandwidth is reserved but not currently used 441 (or is under-utilized) that bandwidth must be available to best- 442 effort (i.e. non-time-sensitive) traffic. For example a single 443 stream may be nailed up (reserved) for specific media content that 444 needs to be presented at different times of the day, ensuring timely 445 delivery of that content, yet in between those times the full 446 bandwidth of the network can be utilized for best-effort tasks such 447 as file transfers. 449 This also addresses a concern of IT network administrators that are 450 considering adding reserved bandwidth traffic to their networks that 451 ("users will reserve large quantities of bandwidth and then never un- 452 reserve it even though they are not using it, and soon the network 453 will have no bandwidth left"). 455 2.3.5. Traffic Segregation 457 Sink devices may be low cost devices with limited processing power. 458 In order to not overwhelm the CPUs in these devices it is important 459 to limit the amount of traffic that these devices must process. 461 As an example, consider the use of individual seat speakers in a 462 cinema. These speakers are typically required to be cost reduced 463 since the quantities in a single theater can reach hundreds of seats. 464 Discovery protocols alone in a one thousand seat theater can generate 465 enough broadcast traffic to overwhelm a low powered CPU. Thus an 466 installation like this will benefit greatly from some type of traffic 467 segregation that can define groups of seats to reduce traffic within 468 each group. All seats in the theater must still be able to 469 communicate with a central controller. 471 There are many techniques that can be used to support this feature 472 including (but not limited to) the following examples. 474 2.3.5.1. Packet Forwarding Rules, VLANs and Subnets 476 Packet forwarding rules can be used to eliminate some extraneous 477 streaming traffic from reaching potentially low powered sink devices, 478 however there may be other types of broadcast traffic that should be 479 eliminated using other means for example VLANs or IP subnets. 481 2.3.5.2. Multicast Addressing (IPv4 and IPv6) 483 Multicast addressing is commonly used to keep bandwidth utilization 484 of shared links to a minimum. 486 Because of the MAC Address forwarding nature of Layer 2 bridges it is 487 important that a multicast MAC address is only associated with one 488 stream. This will prevent reservations from forwarding packets from 489 one stream down a path that has no interested sinks simply because 490 there is another stream on that same path that shares the same 491 multicast MAC address. 493 Since each multicast MAC Address can represent 32 different IPv4 494 multicast addresses there must be a process put in place to make sure 495 this does not occur. Requiring use of IPv6 address can achieve this, 496 however due to their continued prevalence, solutions that are 497 effective for IPv4 installations are also desirable. 499 2.3.6. Latency Optimization by a Central Controller 501 A central network controller might also perform optimizations based 502 on the individual path delays, for example sinks that are closer to 503 the source can inform the controller that they can accept greater 504 latency since they will be buffering packets to match presentation 505 times of farther away sinks. The controller might then move a stream 506 reservation on a short path to a longer path in order to free up 507 bandwidth for other critical streams on that short path. See slides 508 3-5 of [SRP_LATENCY]. 510 Additional optimization can be achieved in cases where sinks have 511 differing latency requirements, for example in a live outdoor concert 512 the speaker sinks have stricter latency requirements than the 513 recording hardware sinks. See slide 7 of [SRP_LATENCY]. 515 2.3.7. Reduced Device Cost Due To Reduced Buffer Memory 517 Device cost can be reduced in a system with guaranteed reservations 518 with a small bounded latency due to the reduced requirements for 519 buffering (i.e. memory) on sink devices. For example, a theme park 520 might broadcast a live event across the globe via a layer 3 protocol; 521 in such cases the size of the buffers required is proportional to the 522 latency bounds and jitter caused by delivery, which depends on the 523 worst case segment of the end-to-end network path. For example on 524 todays open internet the latency is typically unacceptable for audio 525 and video streaming without many seconds of buffering. In such 526 scenarios a single gateway device at the local network that receives 527 the feed from the remote site would provide the expensive buffering 528 required to mask the latency and jitter issues associated with long 529 distance delivery. Sink devices in the local location would have no 530 additional buffering requirements, and thus no additional costs, 531 beyond those required for delivery of local content. The sink device 532 would be receiving the identical packets as those sent by the source 533 and would be unaware that there were any latency or jitter issues 534 along the path. 536 2.4. Pro Audio Asks 538 o Layer 3 routing on top of AVB (and/or other high QoS networks) 540 o Content delivery with bounded, lowest possible latency 542 o IntServ and DiffServ integration with AVB (where practical) 544 o Single network for A/V and IT traffic 546 o Standards-based, interoperable, multi-vendor 548 o IT department friendly 550 o Enterprise-wide networks (e.g. size of San Francisco but not the 551 whole Internet (yet...)) 553 3. Electrical Utilities 555 3.1. Use Case Description 557 Many systems that an electrical utility deploys today rely on high 558 availability and deterministic behavior of the underlying networks. 559 Here we present use cases in Transmission, Generation and 560 Distribution, including key timing and reliability metrics. We also 561 discuss security issues and industry trends which affect the 562 architecture of next generation utility networks 564 3.1.1. Transmission Use Cases 565 3.1.1.1. Protection 567 Protection means not only the protection of human operators but also 568 the protection of the electrical equipment and the preservation of 569 the stability and frequency of the grid. If a fault occurs in the 570 transmission or distribution of electricity then severe damage can 571 occur to human operators, electrical equipment and the grid itself, 572 leading to blackouts. 574 Communication links in conjunction with protection relays are used to 575 selectively isolate faults on high voltage lines, transformers, 576 reactors and other important electrical equipment. The role of the 577 teleprotection system is to selectively disconnect a faulty part by 578 transferring command signals within the shortest possible time. 580 3.1.1.1.1. Key Criteria 582 The key criteria for measuring teleprotection performance are command 583 transmission time, dependability and security. These criteria are 584 defined by the IEC standard 60834 as follows: 586 o Transmission time (Speed): The time between the moment where state 587 changes at the transmitter input and the moment of the 588 corresponding change at the receiver output, including propagation 589 delay. Overall operating time for a teleprotection system 590 includes the time for initiating the command at the transmitting 591 end, the propagation delay over the network (including equipments) 592 and the selection and decision time at the receiving end, 593 including any additional delay due to a noisy environment. 595 o Dependability: The ability to issue and receive valid commands in 596 the presence of interference and/or noise, by minimizing the 597 probability of missing command (PMC). Dependability targets are 598 typically set for a specific bit error rate (BER) level. 600 o Security: The ability to prevent false tripping due to a noisy 601 environment, by minimizing the probability of unwanted commands 602 (PUC). Security targets are also set for a specific bit error 603 rate (BER) level. 605 Additional elements of the the teleprotection system that impact its 606 performance include: 608 o Network bandwidth 610 o Failure recovery capacity (aka resiliency) 612 3.1.1.1.2. Fault Detection and Clearance Timing 614 Most power line equipment can tolerate short circuits or faults for 615 up to approximately five power cycles before sustaining irreversible 616 damage or affecting other segments in the network. This translates 617 to total fault clearance time of 100ms. As a safety precaution, 618 however, actual operation time of protection systems is limited to 619 70- 80 percent of this period, including fault recognition time, 620 command transmission time and line breaker switching time. 622 Some system components, such as large electromechanical switches, 623 require particularly long time to operate and take up the majority of 624 the total clearance time, leaving only a 10ms window for the 625 telecommunications part of the protection scheme, independent of the 626 distance to travel. Given the sensitivity of the issue, new networks 627 impose requirements that are even more stringent: IEC standard 61850 628 limits the transfer time for protection messages to 1/4 - 1/2 cycle 629 or 4 - 8ms (for 60Hz lines) for the most critical messages. 631 3.1.1.1.3. Symmetric Channel Delay 633 Teleprotection channels which are differential must be synchronous, 634 which means that any delays on the transmit and receive paths must 635 match each other. Teleprotection systems ideally support zero 636 asymmetric delay; typical legacy relays can tolerate delay 637 discrepancies of up to 750us. 639 Some tools available for lowering delay variation below this 640 threshold are: 642 o For legacy systems using Time Division Multiplexing (TDM), jitter 643 buffers at the multiplexers on each end of the line can be used to 644 offset delay variation by queuing sent and received packets. The 645 length of the queues must balance the need to regulate the rate of 646 transmission with the need to limit overall delay, as larger 647 buffers result in increased latency. 649 o For jitter-prone IP packet networks, traffic management tools can 650 ensure that the teleprotection signals receive the highest 651 transmission priority to minimize jitter. 653 o Standard packet-based synchronization technologies, such as 654 1588-2008 Precision Time Protocol (PTP) and Synchronous Ethernet 655 (Sync-E), can help keep networks stable by maintaining a highly 656 accurate clock source on the various network devices. 658 3.1.1.1.4. Teleprotection Network Requirements (IEC 61850) 660 The following table captures the main network metrics as based on the 661 IEC 61850 standard. 663 +-----------------------------+-------------------------------------+ 664 | Teleprotection Requirement | Attribute | 665 +-----------------------------+-------------------------------------+ 666 | One way maximum delay | 4-10 ms | 667 | Asymetric delay required | Yes | 668 | Maximum jitter | less than 250 us (750 us for legacy | 669 | | IED) | 670 | Topology | Point to point, point to Multi- | 671 | | point | 672 | Availability | 99.9999 | 673 | precise timing required | Yes | 674 | Recovery time on node | less than 50ms - hitless | 675 | failure | | 676 | performance management | Yes, Mandatory | 677 | Redundancy | Yes | 678 | Packet loss | 0.1% to 1% | 679 +-----------------------------+-------------------------------------+ 681 Table 1: Teleprotection network requirements 683 3.1.1.1.5. Inter-Trip Protection scheme 685 "Inter-tripping" is the signal-controlled tripping of a circuit 686 breaker to complete the isolation of a circuit or piece of apparatus 687 in concert with the tripping of other circuit breakers. 689 +--------------------------------+----------------------------------+ 690 | Inter-Trip protection | Attribute | 691 | Requirement | | 692 +--------------------------------+----------------------------------+ 693 | One way maximum delay | 5 ms | 694 | Asymetric delay required | No | 695 | Maximum jitter | Not critical | 696 | Topology | Point to point, point to Multi- | 697 | | point | 698 | Bandwidth | 64 Kbps | 699 | Availability | 99.9999 | 700 | precise timing required | Yes | 701 | Recovery time on node failure | less than 50ms - hitless | 702 | performance management | Yes, Mandatory | 703 | Redundancy | Yes | 704 | Packet loss | 0.1% | 705 +--------------------------------+----------------------------------+ 707 Table 2: Inter-Trip protection network requirements 709 3.1.1.1.6. Current Differential Protection Scheme 711 Current differential protection is commonly used for line protection, 712 and is typical for protecting parallel circuits. At both end of the 713 lines the current is measured by the differential relays, and both 714 relays will trip the circuit breaker if the current going into the 715 line does not equal the current going out of the line. This type of 716 protection scheme assumes some form of communications being present 717 between the relays at both end of the line, to allow both relays to 718 compare measured current values. Line differential protection 719 schemes assume a very low telecommunications delay between both 720 relays, often as low as 5ms. Moreover, as those systems are often 721 not time-synchronized, they also assume symmetric telecommunications 722 paths with constant delay, which allows comparing current measurement 723 values taken at the exact same time. 725 +----------------------------------+--------------------------------+ 726 | Current Differential protection | Attribute | 727 | Requirement | | 728 +----------------------------------+--------------------------------+ 729 | One way maximum delay | 5 ms | 730 | Asymetric delay Required | Yes | 731 | Maximum jitter | less than 250 us (750us for | 732 | | legacy IED) | 733 | Topology | Point to point, point to | 734 | | Multi-point | 735 | Bandwidth | 64 Kbps | 736 | Availability | 99.9999 | 737 | precise timing required | Yes | 738 | Recovery time on node failure | less than 50ms - hitless | 739 | performance management | Yes, Mandatory | 740 | Redundancy | Yes | 741 | Packet loss | 0.1% | 742 +----------------------------------+--------------------------------+ 744 Table 3: Current Differential Protection metrics 746 3.1.1.1.7. Distance Protection Scheme 748 Distance (Impedance Relay) protection scheme is based on voltage and 749 current measurements. The network metrics are similar (but not 750 identical to) Current Differential protection. 752 +-------------------------------+-----------------------------------+ 753 | Distance protection | Attribute | 754 | Requirement | | 755 +-------------------------------+-----------------------------------+ 756 | One way maximum delay | 5 ms | 757 | Asymetric delay Required | No | 758 | Maximum jitter | Not critical | 759 | Topology | Point to point, point to Multi- | 760 | | point | 761 | Bandwidth | 64 Kbps | 762 | Availability | 99.9999 | 763 | precise timing required | Yes | 764 | Recovery time on node failure | less than 50ms - hitless | 765 | performance management | Yes, Mandatory | 766 | Redundancy | Yes | 767 | Packet loss | 0.1% | 768 +-------------------------------+-----------------------------------+ 770 Table 4: Distance Protection requirements 772 3.1.1.1.8. Inter-Substation Protection Signaling 774 This use case describes the exchange of Sampled Value and/or GOOSE 775 (Generic Object Oriented Substation Events) message between 776 Intelligent Electronic Devices (IED) in two substations for 777 protection and tripping coordination. The two IEDs are in a master- 778 slave mode. 780 The Current Transformer or Voltage Transformer (CT/VT) in one 781 substation sends the sampled analog voltage or current value to the 782 Merging Unit (MU) over hard wire. The MU sends the time-synchronized 783 61850-9-2 sampled values to the slave IED. The slave IED forwards 784 the information to the Master IED in the other substation. The 785 master IED makes the determination (for example based on sampled 786 value differentials) to send a trip command to the originating IED. 787 Once the slave IED/Relay receives the GOOSE trip for breaker 788 tripping, it opens the breaker. It then sends a confirmation message 789 back to the master. All data exchanges between IEDs are either 790 through Sampled Value and/or GOOSE messages. 792 +----------------------------------+--------------------------------+ 793 | Inter-Substation protection | Attribute | 794 | Requirement | | 795 +----------------------------------+--------------------------------+ 796 | One way maximum delay | 5 ms | 797 | Asymetric delay Required | No | 798 | Maximum jitter | Not critical | 799 | Topology | Point to point, point to | 800 | | Multi-point | 801 | Bandwidth | 64 Kbps | 802 | Availability | 99.9999 | 803 | precise timing required | Yes | 804 | Recovery time on node failure | less than 50ms - hitless | 805 | performance management | Yes, Mandatory | 806 | Redundancy | Yes | 807 | Packet loss | 1% | 808 +----------------------------------+--------------------------------+ 810 Table 5: Inter-Substation Protection requirements 812 3.1.1.2. Intra-Substation Process Bus Communications 814 This use case describes the data flow from the CT/VT to the IEDs in 815 the substation via the MU. The CT/VT in the substation send the 816 analog voltage or current values to the MU over hard wire. The MU 817 converts the analog values into digital format (typically time- 818 synchronized Sampled Values as specified by IEC 61850-9-2) and sends 819 them to the IEDs in the substation. The GPS Master Clock can send 820 1PPS or IRIG-B format to the MU through a serial port or IEEE 1588 821 protocol via a network. Process bus communication using 61850 822 simplifies connectivity within the substation and removes the 823 requirement for multiple serial connections and removes the slow 824 serial bus architectures that are typically used. This also ensures 825 increased flexibility and increased speed with the use of multicast 826 messaging between multiple devices. 828 +----------------------------------+--------------------------------+ 829 | Intra-Substation protection | Attribute | 830 | Requirement | | 831 +----------------------------------+--------------------------------+ 832 | One way maximum delay | 5 ms | 833 | Asymetric delay Required | No | 834 | Maximum jitter | Not critical | 835 | Topology | Point to point, point to | 836 | | Multi-point | 837 | Bandwidth | 64 Kbps | 838 | Availability | 99.9999 | 839 | precise timing required | Yes | 840 | Recovery time on Node failure | less than 50ms - hitless | 841 | performance management | Yes, Mandatory | 842 | Redundancy | Yes - No | 843 | Packet loss | 0.1% | 844 +----------------------------------+--------------------------------+ 846 Table 6: Intra-Substation Protection requirements 848 3.1.1.3. Wide Area Monitoring and Control Systems 850 The application of synchrophasor measurement data from Phasor 851 Measurement Units (PMU) to Wide Area Monitoring and Control Systems 852 promises to provide important new capabilities for improving system 853 stability. Access to PMU data enables more timely situational 854 awareness over larger portions of the grid than what has been 855 possible historically with normal SCADA (Supervisory Control and Data 856 Acquisition) data. Handling the volume and real-time nature of 857 synchrophasor data presents unique challenges for existing 858 application architectures. Wide Area management System (WAMS) makes 859 it possible for the condition of the bulk power system to be observed 860 and understood in real-time so that protective, preventative, or 861 corrective action can be taken. Because of the very high sampling 862 rate of measurements and the strict requirement for time 863 synchronization of the samples, WAMS has stringent telecommunications 864 requirements in an IP network that are captured in the following 865 table: 867 +----------------------+--------------------------------------------+ 868 | WAMS Requirement | Attribute | 869 +----------------------+--------------------------------------------+ 870 | One way maximum | 50 ms | 871 | delay | | 872 | Asymetric delay | No | 873 | Required | | 874 | Maximum jitter | Not critical | 875 | Topology | Point to point, point to Multi-point, | 876 | | Multi-point to Multi-point | 877 | Bandwidth | 100 Kbps | 878 | Availability | 99.9999 | 879 | precise timing | Yes | 880 | required | | 881 | Recovery time on | less than 50ms - hitless | 882 | Node failure | | 883 | performance | Yes, Mandatory | 884 | management | | 885 | Redundancy | Yes | 886 | Packet loss | 1% | 887 | Consecutive Packet | At least 1 packet per application cycle | 888 | Loss | must be received. | 889 +----------------------+--------------------------------------------+ 891 Table 7: WAMS Special Communication Requirements 893 3.1.1.4. IEC 61850 WAN engineering guidelines requirement 894 classification 896 The IEC (International Electrotechnical Commission) has recently 897 published a Technical Report which offers guidelines on how to define 898 and deploy Wide Area Networks for the interconnections of electric 899 substations, generation plants and SCADA operation centers. The IEC 900 61850-90-12 is providing a classification of WAN communication 901 requirements into 4 classes. Table 8 summarizes these requirements: 903 +----------------+------------+------------+------------+-----------+ 904 | WAN | Class WA | Class WB | Class WC | Class WD | 905 | Requirement | | | | | 906 +----------------+------------+------------+------------+-----------+ 907 | Application | EHV (Extra | HV (High | MV (Medium | General | 908 | field | High | Voltage) | Voltage) | purpose | 909 | | Voltage) | | | | 910 | Latency | 5 ms | 10 ms | 100 ms | > 100 ms | 911 | Jitter | 10 us | 100 us | 1 ms | 10 ms | 912 | Latency | 100 us | 1 ms | 10 ms | 100 ms | 913 | Asymetry | | | | | 914 | Time Accuracy | 1 us | 10 us | 100 us | 10 to 100 | 915 | | | | | ms | 916 | Bit Error rate | 10-7 to | 10-5 to | 10-3 | | 917 | | 10-6 | 10-4 | | | 918 | Unavailability | 10-7 to | 10-5 to | 10-3 | | 919 | | 10-6 | 10-4 | | | 920 | Recovery delay | Zero | 50 ms | 5 s | 50 s | 921 | Cyber security | extremely | High | Medium | Medium | 922 | | high | | | | 923 +----------------+------------+------------+------------+-----------+ 925 Table 8: 61850-90-12 Communication Requirements; Courtesy of IEC 927 3.1.2. Generation Use Case 929 Energy generation systems are complex infrastructures that require 930 control of both the generated power and the generation 931 infrastructure. 933 3.1.2.1. Control of the Generated Power 935 The electrical power generation frequency must be maintained within a 936 very narrow band. Deviations from the acceptable frequency range are 937 detected and the required signals are sent to the power plants for 938 frequency regulation. 940 Automatic Generation Control (AGC) is a system for adjusting the 941 power output of generators at different power plants, in response to 942 changes in the load. 944 +---------------------------------------------------+---------------+ 945 | FCAG (Frequency Control Automatic Generation) | Attribute | 946 | Requirement | | 947 +---------------------------------------------------+---------------+ 948 | One way maximum delay | 500 ms | 949 | Asymetric delay Required | No | 950 | Maximum jitter | Not critical | 951 | Topology | Point to | 952 | | point | 953 | Bandwidth | 20 Kbps | 954 | Availability | 99.999 | 955 | precise timing required | Yes | 956 | Recovery time on Node failure | N/A | 957 | performance management | Yes, | 958 | | Mandatory | 959 | Redundancy | Yes | 960 | Packet loss | 1% | 961 +---------------------------------------------------+---------------+ 963 Table 9: FCAG Communication Requirements 965 3.1.2.2. Control of the Generation Infrastructure 967 The control of the generation infrastructure combines requirements 968 from industrial automation systems and energy generation systems. In 969 this section we present the use case of the control of the generation 970 infrastructure of a wind turbine. 972 | 973 | 974 | +-----------------+ 975 | | +----+ | 976 | | |WTRM| WGEN | 977 WROT x==|===| | | 978 | | +----+ WCNV| 979 | |WNAC | 980 | +---+---WYAW---+--+ 981 | | | 982 | | | +----+ 983 |WTRF | |WMET| 984 | | | | 985 Wind Turbine | +--+-+ 986 Controller | | 987 WTUR | | | 988 WREP | | | 989 WSLG | | | 990 WALG | WTOW | | 992 Figure 1: Wind Turbine Control Network 994 Figure 1 presents the subsystems that operate a wind turbine. These 995 subsystems include 997 o WROT (Rotor Control) 999 o WNAC (Nacelle Control) (nacelle: housing containing the generator) 1001 o WTRM (Transmission Control) 1003 o WGEN (Generator) 1005 o WYAW (Yaw Controller) (of the tower head) 1007 o WCNV (In-Turbine Power Converter) 1009 o WMET (External Meteorological Station providing real time 1010 information to the controllers of the tower) 1012 Traffic characteristics relevant for the network planning and 1013 dimensioning process in a wind turbine scenario are listed below. 1014 The values in this section are based mainly on the relevant 1015 references [Ahm14] and [Spe09]. Each logical node (Figure 1) is a 1016 part of the metering network and produces analog measurements and 1017 status information which must comply with their respective data rate 1018 constraints. 1020 +-----------+--------+--------+-------------+---------+-------------+ 1021 | Subsystem | Sensor | Analog | Data Rate | Status | Data rate | 1022 | | Count | Sample | (bytes/sec) | Sample | (bytes/sec) | 1023 | | | Count | | Count | | 1024 +-----------+--------+--------+-------------+---------+-------------+ 1025 | WROT | 14 | 9 | 642 | 5 | 10 | 1026 | WTRM | 18 | 10 | 2828 | 8 | 16 | 1027 | WGEN | 14 | 12 | 73764 | 2 | 4 | 1028 | WCNV | 14 | 12 | 74060 | 2 | 4 | 1029 | WTRF | 12 | 5 | 73740 | 2 | 4 | 1030 | WNAC | 12 | 9 | 112 | 3 | 6 | 1031 | WYAW | 7 | 8 | 220 | 4 | 8 | 1032 | WTOW | 4 | 1 | 8 | 3 | 6 | 1033 | WMET | 7 | 7 | 228 | - | - | 1034 +-----------+--------+--------+-------------+---------+-------------+ 1036 Table 10: Wind Turbine Data Rate Constraints 1038 Quality of Service (QoS) constraints for different services are 1039 presented in Table 11. These constraints are defined by IEEE 1646 1040 standard [IEEE1646] and IEC 61400 standard [IEC61400]. 1042 +---------------------+---------+-------------+---------------------+ 1043 | Service | Latency | Reliability | Packet Loss Rate | 1044 +---------------------+---------+-------------+---------------------+ 1045 | Analogue measure | 16 ms | 99.99% | < 10-6 | 1046 | Status information | 16 ms | 99.99% | < 10-6 | 1047 | Protection traffic | 4 ms | 100.00% | < 10-9 | 1048 | Reporting and | 1 s | 99.99% | < 10-6 | 1049 | logging | | | | 1050 | Video surveillance | 1 s | 99.00% | No specific | 1051 | | | | requirement | 1052 | Internet connection | 60 min | 99.00% | No specific | 1053 | | | | requirement | 1054 | Control traffic | 16 ms | 100.00% | < 10-9 | 1055 | Data polling | 16 ms | 99.99% | < 10-6 | 1056 +---------------------+---------+-------------+---------------------+ 1058 Table 11: Wind Turbine Reliability and Latency Constraints 1060 3.1.2.2.1. Intra-Domain Network Considerations 1062 A wind turbine is composed of a large set of subsystems including 1063 sensors and actuators which require time-critical operation. The 1064 reliability and latency constraints of these different subsystems is 1065 shown in Table 11. These subsystems are connected to an intra-domain 1066 network which is used to monitor and control the operation of the 1067 turbine and connect it to the SCADA subsystems. The different 1068 components are interconnected using fiber optics, industrial buses, 1069 industrial Ethernet, EtherCat, or a combination of them. Industrial 1070 signaling and control protocols such as Modbus, Profibus, Profinet 1071 and EtherCat are used directly on top of the Layer 2 transport or 1072 encapsulated over TCP/IP. 1074 The Data collected from the sensors and condition monitoring systems 1075 is multiplexed onto fiber cables for transmission to the base of the 1076 tower, and to remote control centers. The turbine controller 1077 continuously monitors the condition of the wind turbine and collects 1078 statistics on its operation. This controller also manages a large 1079 number of switches, hydraulic pumps, valves, and motors within the 1080 wind turbine. 1082 There is usually a controller both at the bottom of the tower and in 1083 the nacelle. The communication between these two controllers usually 1084 takes place using fiber optics instead of copper links. Sometimes, a 1085 third controller is installed in the hub of the rotor and manages the 1086 pitch of the blades. That unit usually communicates with the nacelle 1087 unit using serial communications. 1089 3.1.2.2.2. Inter-Domain network considerations 1091 A remote control center belonging to a grid operator regulates the 1092 power output, enables remote actuation, and monitors the health of 1093 one or more wind parks in tandem. It connects to the local control 1094 center in a wind park over the Internet (Figure 2) via firewalls at 1095 both ends. The AS path between the local control center and the Wind 1096 Park typically involves several ISPs at different tiers. For 1097 example, a remote control center in Denmark can regulate a wind park 1098 in Greece over the normal public AS path between the two locations. 1100 The remote control center is part of the SCADA system, setting the 1101 desired power output to the wind park and reading back the result 1102 once the new power output level has been set. Traffic between the 1103 remote control center and the wind park typically consists of 1104 protocols like IEC 60870-5-104 [IEC-60870-5-104], OPC XML-DA 1105 [OPCXML], Modbus [MODBUS], and SNMP [RFC3411]. Currently, traffic 1106 flows between the wind farm and the remote control center are best 1107 effort. QoS requirements are not strict, so no SLAs or service 1108 provisioning mechanisms (e.g., VPN) are employed. In case of events 1109 like equipment failure, tolerance for alarm delay is on the order of 1110 minutes, due to redundant systems already in place. 1112 +--------------+ 1113 | | 1114 | | 1115 | Wind Park #1 +----+ 1116 | | | XXXXXX 1117 | | | X XXXXXXXX +----------------+ 1118 +--------------+ | XXXX X XXXXX | | 1119 +---+ XXX | Remote Control | 1120 XXX Internet +----+ Center | 1121 +----+X XXX | | 1122 +--------------+ | XXXXXXX XX | | 1123 | | | XX XXXXXXX +----------------+ 1124 | | | XXXXX 1125 | Wind Park #2 +----+ 1126 | | 1127 | | 1128 +--------------+ 1130 Figure 2: Wind Turbine Control via Internet 1132 We expect future use cases which require bounded latency, bounded 1133 jitter and extraordinary low packet loss for inter-domain traffic 1134 flows due to the softwarization and virtualization of core wind farm 1135 equipment (e.g. switches, firewalls and SCADA server components). 1136 These factors will create opportunities for service providers to 1137 install new services and dynamically manage them from remote 1138 locations. For example, to enable fail-over of a local SCADA server, 1139 a SCADA server in another wind farm site (under the administrative 1140 control of the same operator) could be utilized temporarily 1141 (Figure 3). In that case local traffic would be forwarded to the 1142 remote SCADA server and existing intra-domain QoS and timing 1143 parameters would have to be met for inter-domain traffic flows. 1145 +--------------+ 1146 | | 1147 | | 1148 | Wind Park #1 +----+ 1149 | | | XXXXXX 1150 | | | X XXXXXXXX +----------------+ 1151 +--------------+ | XXXX XXXXX | | 1152 +---+ Operator XXX | Remote Control | 1153 XXX Administered +----+ Center | 1154 +----+X WAN XXX | | 1155 +--------------+ | XXXXXXX XX | | 1156 | | | XX XXXXXXX +----------------+ 1157 | | | XXXXX 1158 | Wind Park #2 +----+ 1159 | | 1160 | | 1161 +--------------+ 1163 Figure 3: Wind Turbine Control via Operator Administered WAN 1165 3.1.3. Distribution use case 1167 3.1.3.1. Fault Location Isolation and Service Restoration (FLISR) 1169 Fault Location, Isolation, and Service Restoration (FLISR) refers to 1170 the ability to automatically locate the fault, isolate the fault, and 1171 restore service in the distribution network. This will likely be the 1172 first widespread application of distributed intelligence in the grid. 1174 Static power switch status (open/closed) in the network dictates the 1175 power flow to secondary substations. Reconfiguring the network in 1176 the event of a fault is typically done manually on site to energize/ 1177 de-energize alternate paths. Automating the operation of substation 1178 switchgear allows the flow of power to be altered automatically under 1179 fault conditions. 1181 FLISR can be managed centrally from a Distribution Management System 1182 (DMS) or executed locally through distributed control via intelligent 1183 switches and fault sensors. 1185 +----------------------+--------------------------------------------+ 1186 | FLISR Requirement | Attribute | 1187 +----------------------+--------------------------------------------+ 1188 | One way maximum | 80 ms | 1189 | delay | | 1190 | Asymetric delay | No | 1191 | Required | | 1192 | Maximum jitter | 40 ms | 1193 | Topology | Point to point, point to Multi-point, | 1194 | | Multi-point to Multi-point | 1195 | Bandwidth | 64 Kbps | 1196 | Availability | 99.9999 | 1197 | precise timing | Yes | 1198 | required | | 1199 | Recovery time on | Depends on customer impact | 1200 | Node failure | | 1201 | performance | Yes, Mandatory | 1202 | management | | 1203 | Redundancy | Yes | 1204 | Packet loss | 0.1% | 1205 +----------------------+--------------------------------------------+ 1207 Table 12: FLISR Communication Requirements 1209 3.2. Electrical Utilities Today 1211 Many utilities still rely on complex environments formed of multiple 1212 application-specific proprietary networks, including TDM networks. 1214 In this kind of environment there is no mixing of OT and IT 1215 applications on the same network, and information is siloed between 1216 operational areas. 1218 Specific calibration of the full chain is required, which is costly. 1220 This kind of environment prevents utility operations from realizing 1221 the operational efficiency benefits, visibility, and functional 1222 integration of operational information across grid applications and 1223 data networks. 1225 In addition, there are many security-related issues as discussed in 1226 the following section. 1228 3.2.1. Security Current Practices and Limitations 1230 Grid monitoring and control devices are already targets for cyber 1231 attacks, and legacy telecommunications protocols have many intrinsic 1232 network-related vulnerabilities. For example, DNP3, Modbus, 1233 PROFIBUS/PROFINET, and other protocols are designed around a common 1234 paradigm of request and respond. Each protocol is designed for a 1235 master device such as an HMI (Human Machine Interface) system to send 1236 commands to subordinate slave devices to retrieve data (reading 1237 inputs) or control (writing to outputs). Because many of these 1238 protocols lack authentication, encryption, or other basic security 1239 measures, they are prone to network-based attacks, allowing a 1240 malicious actor or attacker to utilize the request-and-respond system 1241 as a mechanism for command-and-control like functionality. Specific 1242 security concerns common to most industrial control, including 1243 utility telecommunication protocols include the following: 1245 o Network or transport errors (e.g. malformed packets or excessive 1246 latency) can cause protocol failure. 1248 o Protocol commands may be available that are capable of forcing 1249 slave devices into inoperable states, including powering-off 1250 devices, forcing them into a listen-only state, disabling 1251 alarming. 1253 o Protocol commands may be available that are capable of restarting 1254 communications and otherwise interrupting processes. 1256 o Protocol commands may be available that are capable of clearing, 1257 erasing, or resetting diagnostic information such as counters and 1258 diagnostic registers. 1260 o Protocol commands may be available that are capable of requesting 1261 sensitive information about the controllers, their configurations, 1262 or other need-to-know information. 1264 o Most protocols are application layer protocols transported over 1265 TCP; therefore it is easy to transport commands over non-standard 1266 ports or inject commands into authorized traffic flows. 1268 o Protocol commands may be available that are capable of 1269 broadcasting messages to many devices at once (i.e. a potential 1270 DoS). 1272 o Protocol commands may be available to query the device network to 1273 obtain defined points and their values (i.e. a configuration 1274 scan). 1276 o Protocol commands may be available that will list all available 1277 function codes (i.e. a function scan). 1279 These inherent vulnerabilities, along with increasing connectivity 1280 between IT an OT networks, make network-based attacks very feasible. 1282 Simple injection of malicious protocol commands provides control over 1283 the target process. Altering legitimate protocol traffic can also 1284 alter information about a process and disrupt the legitimate controls 1285 that are in place over that process. A man-in-the-middle attack 1286 could provide both control over a process and misrepresentation of 1287 data back to operator consoles. 1289 3.3. Electrical Utilities Future 1291 The business and technology trends that are sweeping the utility 1292 industry will drastically transform the utility business from the way 1293 it has been for many decades. At the core of many of these changes 1294 is a drive to modernize the electrical grid with an integrated 1295 telecommunications infrastructure. However, interoperability 1296 concerns, legacy networks, disparate tools, and stringent security 1297 requirements all add complexity to the grid transformation. Given 1298 the range and diversity of the requirements that should be addressed 1299 by the next generation telecommunications infrastructure, utilities 1300 need to adopt a holistic architectural approach to integrate the 1301 electrical grid with digital telecommunications across the entire 1302 power delivery chain. 1304 The key to modernizing grid telecommunications is to provide a 1305 common, adaptable, multi-service network infrastructure for the 1306 entire utility organization. Such a network serves as the platform 1307 for current capabilities while enabling future expansion of the 1308 network to accommodate new applications and services. 1310 To meet this diverse set of requirements, both today and in the 1311 future, the next generation utility telecommunnications network will 1312 be based on open-standards-based IP architecture. An end-to-end IP 1313 architecture takes advantage of nearly three decades of IP technology 1314 development, facilitating interoperability and device management 1315 across disparate networks and devices, as it has been already 1316 demonstrated in many mission-critical and highly secure networks. 1318 IPv6 is seen as a future telecommunications technology for the Smart 1319 Grid; the IEC (International Electrotechnical Commission) and 1320 different National Committees have mandated a specific adhoc group 1321 (AHG8) to define the migration strategy to IPv6 for all the IEC TC57 1322 power automation standards. The AHG8 has recently finalised the work 1323 on the migration strategy and the following Technical Report has been 1324 issued: IEC TR 62357-200:2015: Guidelines for migration from Internet 1325 Protocol version 4 (IPv4) to Internet Protocol version 6 (IPv6). 1327 We expect cloud-based SCADA systems to control and monitor the 1328 critical and non-critical subsystems of generation systems, for 1329 example wind farms. 1331 3.3.1. Migration to Packet-Switched Network 1333 Throughout the world, utilities are increasingly planning for a 1334 future based on smart grid applications requiring advanced 1335 telecommunications systems. Many of these applications utilize 1336 packet connectivity for communicating information and control signals 1337 across the utility's Wide Area Network (WAN), made possible by 1338 technologies such as multiprotocol label switching (MPLS). The data 1339 that traverses the utility WAN includes: 1341 o Grid monitoring, control, and protection data 1343 o Non-control grid data (e.g. asset data for condition-based 1344 monitoring) 1346 o Physical safety and security data (e.g. voice and video) 1348 o Remote worker access to corporate applications (voice, maps, 1349 schematics, etc.) 1351 o Field area network backhaul for smart metering, and distribution 1352 grid management 1354 o Enterprise traffic (email, collaboration tools, business 1355 applications) 1357 WANs support this wide variety of traffic to and from substations, 1358 the transmission and distribution grid, generation sites, between 1359 control centers, and between work locations and data centers. To 1360 maintain this rapidly expanding set of applications, many utilities 1361 are taking steps to evolve present time-division multiplexing (TDM) 1362 based and frame relay infrastructures to packet systems. Packet- 1363 based networks are designed to provide greater functionalities and 1364 higher levels of service for applications, while continuing to 1365 deliver reliability and deterministic (real-time) traffic support. 1367 3.3.2. Telecommunications Trends 1369 These general telecommunications topics are in addition to the use 1370 cases that have been addressed so far. These include both current 1371 and future telecommunications related topics that should be factored 1372 into the network architecture and design. 1374 3.3.2.1. General Telecommunications Requirements 1376 o IP Connectivity everywhere 1378 o Monitoring services everywhere and from different remote centers 1379 o Move services to a virtual data center 1381 o Unify access to applications / information from the corporate 1382 network 1384 o Unify services 1386 o Unified Communications Solutions 1388 o Mix of fiber and microwave technologies - obsolescence of SONET/ 1389 SDH or TDM 1391 o Standardize grid telecommunications protocol to opened standard to 1392 ensure interoperability 1394 o Reliable Telecommunications for Transmission and Distribution 1395 Substations 1397 o IEEE 1588 time synchronization Client / Server Capabilities 1399 o Integration of Multicast Design 1401 o QoS Requirements Mapping 1403 o Enable Future Network Expansion 1405 o Substation Network Resilience 1407 o Fast Convergence Design 1409 o Scalable Headend Design 1411 o Define Service Level Agreements (SLA) and Enable SLA Monitoring 1413 o Integration of 3G/4G Technologies and future technologies 1415 o Ethernet Connectivity for Station Bus Architecture 1417 o Ethernet Connectivity for Process Bus Architecture 1419 o Protection, teleprotection and PMU (Phaser Measurement Unit) on IP 1421 3.3.2.2. Specific Network topologies of Smart Grid Applications 1423 Utilities often have very large private telecommunications networks. 1424 It covers an entire territory / country. The main purpose of the 1425 network, until now, has been to support transmission network 1426 monitoring, control, and automation, remote control of generation 1427 sites, and providing FCAPS (Fault, Configuration, Accounting, 1428 Performance, Security) services from centralized network operation 1429 centers. 1431 Going forward, one network will support operation and maintenance of 1432 electrical networks (generation, transmission, and distribution), 1433 voice and data services for ten of thousands of employees and for 1434 exchange with neighboring interconnections, and administrative 1435 services. To meet those requirements, utility may deploy several 1436 physical networks leveraging different technologies across the 1437 country: an optical network and a microwave network for instance. 1438 Each protection and automatism system between two points has two 1439 telecommunications circuits, one on each network. Path diversity 1440 between two substations is key. Regardless of the event type 1441 (hurricane, ice storm, etc.), one path shall stay available so the 1442 system can still operate. 1444 In the optical network, signals are transmitted over more than tens 1445 of thousands of circuits using fiber optic links, microwave and 1446 telephone cables. This network is the nervous system of the 1447 utility's power transmission operations. The optical network 1448 represents ten of thousands of km of cable deployed along the power 1449 lines, with individual runs as long as 280 km. 1451 3.3.2.3. Precision Time Protocol 1453 Some utilities do not use GPS clocks in generation substations. One 1454 of the main reasons is that some of the generation plants are 30 to 1455 50 meters deep under ground and the GPS signal can be weak and 1456 unreliable. Instead, atomic clocks are used. Clocks are 1457 synchronized amongst each other. Rubidium clocks provide clock and 1458 1ms timestamps for IRIG-B. 1460 Some companies plan to transition to the Precision Time Protocol 1461 (PTP, [IEEE1588]), distributing the synchronization signal over the 1462 IP/MPLS network. PTP provides a mechanism for synchronizing the 1463 clocks of participating nodes to a high degree of accuracy and 1464 precision. 1466 PTP operates based on the following assumptions: 1468 It is assumed that the network eliminates cyclic forwarding of PTP 1469 messages within each communication path (e.g. by using a spanning 1470 tree protocol). 1472 PTP is tolerant of an occasional missed message, duplicated 1473 message, or message that arrived out of order. However, PTP 1474 assumes that such impairments are relatively rare. 1476 PTP was designed assuming a multicast communication model, however 1477 PTP also supports a unicast communication model as long as the 1478 behavior of the protocol is preserved. 1480 Like all message-based time transfer protocols, PTP time accuracy 1481 is degraded by delay asymmetry in the paths taken by event 1482 messages. Asymmetry is not detectable by PTP, however, if such 1483 delays are known a priori, PTP can correct for asymmetry. 1485 IEC 61850 defines the use of IEC/IEEE 61850-9-3:2016. The title is: 1486 Precision time protocol profile for power utility automation. It is 1487 based on Annex B/IEC 62439 which offers the support of redundant 1488 attachment of clocks to Parallel Redundancy Protocol (PRP) and High- 1489 availability Seamless Redundancy (HSR) networks. 1491 3.3.3. Security Trends in Utility Networks 1493 Although advanced telecommunications networks can assist in 1494 transforming the energy industry by playing a critical role in 1495 maintaining high levels of reliability, performance, and 1496 manageability, they also introduce the need for an integrated 1497 security infrastructure. Many of the technologies being deployed to 1498 support smart grid projects such as smart meters and sensors can 1499 increase the vulnerability of the grid to attack. Top security 1500 concerns for utilities migrating to an intelligent smart grid 1501 telecommunications platform center on the following trends: 1503 o Integration of distributed energy resources 1505 o Proliferation of digital devices to enable management, automation, 1506 protection, and control 1508 o Regulatory mandates to comply with standards for critical 1509 infrastructure protection 1511 o Migration to new systems for outage management, distribution 1512 automation, condition-based maintenance, load forecasting, and 1513 smart metering 1515 o Demand for new levels of customer service and energy management 1517 This development of a diverse set of networks to support the 1518 integration of microgrids, open-access energy competition, and the 1519 use of network-controlled devices is driving the need for a converged 1520 security infrastructure for all participants in the smart grid, 1521 including utilities, energy service providers, large commercial and 1522 industrial, as well as residential customers. Securing the assets of 1523 electric power delivery systems (from the control center to the 1524 substation, to the feeders and down to customer meters) requires an 1525 end-to-end security infrastructure that protects the myriad of 1526 telecommunications assets used to operate, monitor, and control power 1527 flow and measurement. 1529 "Cyber security" refers to all the security issues in automation and 1530 telecommunications that affect any functions related to the operation 1531 of the electric power systems. Specifically, it involves the 1532 concepts of: 1534 o Integrity : data cannot be altered undetectably 1536 o Authenticity : the telecommunications parties involved must be 1537 validated as genuine 1539 o Authorization : only requests and commands from the authorized 1540 users can be accepted by the system 1542 o Confidentiality : data must not be accessible to any 1543 unauthenticated users 1545 When designing and deploying new smart grid devices and 1546 telecommunications systems, it is imperative to understand the 1547 various impacts of these new components under a variety of attack 1548 situations on the power grid. Consequences of a cyber attack on the 1549 grid telecommunications network can be catastrophic. This is why 1550 security for smart grid is not just an ad hoc feature or product, 1551 it's a complete framework integrating both physical and Cyber 1552 security requirements and covering the entire smart grid networks 1553 from generation to distribution. Security has therefore become one 1554 of the main foundations of the utility telecom network architecture 1555 and must be considered at every layer with a defense-in-depth 1556 approach. Migrating to IP based protocols is key to address these 1557 challenges for two reasons: 1559 o IP enables a rich set of features and capabilities to enhance the 1560 security posture 1562 o IP is based on open standards, which allows interoperability 1563 between different vendors and products, driving down the costs 1564 associated with implementing security solutions in OT networks. 1566 Securing OT (Operation technology) telecommunications over packet- 1567 switched IP networks follow the same principles that are foundational 1568 for securing the IT infrastructure, i.e., consideration must be given 1569 to enforcing electronic access control for both person-to-machine and 1570 machine-to-machine communications, and providing the appropriate 1571 levels of data privacy, device and platform integrity, and threat 1572 detection and mitigation. 1574 3.4. Electrical Utilities Asks 1576 o Mixed L2 and L3 topologies 1578 o Deterministic behavior 1580 o Bounded latency and jitter 1582 o Tight feedback intervals 1584 o High availability, low recovery time 1586 o Redundancy, low packet loss 1588 o Precise timing 1590 o Centralized computing of deterministic paths 1592 o Distributed configuration may also be useful 1594 4. Building Automation Systems 1596 4.1. Use Case Description 1598 A Building Automation System (BAS) manages equipment and sensors in a 1599 building for improving residents' comfort, reducing energy 1600 consumption, and responding to failures and emergencies. For 1601 example, the BAS measures the temperature of a room using sensors and 1602 then controls the HVAC (heating, ventilating, and air conditioning) 1603 to maintain a set temperature and minimize energy consumption. 1605 A BAS primarily performs the following functions: 1607 o Periodically measures states of devices, for example humidity and 1608 illuminance of rooms, open/close state of doors, FAN speed, etc. 1610 o Stores the measured data. 1612 o Provides the measured data to BAS systems and operators. 1614 o Generates alarms for abnormal state of devices. 1616 o Controls devices (e.g. turn off room lights at 10:00 PM). 1618 4.2. Building Automation Systems Today 1620 4.2.1. BAS Architecture 1622 A typical BAS architecture of today is shown in Figure 4. 1624 +----------------------------+ 1625 | | 1626 | BMS HMI | 1627 | | | | 1628 | +----------------------+ | 1629 | | Management Network | | 1630 | +----------------------+ | 1631 | | | | 1632 | LC LC | 1633 | | | | 1634 | +----------------------+ | 1635 | | Field Network | | 1636 | +----------------------+ | 1637 | | | | | | 1638 | Dev Dev Dev Dev | 1639 | | 1640 +----------------------------+ 1642 BMS := Building Management Server 1643 HMI := Human Machine Interface 1644 LC := Local Controller 1646 Figure 4: BAS architecture 1648 There are typically two layers of network in a BAS. The upper one is 1649 called the Management Network and the lower one is called the Field 1650 Network. In management networks an IP-based communication protocol 1651 is used, while in field networks non-IP based communication protocols 1652 ("field protocols") are mainly used. Field networks have specific 1653 timing requirements, whereas management networks can be best-effort. 1655 A Human Machine Interface (HMI) is typically a desktop PC used by 1656 operators to monitor and display device states, send device control 1657 commands to Local Controllers (LCs), and configure building schedules 1658 (for example "turn off all room lights in the building at 10:00 PM"). 1660 A Building Management Server (BMS) performs the following operations. 1662 o Collect and store device states from LCs at regular intervals. 1664 o Send control values to LCs according to a building schedule. 1666 o Send an alarm signal to operators if it detects abnormal devices 1667 states. 1669 The BMS and HMI communicate with LCs via IP-based "management 1670 protocols" (see standards [bacnetip], [knx]). 1672 A LC is typically a Programmable Logic Controller (PLC) which is 1673 connected to several tens or hundreds of devices using "field 1674 protocols". An LC performs the following kinds of operations: 1676 o Measure device states and provide the information to BMS or HMI. 1678 o Send control values to devices, unilaterally or as part of a 1679 feedback control loop. 1681 There are many field protocols used today; some are standards-based 1682 and others are proprietary (see standards [lontalk], [modbus], 1683 [profibus] and [flnet]). The result is that BASs have multiple MAC/ 1684 PHY modules and interfaces. This makes BASs more expensive, slower 1685 to develop, and can result in "vendor lock-in" with multiple types of 1686 management applications. 1688 4.2.2. BAS Deployment Model 1690 An example BAS for medium or large buildings is shown in Figure 5. 1691 The physical layout spans multiple floors, and there is a monitoring 1692 room where the BAS management entities are located. Each floor will 1693 have one or more LCs depending upon the number of devices connected 1694 to the field network. 1696 +--------------------------------------------------+ 1697 | Floor 3 | 1698 | +----LC~~~~+~~~~~+~~~~~+ | 1699 | | | | | | 1700 | | Dev Dev Dev | 1701 | | | 1702 |--- | ------------------------------------------| 1703 | | Floor 2 | 1704 | +----LC~~~~+~~~~~+~~~~~+ Field Network | 1705 | | | | | | 1706 | | Dev Dev Dev | 1707 | | | 1708 |--- | ------------------------------------------| 1709 | | Floor 1 | 1710 | +----LC~~~~+~~~~~+~~~~~+ +-----------------| 1711 | | | | | | Monitoring Room | 1712 | | Dev Dev Dev | | 1713 | | | BMS HMI | 1714 | | Management Network | | | | 1715 | +--------------------------------+-----+ | 1716 | | | 1717 +--------------------------------------------------+ 1719 Figure 5: BAS Deployment model for Medium/Large Buildings 1721 Each LC is connected to the monitoring room via the Management 1722 network, and the management functions are performed within the 1723 building. In most cases, fast Ethernet (e.g. 100BASE-T) is used for 1724 the management network. Since the management network is non- 1725 realtime, use of Ethernet without quality of service is sufficient 1726 for today's deployment. 1728 In the field network a variety of physical interfaces such as RS232C 1729 and RS485 are used, which have specific timing requirements. Thus if 1730 a field network is to be replaced with an Ethernet or wireless 1731 network, such networks must support time-critical deterministic 1732 flows. 1734 In Figure 6, another deployment model is presented in which the 1735 management system is hosted remotely. This is becoming popular for 1736 small office and residential buildings in which a standalone 1737 monitoring system is not cost-effective. 1739 +---------------+ 1740 | Remote Center | 1741 | | 1742 | BMS HMI | 1743 +------------------------------------+ | | | | 1744 | Floor 2 | | +---+---+ | 1745 | +----LC~~~~+~~~~~+ Field Network| | | | 1746 | | | | | | Router | 1747 | | Dev Dev | +-------|-------+ 1748 | | | | 1749 |--- | ------------------------------| | 1750 | | Floor 1 | | 1751 | +----LC~~~~+~~~~~+ | | 1752 | | | | | | 1753 | | Dev Dev | | 1754 | | | | 1755 | | Management Network | WAN | 1756 | +------------------------Router-------------+ 1757 | | 1758 +------------------------------------+ 1760 Figure 6: Deployment model for Small Buildings 1762 Some interoperability is possible today in the Management Network, 1763 but not in today's field networks due to their non-IP-based design. 1765 4.2.3. Use Cases for Field Networks 1767 Below are use cases for Environmental Monitoring, Fire Detection, and 1768 Feedback Control, and their implications for field network 1769 performance. 1771 4.2.3.1. Environmental Monitoring 1773 The BMS polls each LC at a maximum measurement interval of 100ms (for 1774 example to draw a historical chart of 1 second granularity with a 10x 1775 sampling interval) and then performs the operations as specified by 1776 the operator. Each LC needs to measure each of its several hundred 1777 sensors once per measurement interval. Latency is not critical in 1778 this scenario as long as all sensor values are completed in the 1779 measurement interval. Availability is expected to be 99.999 %. 1781 4.2.3.2. Fire Detection 1783 On detection of a fire, the BMS must stop the HVAC, close the fire 1784 shutters, turn on the fire sprinklers, send an alarm, etc. There are 1785 typically ~10s of sensors per LC that BMS needs to manage. In this 1786 scenario the measurement interval is 10-50ms, the communication delay 1787 is 10ms, and the availability must be 99.9999 %. 1789 4.2.3.3. Feedback Control 1791 BAS systems utilize feedback control in various ways; the most time- 1792 critial is control of DC motors, which require a short feedback 1793 interval (1-5ms) with low communication delay (10ms) and jitter 1794 (1ms). The feedback interval depends on the characteristics of the 1795 device and a target quality of control value. There are typically 1796 ~10s of such devices per LC. 1798 Communication delay is expected to be less than 10ms, jitter less 1799 than 1ms while the availability must be 99.9999% . 1801 4.2.4. Security Considerations 1803 When BAS field networks were developed it was assumed that the field 1804 networks would always be physically isolated from external networks 1805 and therefore security was not a concern. In today's world many BASs 1806 are managed remotely and are thus connected to shared IP networks and 1807 so security is definitely a concern, yet security features are not 1808 available in the majority of BAS field network deployments . 1810 The management network, being an IP-based network, has the protocols 1811 available to enable network security, but in practice many BAS 1812 systems do not implement even the available security features such as 1813 device authentication or encryption for data in transit. 1815 4.3. BAS Future 1817 In the future we expect more fine-grained environmental monitoring 1818 and lower energy consumption, which will require more sensors and 1819 devices, thus requiring larger and more complex building networks. 1821 We expect building networks to be connected to or converged with 1822 other networks (Enterprise network, Home network, and Internet). 1824 Therefore better facilities for network management, control, 1825 reliability and security are critical in order to improve resident 1826 and operator convenience and comfort. For example the ability to 1827 monitor and control building devices via the internet would enable 1828 (for example) control of room lights or HVAC from a resident's 1829 desktop PC or phone application. 1831 4.4. BAS Asks 1833 The community would like to see an interoperable protocol 1834 specification that can satisfy the timing, security, availability and 1835 QoS constraints described above, such that the resulting converged 1836 network can replace the disparate field networks. Ideally this 1837 connectivity could extend to the open Internet. 1839 This would imply an architecture that can guarantee 1841 o Low communication delays (from <10ms to 100ms in a network of 1842 several hundred devices) 1844 o Low jitter (< 1 ms) 1846 o Tight feedback intervals (1ms - 10ms) 1848 o High network availability (up to 99.9999% ) 1850 o Availability of network data in disaster scenario 1852 o Authentication between management and field devices (both local 1853 and remote) 1855 o Integrity and data origin authentication of communication data 1856 between field and management devices 1858 o Confidentiality of data when communicated to a remote device 1860 5. Wireless for Industrial 1862 5.1. Use Case Description 1864 Wireless networks are useful for industrial applications, for example 1865 when portable, fast-moving or rotating objects are involved, and for 1866 the resource-constrained devices found in the Internet of Things 1867 (IoT). 1869 Such network-connected sensors, actuators, control loops (etc.) 1870 typically require that the underlying network support real-time 1871 quality of service (QoS), as well as specific classes of other 1872 network properties such as reliability, redundancy, and security. 1874 These networks may also contain very large numbers of devices, for 1875 example for factories, "big data" acquisition, and the IoT. Given 1876 the large numbers of devices installed, and the potential 1877 pervasiveness of the IoT, this is a huge and very cost-sensitive 1878 market. For example, a 1% cost reduction in some areas could save 1879 $100B 1881 5.1.1. Network Convergence using 6TiSCH 1883 Some wireless network technologies support real-time QoS, and are 1884 thus useful for these kinds of networks, but others do not. For 1885 example WiFi is pervasive but does not provide guaranteed timing or 1886 delivery of packets, and thus is not useful in this context. 1888 In this use case we focus on one specific wireless network technology 1889 which does provide the required deterministic QoS, which is "IPv6 1890 over the TSCH mode of IEEE 802.15.4e" (6TiSCH, where TSCH stands for 1891 "Time-Slotted Channel Hopping", see [I-D.ietf-6tisch-architecture], 1892 [IEEE802154], [IEEE802154e], and [RFC7554]). 1894 There are other deterministic wireless busses and networks available 1895 today, however they are imcompatible with each other, and 1896 incompatible with IP traffic (for example [ISA100], [WirelessHART]). 1898 Thus the primary goal of this use case is to apply 6TiSCH as a 1899 converged IP- and standards-based wireless network for industrial 1900 applications, i.e. to replace multiple proprietary and/or 1901 incompatible wireless networking and wireless network management 1902 standards. 1904 5.1.2. Common Protocol Development for 6TiSCH 1906 Today there are a number of protocols required by 6TiSCH which are 1907 still in development, and a second intent of this use case is to 1908 highlight the ways in which these "missing" protocols share goals in 1909 common with DetNet. Thus it is possible that some of the protocol 1910 technology developed for DetNet will also be applicable to 6TiSCH. 1912 These protocol goals are identified here, along with their 1913 relationship to DetNet. It is likely that ultimately the resulting 1914 protocols will not be identical, but will share design principles 1915 which contribute to the eficiency of enabling both DetNet and 6TiSCH. 1917 One such commonality is that although at a different time scale, in 1918 both TSN [IEEE802.1TSNTG] and TSCH a packet crosses the network from 1919 node to node follows a precise schedule, as a train that leaves 1920 intermediate stations at precise times along its path. This kind of 1921 operation reduces collisions, saves energy, and enables engineering 1922 the network for deterministic properties. 1924 Another commonality is remote monitoring and scheduling management of 1925 a TSCH network by a Path Computation Element (PCE) and Network 1926 Management Entity (NME). The PCE/NME manage timeslots and device 1927 resources in a manner that minimizes the interaction with and the 1928 load placed on resource-constrained devices. For example, a tiny IoT 1929 device may have just enough buffers to store one or a few IPv6 1930 packets, and will have limited bandwidth between peers such that it 1931 can maintain only a small amount of peer information, and will not be 1932 able to store many packets waiting to be forwarded. It is 1933 advantageous then for it to only be required to carry out the 1934 specific behavior assigned to it by the PCE/NME (as opposed to 1935 maintaining its own IP stack, for example). 1937 Note: Current WG discussion indicates that some peer-to-peer 1938 communication must be assumed, i.e. the PCE may communicate only 1939 indirectly with any given device, enabling hierarchical configuration 1940 of the system. 1942 6TiSCH depends on [PCE] and [I-D.ietf-detnet-architecture]. 1944 6TiSCH also depends on the fact that DetNet will maintain consistency 1945 with [IEEE802.1TSNTG]. 1947 5.2. Wireless Industrial Today 1949 Today industrial wireless is accomplished using multiple 1950 deterministic wireless networks which are incompatible with each 1951 other and with IP traffic. 1953 6TiSCH is not yet fully specified, so it cannot be used in today's 1954 applications. 1956 5.3. Wireless Industrial Future 1958 5.3.1. Unified Wireless Network and Management 1960 We expect DetNet and 6TiSCH together to enable converged transport of 1961 deterministic and best-effort traffic flows between real-time 1962 industrial devices and wide area networks via IP routing. A high 1963 level view of a basic such network is shown in Figure 7. 1965 ---+-------- ............ ------------ 1966 | External Network | 1967 | +-----+ 1968 +-----+ | NME | 1969 | | LLN Border | | 1970 | | router +-----+ 1971 +-----+ 1972 o o o 1973 o o o o 1974 o o LLN o o o 1975 o o o o 1976 o 1978 Figure 7: Basic 6TiSCH Network 1980 Figure 8 shows a backbone router federating multiple synchronized 1981 6TiSCH subnets into a single subnet connected to the external 1982 network. 1984 ---+-------- ............ ------------ 1985 | External Network | 1986 | +-----+ 1987 | +-----+ | NME | 1988 +-----+ | +-----+ | | 1989 | | Router | | PCE | +-----+ 1990 | | +--| | 1991 +-----+ +-----+ 1992 | | 1993 | Subnet Backbone | 1994 +--------------------+------------------+ 1995 | | | 1996 +-----+ +-----+ +-----+ 1997 | | Backbone | | Backbone | | Backbone 1998 o | | router | | router | | router 1999 +-----+ +-----+ +-----+ 2000 o o o o o 2001 o o o o o o o o o o o 2002 o o o LLN o o o o 2003 o o o o o o o o o o o o 2005 Figure 8: Extended 6TiSCH Network 2007 The backbone router must ensure end-to-end deterministic behavior 2008 between the LLN and the backbone. We would like to see this 2009 accomplished in conformance with the work done in 2010 [I-D.ietf-detnet-architecture] with respect to Layer-3 aspects of 2011 deterministic networks that span multiple Layer-2 domains. 2013 The PCE must compute a deterministic path end-to-end across the TSCH 2014 network and IEEE802.1 TSN Ethernet backbone, and DetNet protocols are 2015 expected to enable end-to-end deterministic forwarding. 2017 +-----+ 2018 | IoT | 2019 | G/W | 2020 +-----+ 2021 ^ <---- Elimination 2022 | | 2023 Track branch | | 2024 +-------+ +--------+ Subnet Backbone 2025 | | 2026 +--|--+ +--|--+ 2027 | | | Backbone | | | Backbone 2028 o | | | router | | | router 2029 +--/--+ +--|--+ 2030 o / o o---o----/ o 2031 o o---o--/ o o o o o 2032 o \ / o o LLN o 2033 o v <---- Replication 2034 o 2036 Figure 9: 6TiSCH Network with PRE 2038 5.3.1.1. PCE and 6TiSCH ARQ Retries 2040 Note: The possible use of ARQ techniques in DetNet is currently 2041 considered a possible design alternative. 2043 6TiSCH uses the IEEE802.15.4 Automatic Repeat-reQuest (ARQ) mechanism 2044 to provide higher reliability of packet delivery. ARQ is related to 2045 packet replication and elimination because there are two independent 2046 paths for packets to arrive at the destination, and if an expected 2047 packed does not arrive on one path then it checks for the packet on 2048 the second path. 2050 Although to date this mechanism is only used by wireless networks, 2051 this may be a technique that would be appropriate for DetNet and so 2052 aspects of the enabling protocol could be co-developed. 2054 For example, in Figure 9, a Track is laid out from a field device in 2055 a 6TiSCH network to an IoT gateway that is located on a IEEE802.1 TSN 2056 backbone. 2058 In ARQ the Replication function in the field device sends a copy of 2059 each packet over two different branches, and the PCE schedules each 2060 hop of both branches so that the two copies arrive in due time at the 2061 gateway. In case of a loss on one branch, hopefully the other copy 2062 of the packet still arrives within the allocated time. If two copies 2063 make it to the IoT gateway, the Elimination function in the gateway 2064 ignores the extra packet and presents only one copy to upper layers. 2066 At each 6TiSCH hop along the Track, the PCE may schedule more than 2067 one timeSlot for a packet, so as to support Layer-2 retries (ARQ). 2069 In current deployments, a TSCH Track does not necessarily support PRE 2070 but is systematically multi-path. This means that a Track is 2071 scheduled so as to ensure that each hop has at least two forwarding 2072 solutions, and the forwarding decision is to try the preferred one 2073 and use the other in case of Layer-2 transmission failure as detected 2074 by ARQ. 2076 5.3.2. Schedule Management by a PCE 2078 A common feature of 6TiSCH and DetNet is the action of a PCE to 2079 configure paths through the network. Specifically, what is needed is 2080 a protocol and data model that the PCE will use to get/set the 2081 relevant configuration from/to the devices, as well as perform 2082 operations on the devices. We expect that this protocol will be 2083 developed by DetNet with consideration for its reuse by 6TiSCH. The 2084 remainder of this section provides a bit more context from the 6TiSCH 2085 side. 2087 5.3.2.1. PCE Commands and 6TiSCH CoAP Requests 2089 The 6TiSCH device does not expect to place the request for bandwidth 2090 between itself and another device in the network. Rather, an 2091 operation control system invoked through a human interface specifies 2092 the required traffic specification and the end nodes (in terms of 2093 latency and reliability). Based on this information, the PCE must 2094 compute a path between the end nodes and provision the network with 2095 per-flow state that describes the per-hop operation for a given 2096 packet, the corresponding timeslots, and the flow identification that 2097 enables recognizing that a certain packet belongs to a certain path, 2098 etc. 2100 For a static configuration that serves a certain purpose for a long 2101 period of time, it is expected that a node will be provisioned in one 2102 shot with a full schedule, which incorporates the aggregation of its 2103 behavior for multiple paths. 6TiSCH expects that the programing of 2104 the schedule will be done over COAP as discussed in 2105 [I-D.ietf-6tisch-coap]. 2107 6TiSCH expects that the PCE commands will be mapped back and forth 2108 into CoAP by a gateway function at the edge of the 6TiSCH network. 2109 For instance, it is possible that a mapping entity on the backbone 2110 transforms a non-CoAP protocol such as PCEP into the RESTful 2111 interfaces that the 6TiSCH devices support. This architecture will 2112 be refined to comply with DetNet [I-D.ietf-detnet-architecture] when 2113 the work is formalized. Related information about 6TiSCH can be 2114 found at [I-D.ietf-6tisch-6top-interface] and RPL [RFC6550]. 2116 A protocol may be used to update the state in the devices during 2117 runtime, for example if it appears that a path through the network 2118 has ceased to perform as expected, but in 6TiSCH that flow was not 2119 designed and no protocol was selected. We would like to see DetNet 2120 define the appropriate end-to-end protocols to be used in that case. 2121 The implication is that these state updates take place once the 2122 system is configured and running, i.e. they are not limited to the 2123 initial communication of the configuration of the system. 2125 A "slotFrame" is the base object that a PCE would manipulate to 2126 program a schedule into an LLN node ([I-D.ietf-6tisch-architecture]). 2128 We would like to see the PCE read energy data from devices, and 2129 compute paths that will implement policies on how energy in devices 2130 is consumed, for instance to ensure that the spent energy does not 2131 exceeded the available energy over a period of time. Note: this 2132 statement implies that an extensible protocol for communicating 2133 device info to the PCE and enabling the PCE to act on it will be part 2134 of the DetNet architecture, however for subnets with specific 2135 protocols (e.g. CoAP) a gateway may be required. 2137 6TiSCH devices can discover their neighbors over the radio using a 2138 mechanism such as beacons, but even though the neighbor information 2139 is available in the 6TiSCH interface data model, 6TiSCH does not 2140 describe a protocol to proactively push the neighborhood information 2141 to a PCE. We would like to see DetNet define such a protocol; one 2142 possible design alternative is that it could operate over CoAP, 2143 alternatively it could be converted to/from CoAP by a gateway. We 2144 would like to see such a protocol carry multiple metrics, for example 2145 similar to those used for RPL operations [RFC6551] 2147 5.3.2.2. 6TiSCH IP Interface 2149 "6top" ([I-D.wang-6tisch-6top-sublayer]) is a logical link control 2150 sitting between the IP layer and the TSCH MAC layer which provides 2151 the link abstraction that is required for IP operations. The 6top 2152 data model and management interfaces are further discussed in 2153 [I-D.ietf-6tisch-6top-interface] and [I-D.ietf-6tisch-coap]. 2155 An IP packet that is sent along a 6TiSCH path uses the Differentiated 2156 Services Per-Hop-Behavior Group called Deterministic Forwarding, as 2157 described in [I-D.svshah-tsvwg-deterministic-forwarding]. 2159 5.3.3. 6TiSCH Security Considerations 2161 On top of the classical requirements for protection of control 2162 signaling, it must be noted that 6TiSCH networks operate on limited 2163 resources that can be depleted rapidly in a DoS attack on the system, 2164 for instance by placing a rogue device in the network, or by 2165 obtaining management control and setting up unexpected additional 2166 paths. 2168 5.4. Wireless Industrial Asks 2170 6TiSCH depends on DetNet to define: 2172 o Configuration (state) and operations for deterministic paths 2174 o End-to-end protocols for deterministic forwarding (tagging, IP) 2176 o Protocol for packet replication and elimination 2178 6. Cellular Radio 2180 6.1. Use Case Description 2182 This use case describes the application of deterministic networking 2183 in the context of cellular telecom transport networks. Important 2184 elements include time synchronization, clock distribution, and ways 2185 of establishing time-sensitive streams for both Layer-2 and Layer-3 2186 user plane traffic. 2188 6.1.1. Network Architecture 2190 Figure 10 illustrates a typical 3GPP-defined cellular network 2191 architecture, which includes "Fronthaul", "Midhaul" and "Backhaul" 2192 network segments. The "Fronthaul" is the network connecting base 2193 stations (baseband processing units) to the remote radio heads 2194 (antennas). The "Midhaul" is the network inter-connecting base 2195 stations (or small cell sites). The "Backhaul" is the network or 2196 links connecting the radio base station sites to the network 2197 controller/gateway sites (i.e. the core of the 3GPP cellular 2198 network). 2200 In Figure 10 "eNB" ("E-UTRAN Node B") is the hardware that is 2201 connected to the mobile phone network which communicates directly 2202 with mobile handsets ([TS36300]). 2204 Y (remote radio heads (antennas)) 2205 \ 2206 Y__ \.--. .--. +------+ 2207 \_( `. +---+ _(Back`. | 3GPP | 2208 Y------( Front )----|eNB|----( Haul )----| core | 2209 ( ` .Haul ) +---+ ( ` . ) ) | netw | 2210 /`--(___.-' \ `--(___.-' +------+ 2211 Y_/ / \.--. \ 2212 Y_/ _( Mid`. \ 2213 ( Haul ) \ 2214 ( ` . ) ) \ 2215 `--(___.-'\_____+---+ (small cell sites) 2216 \ |SCe|__Y 2217 +---+ +---+ 2218 Y__|eNB|__Y 2219 +---+ 2220 Y_/ \_Y ("local" radios) 2222 Figure 10: Generic 3GPP-based Cellular Network Architecture 2224 6.1.2. Delay Constraints 2226 The available processing time for Fronthaul networking overhead is 2227 limited to the available time after the baseband processing of the 2228 radio frame has completed. For example in Long Term Evolution (LTE) 2229 radio, processing of a radio frame is allocated 3ms but typically the 2230 processing uses most of it, allowing only a small fraction to be used 2231 by the Fronthaul network (e.g. up to 250us one-way delay, though the 2232 existing spec ([NGMN-fronth]) supports delay only up to 100us). This 2233 ultimately determines the distance the remote radio heads can be 2234 located from the base stations (e.g., 100us equals roughly 20 km of 2235 optical fiber-based transport). Allocation options of the available 2236 time budget between processing and transport are under heavy 2237 discussions in the mobile industry. 2239 For packet-based transport the allocated transport time (e.g. CPRI 2240 would allow for 100us delay [CPRI]) is consumed by all nodes and 2241 buffering between the remote radio head and the baseband processing 2242 unit, plus the distance-incurred delay. 2244 The baseband processing time and the available "delay budget" for the 2245 fronthaul is likely to change in the forthcoming "5G" due to reduced 2246 radio round trip times and other architectural and service 2247 requirements [NGMN]. 2249 The transport time budget, as noted above, places limitations on the 2250 distance that remote radio heads can be located from base stations 2251 (i.e. the link length). In the above analysis, the entire transport 2252 time budget is assumed to be available for link propagation delay. 2253 However the transport time budget can be broken down into three 2254 components: scheduling /queueing delay, transmission delay, and link 2255 propagation delay. Using today's Fronthaul networking technology, 2256 the queuing, scheduling and transmission components might become the 2257 dominant factors in the total transport time rather than the link 2258 propagation delay. This is especially true in cases where the 2259 Fronthaul link is relatively short and it is shared among multiple 2260 Fronthaul flows, for example in indoor and small cell networks, 2261 massive MIMO antenna networks, and split Fronthaul architectures. 2263 DetNet technology can improve this application by controlling and 2264 reducing the time required for the queuing, scheduling and 2265 transmission operations by properly assigning the network resources, 2266 thus leaving more of the transport time budget available for link 2267 propagation, and thus enabling longer link lengths. However, link 2268 length is usually a given parameter and is not a controllable network 2269 parameter, since RRH and BBU sights are usually located in 2270 predetermined locations. However, the number of antennas in an RRH 2271 sight might increase for example by adding more antennas, increasing 2272 the MIMO capability of the network or support of massive MIMO. This 2273 means increasing the number of the fronthaul flows sharing the same 2274 fronthaul link. DetNet can now control the bandwidth assignment of 2275 the fronthaul link and the scheduling of fronthaul packets over this 2276 link and provide adequate buffer provisioning for each flow to reduce 2277 the packet loss rate. 2279 Another way in which DetNet technology can aid Fronthaul networks is 2280 by providing effective isolation from best-effort (and other classes 2281 of) traffic, which can arise as a result of network slicing in 5G 2282 networks where Fronthaul traffic generated in different network 2283 slices might have differing performance requirements. DetNet 2284 technology can also dynamically control the bandwidth assignment, 2285 scheduling and packet forwarding decisions and the buffer 2286 provisioning of the Fronthaul flows to guarantee the end-to-end delay 2287 of the Fronthaul packets and minimize the packet loss rate. 2289 [METIS] documents the fundamental challenges as well as overall 2290 technical goals of the future 5G mobile and wireless system as the 2291 starting point. These future systems should support much higher data 2292 volumes and rates and significantly lower end-to-end latency for 100x 2293 more connected devices (at similar cost and energy consumption levels 2294 as today's system). 2296 For Midhaul connections, delay constraints are driven by Inter-Site 2297 radio functions like Coordinated Multipoint Processing (CoMP, see 2298 [CoMP]). CoMP reception and transmission is a framework in which 2299 multiple geographically distributed antenna nodes cooperate to 2300 improve the performance of the users served in the common cooperation 2301 area. The design principal of CoMP is to extend the current single- 2302 cell to multi-UE (User Equipment) transmission to a multi-cell-to- 2303 multi-UEs transmission by base station cooperation. 2305 CoMP has delay-sensitive performance parameters, which are "midhaul 2306 latency" and "CSI (Channel State Information) reporting and 2307 accuracy". The essential feature of CoMP is signaling between eNBs, 2308 so Midhaul latency is the dominating limitation of CoMP performance. 2309 Generally, CoMP can benefit from coordinated scheduling (either 2310 distributed or centralized) of different cells if the signaling delay 2311 between eNBs is within 1-10ms. This delay requirement is both rigid 2312 and absolute because any uncertainty in delay will degrade the 2313 performance significantly. 2315 Inter-site CoMP is one of the key requirements for 5G and is also a 2316 near-term goal for the current 4.5G network architecture. 2318 6.1.3. Time Synchronization Constraints 2320 Fronthaul time synchronization requirements are given by [TS25104], 2321 [TS36104], [TS36211], and [TS36133]. These can be summarized for the 2322 current 3GPP LTE-based networks as: 2324 Delay Accuracy: 2325 +-8ns (i.e. +-1/32 Tc, where Tc is the UMTS Chip time of 1/3.84 2326 MHz) resulting in a round trip accuracy of +-16ns. The value is 2327 this low to meet the 3GPP Timing Alignment Error (TAE) measurement 2328 requirements. Note: performance guarantees of low nanosecond 2329 values such as these are considered to be below the DetNet layer - 2330 it is assumed that the underlying implementation, e.g. the 2331 hardware, will provide sufficient support (e.g. buffering) to 2332 enable this level of accuracy. These values are maintained in the 2333 use case to give an indication of the overall application. 2335 Timing Alignment Error: 2336 Timing Alignment Error (TAE) is problematic to Fronthaul networks 2337 and must be minimized. If the transport network cannot guarantee 2338 low enough TAE then additional buffering has to be introduced at 2339 the edges of the network to buffer out the jitter. Buffering is 2340 not desirable as it reduces the total available delay budget. 2341 Packet Delay Variation (PDV) requirements can be derived from TAE 2342 for packet based Fronthaul networks. 2344 * For multiple input multiple output (MIMO) or TX diversity 2345 transmissions, at each carrier frequency, TAE shall not exceed 2346 65 ns (i.e. 1/4 Tc). 2348 * For intra-band contiguous carrier aggregation, with or without 2349 MIMO or TX diversity, TAE shall not exceed 130 ns (i.e. 1/2 2350 Tc). 2352 * For intra-band non-contiguous carrier aggregation, with or 2353 without MIMO or TX diversity, TAE shall not exceed 260 ns (i.e. 2354 one Tc). 2356 * For inter-band carrier aggregation, with or without MIMO or TX 2357 diversity, TAE shall not exceed 260 ns. 2359 Transport link contribution to radio frequency error: 2360 +-2 PPB. This value is considered to be "available" for the 2361 Fronthaul link out of the total 50 PPB budget reserved for the 2362 radio interface. Note: the reason that the transport link 2363 contributes to radio frequency error is as follows. The current 2364 way of doing Fronthaul is from the radio unit to remote radio head 2365 directly. The remote radio head is essentially a passive device 2366 (without buffering etc.) The transport drives the antenna 2367 directly by feeding it with samples and everything the transport 2368 adds will be introduced to radio as-is. So if the transport 2369 causes additional frequency error that shows immediately on the 2370 radio as well. Note: performance guarantees of low nanosecond 2371 values such as these are considered to be below the DetNet layer - 2372 it is assumed that the underlying implementation, e.g. the 2373 hardware, will provide sufficient support to enable this level of 2374 performance. These values are maintained in the use case to give 2375 an indication of the overall application. 2377 The above listed time synchronization requirements are difficult to 2378 meet with point-to-point connected networks, and more difficult when 2379 the network includes multiple hops. It is expected that networks 2380 must include buffering at the ends of the connections as imposed by 2381 the jitter requirements, since trying to meet the jitter requirements 2382 in every intermediate node is likely to be too costly. However, 2383 every measure to reduce jitter and delay on the path makes it easier 2384 to meet the end-to-end requirements. 2386 In order to meet the timing requirements both senders and receivers 2387 must remain time synchronized, demanding very accurate clock 2388 distribution, for example support for IEEE 1588 transparent clocks or 2389 boundary clocks in every intermediate node. 2391 In cellular networks from the LTE radio era onward, phase 2392 synchronization is needed in addition to frequency synchronization 2393 ([TS36300], [TS23401]). Time constraints are also important due to 2394 their impact on packet loss. If a packet is delivered too late, then 2395 the packet may be dropped by the host. 2397 6.1.4. Transport Loss Constraints 2399 Fronthaul and Midhaul networks assume almost error-free transport. 2400 Errors can result in a reset of the radio interfaces, which can cause 2401 reduced throughput or broken radio connectivity for mobile customers. 2403 For packetized Fronthaul and Midhaul connections packet loss may be 2404 caused by BER, congestion, or network failure scenarios. Different 2405 fronthaul functional splits are being considered by 3GPP, requiring 2406 strict frame loss ratio (FLR) guarantees. As one example (referring 2407 to the legacy CPRI split which is option 8 in 3GPP) lower layers 2408 splits may imply an FLR of less than 10E-7 for data traffic and less 2409 than 10E-6 for control and management traffic. Current tools for 2410 eliminating packet loss for Fronthaul and Midhaul networks have 2411 serious challenges, for example retransmitting lost packets and/or 2412 using forward error correction (FEC) to circumvent bit errors is 2413 practically impossible due to the additional delay incurred. Using 2414 redundant streams for better guarantees for delivery is also 2415 practically impossible in many cases due to high bandwidth 2416 requirements of Fronthaul and Midhaul networks. Protection switching 2417 is also a candidate but current technologies for the path switch are 2418 too slow to avoid reset of mobile interfaces. 2420 Fronthaul links are assumed to be symmetric, and all Fronthaul 2421 streams (i.e. those carrying radio data) have equal priority and 2422 cannot delay or pre-empt each other. This implies that the network 2423 must guarantee that each time-sensitive flow meets their schedule. 2425 6.1.5. Security Considerations 2427 Establishing time-sensitive streams in the network entails reserving 2428 networking resources for long periods of time. It is important that 2429 these reservation requests be authenticated to prevent malicious 2430 reservation attempts from hostile nodes (or accidental 2431 misconfiguration). This is particularly important in the case where 2432 the reservation requests span administrative domains. Furthermore, 2433 the reservation information itself should be digitally signed to 2434 reduce the risk of a legitimate node pushing a stale or hostile 2435 configuration into another networking node. 2437 Note: This is considered important for the security policy of the 2438 network, but does not affect the core DetNet architecture and design. 2440 6.2. Cellular Radio Networks Today 2442 6.2.1. Fronthaul 2444 Today's Fronthaul networks typically consist of: 2446 o Dedicated point-to-point fiber connection is common 2448 o Proprietary protocols and framings 2450 o Custom equipment and no real networking 2452 Current solutions for Fronthaul are direct optical cables or 2453 Wavelength-Division Multiplexing (WDM) connections. 2455 6.2.2. Midhaul and Backhaul 2457 Today's Midhaul and Backhaul networks typically consist of: 2459 o Mostly normal IP networks, MPLS-TP, etc. 2461 o Clock distribution and sync using 1588 and SyncE 2463 Telecommunication networks in the Mid- and Backhaul are already 2464 heading towards transport networks where precise time synchronization 2465 support is one of the basic building blocks. While the transport 2466 networks themselves have practically transitioned to all-IP packet- 2467 based networks to meet the bandwidth and cost requirements, highly 2468 accurate clock distribution has become a challenge. 2470 In the past, Mid- and Backhaul connections were typically based on 2471 Time Division Multiplexing (TDM-based) and provided frequency 2472 synchronization capabilities as a part of the transport media. 2473 Alternatively other technologies such as Global Positioning System 2474 (GPS) or Synchronous Ethernet (SyncE) are used [SyncE]. 2476 Both Ethernet and IP/MPLS [RFC3031] (and PseudoWires (PWE) [RFC3985] 2477 for legacy transport support) have become popular tools to build and 2478 manage new all-IP Radio Access Networks (RANs) 2479 [I-D.kh-spring-ip-ran-use-case]. Although various timing and 2480 synchronization optimizations have already been proposed and 2481 implemented including 1588 PTP enhancements 2482 [I-D.ietf-tictoc-1588overmpls] and [RFC8169], these solution are not 2483 necessarily sufficient for the forthcoming RAN architectures nor do 2484 they guarantee the more stringent time-synchronization requirements 2485 such as [CPRI]. 2487 There are also existing solutions for TDM over IP such as [RFC4553], 2488 [RFC5086], and [RFC5087], as well as TDM over Ethernet transports 2489 such as [MEF8]. 2491 6.3. Cellular Radio Networks Future 2493 Future Cellular Radio Networks will be based on a mix of different 2494 xHaul networks (xHaul = front-, mid- and backhaul), and future 2495 transport networks should be able to support all of them 2496 simultaneously. It is already envisioned today that: 2498 o Not all "cellular radio network" traffic will be IP, for example 2499 some will remain at Layer 2 (e.g. Ethernet based). DetNet 2500 solutions must address all traffic types (Layer 2, Layer 3) with 2501 the same tools and allow their transport simultaneously. 2503 o All forms of xHaul networks will need some form of DetNet 2504 solutions. For example with the advent of 5G some Backhaul 2505 traffic will also have DetNet requirements, for example traffic 2506 belonging to time-critical 5G applications. 2508 o Different splits of the functionality run on the base stations and 2509 the on-site units could co-exist on the same Fronthaul and 2510 Backhaul network. 2512 We would like to see the following in future Cellular Radio networks: 2514 o Unified standards-based transport protocols and standard 2515 networking equipment that can make use of underlying deterministic 2516 link-layer services 2518 o Unified and standards-based network management systems and 2519 protocols in all parts of the network (including Fronthaul) 2521 New radio access network deployment models and architectures may 2522 require time- sensitive networking services with strict requirements 2523 on other parts of the network that previously were not considered to 2524 be packetized at all. Time and synchronization support are already 2525 topical for Backhaul and Midhaul packet networks [MEF22.1.1] and are 2526 becoming a real issue for Fronthaul networks also. Specifically in 2527 Fronthaul networks the timing and synchronization requirements can be 2528 extreme for packet based technologies, for example, on the order of 2529 sub +-20 ns packet delay variation (PDV) and frequency accuracy of 2530 +0.002 PPM [Fronthaul]. 2532 The actual transport protocols and/or solutions to establish required 2533 transport "circuits" (pinned-down paths) for Fronthaul traffic are 2534 still undefined. Those are likely to include (but are not limited 2535 to) solutions directly over Ethernet, over IP, and using MPLS/ 2536 PseudoWire transport. 2538 Even the current time-sensitive networking features may not be 2539 sufficient for Fronthaul traffic. Therefore, having specific 2540 profiles that take the requirements of Fronthaul into account is 2541 desirable [IEEE8021CM]. 2543 Interesting and important work for time-sensitive networking has been 2544 done for Ethernet [TSNTG], which specifies the use of IEEE 1588 time 2545 precision protocol (PTP) [IEEE1588] in the context of IEEE 802.1D and 2546 IEEE 802.1Q. [IEEE8021AS] specifies a Layer 2 time synchronizing 2547 service, and other specifications such as IEEE 1722 [IEEE1722] 2548 specify Ethernet-based Layer-2 transport for time-sensitive streams. 2550 New promising work seeks to enable the transport of time-sensitive 2551 fronthaul streams in Ethernet bridged networks [IEEE8021CM]. 2552 Analogous to IEEE 1722 there is an ongoing standardization effort to 2553 define the Layer-2 transport encapsulation format for transporting 2554 radio over Ethernet (RoE) in the IEEE 1904.3 Task Force [IEEE19143]. 2556 As mentioned in Section 6.1.2, 5G communications will provide one of 2557 the most challenging cases for delay sensitive networking. In order 2558 to meet the challenges of ultra-low latency and ultra-high 2559 throughput, 3GPP has studied various "functional splits" for 5G, 2560 i.e., physical decomposition of the gNodeB base station and 2561 deployment of its functional blocks in different locations [TR38801]. 2563 These splits are numbered from split option 1 (Dual Connectivity, a 2564 split in which the radio resource control is centralized and other 2565 radio stack layers are in distributed units) to split option 8 (a 2566 PHY-RF split in which RF functionality is in a distributed unit and 2567 the rest of the radio stack is in the centralized unit), with each 2568 intermediate split having its own data rate and delay requirements. 2569 Packetized versions of different splits have recently been proposed 2570 including eCPRI [eCPRI] and RoE (as previously noted). Both provide 2571 Ethernet encapsulations, and eCPRI is also capable of IP 2572 encapsulation. 2574 All-IP RANs and xHaul networks would benefit from time 2575 synchronization and time-sensitive transport services. Although 2576 Ethernet appears to be the unifying technology for the transport, 2577 there is still a disconnect providing Layer 3 services. The protocol 2578 stack typically has a number of layers below the Ethernet Layer 2 2579 that shows up to the Layer 3 IP transport. It is not uncommon that 2580 on top of the lowest layer (optical) transport there is the first 2581 layer of Ethernet followed one or more layers of MPLS, PseudoWires 2582 and/or other tunneling protocols finally carrying the Ethernet layer 2583 visible to the user plane IP traffic. 2585 While there are existing technologies to establish circuits through 2586 the routed and switched networks (especially in MPLS/PWE space), 2587 there is still no way to signal the time synchronization and time- 2588 sensitive stream requirements/reservations for Layer-3 flows in a way 2589 that addresses the entire transport stack, including the Ethernet 2590 layers that need to be configured. 2592 Furthermore, not all "user plane" traffic will be IP. Therefore, the 2593 same solution also must address the use cases where the user plane 2594 traffic is a different layer, for example Ethernet frames. 2596 There is existing work describing the problem statement 2597 [I-D.ietf-detnet-problem-statement] and the architecture 2598 [I-D.ietf-detnet-architecture] for deterministic networking (DetNet) 2599 that targets solutions for time-sensitive (IP/transport) streams with 2600 deterministic properties over Ethernet-based switched networks. 2602 6.4. Cellular Radio Networks Asks 2604 A standard for data plane transport specification which is: 2606 o Unified among all xHauls (meaning that different flows with 2607 diverse DetNet requirements can coexist in the same network and 2608 traverse the same nodes without interfering with each other) 2610 o Deployed in a highly deterministic network environment 2612 o Capable of supporting multiple functional splits simultaneously, 2613 including existing Backhaul and CPRI Fronthaul and potentially new 2614 modes as defined for example in 3GPP; these goals can be supported 2615 by the existing DetNet Use Case Common Themes, notably "Mix of 2616 Deterministic and Best-Effort Traffic", "Bounded Latency", "Low 2617 Latency", "Symmetrical Path Delays", and "Deterministic Flows". 2619 o Capable of supporting Network Slicing and Multi-tenancy; these 2620 goals can be supported by the same DetNet themes noted above. 2622 o Capable of transporting both in-band and out-band control traffic 2623 (OAM info, ...). 2625 o Deployable over multiple data link technologies (e.g., IEEE 802.3, 2626 mmWave, etc.). 2628 A standard for data flow information models that are: 2630 o Aware of the time sensitivity and constraints of the target 2631 networking environment 2633 o Aware of underlying deterministic networking services (e.g., on 2634 the Ethernet layer) 2636 7. Industrial M2M 2638 7.1. Use Case Description 2640 Industrial Automation in general refers to automation of 2641 manufacturing, quality control and material processing. In this 2642 "machine to machine" (M2M) use case we consider machine units in a 2643 plant floor which periodically exchange data with upstream or 2644 downstream machine modules and/or a supervisory controller within a 2645 local area network. 2647 The actors of M2M communication are Programmable Logic Controllers 2648 (PLCs). Communication between PLCs and between PLCs and the 2649 supervisory PLC (S-PLC) is achieved via critical control/data streams 2650 Figure 11. 2652 S (Sensor) 2653 \ +-----+ 2654 PLC__ \.--. .--. ---| MES | 2655 \_( `. _( `./ +-----+ 2656 A------( Local )-------------( L2 ) 2657 ( Net ) ( Net ) +-------+ 2658 /`--(___.-' `--(___.-' ----| S-PLC | 2659 S_/ / PLC .--. / +-------+ 2660 A_/ \_( `. 2661 (Actuator) ( Local ) 2662 ( Net ) 2663 /`--(___.-'\ 2664 / \ A 2665 S A 2667 Figure 11: Current Generic Industrial M2M Network Architecture 2669 This use case focuses on PLC-related communications; communication to 2670 Manufacturing-Execution-Systems (MESs) are not addressed. 2672 This use case covers only critical control/data streams; non-critical 2673 traffic between industrial automation applications (such as 2674 communication of state, configuration, set-up, and database 2675 communication) are adequately served by currently available 2676 prioritizing techniques. Such traffic can use up to 80% of the total 2677 bandwidth required. There is also a subset of non-time-critical 2678 traffic that must be reliable even though it is not time sensitive. 2680 In this use case the primary need for deterministic networking is to 2681 provide end-to-end delivery of M2M messages within specific timing 2682 constraints, for example in closed loop automation control. Today 2683 this level of determinism is provided by proprietary networking 2684 technologies. In addition, standard networking technologies are used 2685 to connect the local network to remote industrial automation sites, 2686 e.g. over an enterprise or metro network which also carries other 2687 types of traffic. Therefore, flows that should be forwarded with 2688 deterministic guarantees need to be sustained regardless of the 2689 amount of other flows in those networks. 2691 7.2. Industrial M2M Communication Today 2693 Today, proprietary networks fulfill the needed timing and 2694 availability for M2M networks. 2696 The network topologies used today by industrial automation are 2697 similar to those used by telecom networks: Daisy Chain, Ring, Hub and 2698 Spoke, and Comb (a subset of Daisy Chain). 2700 PLC-related control/data streams are transmitted periodically and 2701 carry either a pre-configured payload or a payload configured during 2702 runtime. 2704 Some industrial applications require time synchronization at the end 2705 nodes. For such time-coordinated PLCs, accuracy of 1 microsecond is 2706 required. Even in the case of "non-time-coordinated" PLCs time sync 2707 may be needed e.g. for timestamping of sensor data. 2709 Industrial network scenarios require advanced security solutions. 2710 Many of the current industrial production networks are physically 2711 separated. Preventing critical flows from be leaked outside a domain 2712 is handled today by filtering policies that are typically enforced in 2713 firewalls. 2715 7.2.1. Transport Parameters 2717 The Cycle Time defines the frequency of message(s) between industrial 2718 actors. The Cycle Time is application dependent, in the range of 1ms 2719 - 100ms for critical control/data streams. 2721 Because industrial applications assume deterministic transport for 2722 critical Control-Data-Stream parameters (instead of defining latency 2723 and delay variation parameters) it is sufficient to fulfill the upper 2724 bound of latency (maximum latency). The underlying networking 2725 infrastructure must ensure a maximum end-to-end delivery time of 2726 messages in the range of 100 microseconds to 50 milliseconds 2727 depending on the control loop application. 2729 The bandwidth requirements of control/data streams are usually 2730 calculated directly from the bytes-per-cycle parameter of the control 2731 loop. For PLC-to-PLC communication one can expect 2 - 32 streams 2732 with packet size in the range of 100 - 700 bytes. For S-PLC to PLCs 2733 the number of streams is higher - up to 256 streams. Usually no more 2734 than 20% of available bandwidth is used for critical control/data 2735 streams. In today's networks 1Gbps links are commonly used. 2737 Most PLC control loops are rather tolerant of packet loss, however 2738 critical control/data streams accept no more than 1 packet loss per 2739 consecutive communication cycle (i.e. if a packet gets lost in cycle 2740 "n", then the next cycle ("n+1") must be lossless). After two or 2741 more consecutive packet losses the network may be considered to be 2742 "down" by the Application. 2744 As network downtime may impact the whole production system the 2745 required network availability is rather high (99,999%). 2747 Based on the above parameters we expect that some form of redundancy 2748 will be required for M2M communications, however any individual 2749 solution depends on several parameters including cycle time, delivery 2750 time, etc. 2752 7.2.2. Stream Creation and Destruction 2754 In an industrial environment, critical control/data streams are 2755 created rather infrequently, on the order of ~10 times per day / week 2756 / month. Most of these critical control/data streams get created at 2757 machine startup, however flexibility is also needed during runtime, 2758 for example when adding or removing a machine. Going forward as 2759 production systems become more flexible, we expect a significant 2760 increase in the rate at which streams are created, changed and 2761 destroyed. 2763 7.3. Industrial M2M Future 2765 We would like to see a converged IP-standards-based network with 2766 deterministic properties that can satisfy the timing, security and 2767 reliability constraints described above. Today's proprietary 2768 networks could then be interfaced to such a network via gateways or, 2769 in the case of new installations, devices could be connected directly 2770 to the converged network. 2772 For this use case we expect time synchronization accuracy on the 2773 order of 1us. 2775 7.4. Industrial M2M Asks 2777 o Converged IP-based network 2779 o Deterministic behavior (bounded latency and jitter ) 2781 o High availability (presumably through redundancy) (99.999 %) 2783 o Low message delivery time (100us - 50ms) 2785 o Low packet loss (burstless, 0.1-1 %) 2787 o Security (e.g. prevent critical flows from being leaked between 2788 physically separated networks) 2790 8. Mining Industry 2792 8.1. Use Case Description 2794 The mining industry is highly dependent on networks to monitor and 2795 control their systems both in open-pit and underground extraction, 2796 transport and refining processes. In order to reduce risks and 2797 increase operational efficiency in mining operations, a number of 2798 processes have migrated the operators from the extraction site to 2799 remote control and monitoring. 2801 In the case of open pit mining, autonomous trucks are used to 2802 transport the raw materials from the open pit to the refining factory 2803 where the final product (e.g. Copper) is obtained. Although the 2804 operation is autonomous, the tracks are remotely monitored from a 2805 central facility. 2807 In pit mines, the monitoring of the tailings or mine dumps is 2808 critical in order to avoid any environmental pollution. In the past, 2809 monitoring has been conducted through manual inspection of pre- 2810 installed dataloggers. Cabling is not usually exploited in such 2811 scenarios due to the cost and complex deployment requirements. 2812 Currently, wireless technologies are being employed to monitor these 2813 cases permanently. Slopes are also monitored in order to anticipate 2814 possible mine collapse. Due to the unstable terrain, cable 2815 maintenance is costly and complex and hence wireless technologies are 2816 employed. 2818 In the underground monitoring case, autonomous vehicles with 2819 extraction tools travel autonomously through the tunnels, but their 2820 operational tasks (such as excavation, stone breaking and transport) 2821 are controlled remotely from a central facility. This generates 2822 video and feedback upstream traffic plus downstream actuator control 2823 traffic. 2825 8.2. Mining Industry Today 2827 Currently the mining industry uses a packet switched architecture 2828 supported by high speed ethernet. However in order to achieve the 2829 delay and packet loss requirements the network bandwidth is 2830 overestimated, thus providing very low efficiency in terms of 2831 resource usage. 2833 QoS is implemented at the Routers to separate video, management, 2834 monitoring and process control traffic for each stream. 2836 Since mobility is involved in this process, the connection between 2837 the backbone and the mobile devices (e.g. trucks, trains and 2838 excavators) is solved using a wireless link. These links are based 2839 on 802.11 for open-pit mining and leaky feeder for underground 2840 mining. 2842 Lately in pit mines the use of LPWAN technologies has been extended: 2843 Tailings, slopes and mine dumps are monitored by battery-powered 2844 dataloggers that make use of robust long range radio technologies. 2845 Reliability is usually ensured through retransmissions at L2. 2846 Gateways or concentrators act as bridges forwarding the data to the 2847 backbone ethernet network. Deterministic requirements are biased 2848 towards reliability rather than latency as events are slowly 2849 triggered or can be anticipated in advance. 2851 At the mineral processing stage, conveyor belts and refining 2852 processes are controlled by a SCADA system, which provides the in- 2853 factory delay-constrained networking requirements. 2855 Voice communications are currently served by a redundant trunking 2856 infrastructure, independent from current data networks. 2858 8.3. Mining Industry Future 2860 Mining operations and management are currently converging towards a 2861 combination of autonomous operation and teleoperation of transport 2862 and extraction machines. This means that video, audio, monitoring 2863 and process control traffic will increase dramatically. Ideally, all 2864 activities on the mine will rely on network infrastructure. 2866 Wireless for open-pit mining is already a reality with LPWAN 2867 technologies and it is expected to evolve to more advanced LPWAN 2868 technologies such as those based on LTE to increase last hop 2869 reliability or novel LPWAN flavours with deterministic access. 2871 One area in which DetNet can improve this use case is in the wired 2872 networks that make up the "backbone network" of the system, which 2873 connect together many wireless access points (APs). The mobile 2874 machines (which are connected to the network via wireless) transition 2875 from one AP to the next as they move about. A deterministic, 2876 reliable, low latency backbone can enable these transitions to be 2877 more reliable. 2879 Connections which extend all the way from the base stations to the 2880 machinery via a mix of wired and wireless hops would also be 2881 beneficial, for example to improve remote control responsiveness of 2882 digging machines. However to guarantee deterministic performance of 2883 a DetNet, the end-to-end underlying network must be deterministic. 2884 Thus for this use case if a deterministic wireless transport is 2885 integrated with a wire-based DetNet network, it could create the 2886 desired wired plus wireless end-to-end deterministic network. 2888 8.4. Mining Industry Asks 2890 o Improved bandwidth efficiency 2892 o Very low delay to enable machine teleoperation 2894 o Dedicated bandwidth usage for high resolution video streams 2896 o Predictable delay to enable realtime monitoring 2898 o Potential to construct a unified DetNet network over a combination 2899 of wired and deterministic wireless links 2901 9. Private Blockchain 2903 9.1. Use Case Description 2905 Blockchain was created with bitcoin, as a 'public' blockchain on the 2906 open Internet, however blockchain has also spread far beyond its 2907 original host into various industries such as smart manufacturing, 2908 logistics, security, legal rights and others. In these industries 2909 blockchain runs in designated and carefully managed network in which 2910 deterministic networking requirements could be addressed by Detnet. 2911 Such implementations are referred to as 'private' blockchain. 2913 The sole distinction between public and private blockchain is related 2914 to who is allowed to participate in the network, execute the 2915 consensus protocol and maintain the shared ledger. 2917 Today's networks treat the traffic from blockchain on a best-effort 2918 basis, but blockchain operation could be made much more efficient if 2919 deterministic networking service were available to minimize latency 2920 and packet loss in the network. 2922 9.1.1. Blockchain Operation 2924 A 'block' runs as a container of a batch of primary items such as 2925 transactions, property records etc. The blocks are chained in such a 2926 way that the hash of the previous block works as the pointer header 2927 of the new block, where confirmation of each block requires a 2928 consensus mechanism. When an item arrives at a blockchain node, the 2929 latter broadcasts this item to the rest of nodes which receive and 2930 verify it and put it in the ongoing block. Block confirmation 2931 process begins as the amount of items reaches the predefined block 2932 capacity, and the node broadcasts its proved block to the rest of 2933 nodes to be verified and chained. 2935 9.1.2. Blockchain Network Architecture 2937 Blockchain node communication and coordination is achieved mainly 2938 through frequent point to multi-point communication, however 2939 persistent point-to-point connections are used to transport both the 2940 items and the blocks to the other nodes. 2942 When a node initiates, it first requests the other nodes' address 2943 from a specific entity such as DNS, then it creates persistent 2944 connections each of with other nodes. If node A confirms an item, it 2945 sends the item to the other nodes via the persistent connections. 2947 As a new block in a node completes and gets proved among the nodes, 2948 it starts propagating this block towards its neighbor nodes. Assume 2949 node A receives a block, it sends invite message after verification 2950 to its neighbor B, B checks if the designated block is available, it 2951 responds get message to A if it is unavailable, and A send the 2952 complete block to B. B repeats the process as A to start the next 2953 round of block propagation. 2955 The challenge of blockchain network operation is not overall data 2956 rates, since the volume from both block and item stays between 2957 hundreds of bytes to a couple of mega bytes per second, but is in 2958 transporting the blocks with minimum latency to maximize efficiency 2959 of the blockchain consensus process. 2961 9.1.3. Security Considerations 2963 Security is crucial to blockchain applications, and todayt blockchain 2964 addresses its security issues mainly at the application level, where 2965 cryptography as well as hash-based consensus play a leading role 2966 preventing both double-spending and malicious service attack. 2967 However, there is concern that in the proposed use case of a private 2968 blockchain network which is dependent on deterministic properties, 2969 the network could be vulnerable to delays and other specific attacks 2970 against determinism which could interrupt service. 2972 9.2. Private Blockchain Today 2974 Today private blockchain runs in L2 or L3 VPN, in general without 2975 guaranteed determinism. The industry players are starting to realize 2976 that improving determinism in their blockchain networks could improve 2977 the performance of their service, but as of today these goals are not 2978 being met. 2980 9.3. Private Blockchain Future 2982 Blockchain system performance can be greatly improved through 2983 deterministic networking service primarily because it would 2984 accelerate the consensus process. It would be valuable to be able to 2985 design a private blockchain network with the following properties: 2987 o Transport of point to multi-point traffic in a coordinated network 2988 architecture rather than at the application layer (which typically 2989 uses point-to-point connections) 2991 o Guaranteed transport latency 2993 o Reduced packet loss (to the point where packet retransmission- 2994 incurred delay would be negligible.) 2996 9.4. Private Blockchain Asks 2998 o Layer 2 and Layer 3 multicast of blockchain traffic 3000 o Item and block delivery with bounded, low latency and negligible 3001 packet loss 3003 o Coexistence in a single network of blockchain and IT traffic. 3005 o Ability to scale the network by distributing the centralized 3006 control of the network across multiple control entities. 3008 10. Network Slicing 3010 10.1. Use Case Description 3012 Network Slicing divides one physical network infrastructure into 3013 multiple logical networks. Each slice, corresponding to a logical 3014 network, uses resources and network functions independently from each 3015 other. Network Slicing provides flexibility of resource allocation 3016 and service quality customization. 3018 Future services will demand network performance with a wide variety 3019 of characteristics such as high data rate, low latency, low loss 3020 rate, security and many other parameters. Ideally every service 3021 would have its own physical network satisfying its particular 3022 performance requirements, however that would be prohibitively 3023 expensive. Network Slicing can provide a customized slice for a 3024 single service, and multiple slices can share the same physical 3025 network. This method can optimize the performance for the service at 3026 lower cost, and the flexibility of setting up and release the slices 3027 also allows the user to allocate the network resources dynamically. 3029 Unlike the other use cases presented here, Network Slicing is not a 3030 specific application that depends on specific deterministic 3031 properties; rather it is introduced as an area of networking to which 3032 DetNet might be applicable. 3034 10.2. DetNet Applied to Network Slicing 3036 10.2.1. Resource Isolation Across Slices 3038 One of the requirements discussed for Network Slicing is the "hard" 3039 separation of various users' deterministic performance. That is, it 3040 should be impossible for activity, lack of activity, or changes in 3041 activity of one or more users to have any appreciable effect on the 3042 deterministic performance parameters of any other slices. Typical 3043 techniques used today, which share a physical network among users, do 3044 not offer this level of isolation. DetNet can supply point-to-point 3045 or point-to-multipoint paths that offer bandwidth and latency 3046 guarantees to a user that cannot be affected by other users' data 3047 traffic. Thus DetNet is a powerful tool when latency and reliability 3048 are required in Network Slicing. 3050 10.2.2. Deterministic Services Within Slices 3052 Slices may need to provide services with DetNet-type performance 3053 guarantees, however we note that a system can be implemented to 3054 provide such services in more than one way. For example the slice 3055 itself might be implemented using DetNet, and thus the slice can 3056 provide service guarantees and isolation to its users without any 3057 particular DetNet awareness on the part of the users' applications. 3058 Alternatively, a "non-DetNet-aware" slice may host an application 3059 that itself implements DetNet services and thus can enjoy similar 3060 service guarantees. 3062 10.3. A Network Slicing Use Case Example - 5G Bearer Network 3064 Network Slicing is a core feature of 5G defined in 3GPP, which is 3065 currently under development. A network slice in a mobile network is 3066 a complete logical network including Radio Access Network (RAN) and 3067 Core Network (CN). It provides telecommunication services and 3068 network capabilities, which may vary from slice to slice. A 5G 3069 bearer network is a typical use case of Network Slicing; for example 3070 consider three 5G service scenarios: eMMB, URLLC, and mMTC. 3072 o eMBB (Enhanced Mobile Broadband) focuses on services characterized 3073 by high data rates, such as high definition videos, virtual 3074 reality, augmented reality, and fixed mobile convergence. 3076 o URLLC (Ultra-Reliable and Low Latency Communications) focuses on 3077 latency-sensitive services, such as self-driving vehicles, remote 3078 surgery, or drone control. 3080 o mMTC (massive Machine Type Communications) focuses on services 3081 that have high requirements for connection density, such as those 3082 typical for smart city and smart agriculture use cases. 3084 A 5G bearer network could use DetNet to provide hard resource 3085 isolation across slices and within the slice. For example consider 3086 Slice-A and Slice-B, with DetNet used to transit services URLLC-A and 3087 URLLC-B over them. Without DetNet, URLLC-A and URLLC-B would compete 3088 for bandwidth resource, and latency and reliability would not be 3089 guaranteed. With DetNet, URLLC-A and URLLC-B have separate bandwidth 3090 reservation and there is no resource conflict between them, as though 3091 they were in different logical networks. 3093 10.4. Non-5G Applications of Network Slicing 3095 Although operation of services not related to 5G is not part of the 3096 5G Network Slicing definition and scope, Network Slicing is likely to 3097 become a preferred approach to providing various services across a 3098 shared physical infrastructure. Examples include providing 3099 electrical utilities services and pro audio services via slices. Use 3100 cases like these could become more common once the work for the 5G 3101 core network evolves to include wired as well as wireless access. 3103 10.5. Limitations of DetNet in Network Slicing 3105 DetNet cannot cover every Network Slicing use case. One issue is 3106 that DetNet is a point-to-point or point-to-multipoint technology, 3107 however Network Slicing ultimately needs multi-point to multi-point 3108 guarantees. Another issue is that the number of flows that can be 3109 carried by DetNet is limited by DetNet scalability; flow aggregation 3110 and queuing management modification may help address this. 3111 Additional work and discussion are needed to address these topics. 3113 10.6. Network Slicing Today and Future 3115 Network Slicing has the promise to satisfy many requirements of 3116 future network deployment scenarios, but it is still a collection of 3117 ideas and analysis, without a specific technical solution. DetNet is 3118 one of various technologies that have potential to be used in Network 3119 Slicing, along with for example Flex-E and Segment Routing. For more 3120 information please see the IETF99 Network Slicing BOF session agenda 3121 and materials. 3123 10.7. Network Slicing Asks 3125 o Isolation from other flows through Queuing Management 3127 o Service Quality Customization and Guarantee 3129 o Security 3131 11. Use Case Common Themes 3133 This section summarizes the expected properties of a DetNet network, 3134 based on the use cases as described in this draft. 3136 11.1. Unified, standards-based network 3138 11.1.1. Extensions to Ethernet 3140 A DetNet network is not "a new kind of network" - it based on 3141 extensions to existing Ethernet standards, including elements of IEEE 3142 802.1 AVB/TSN and related standards. Presumably it will be possible 3143 to run DetNet over other underlying transports besides Ethernet, but 3144 Ethernet is explicitly supported. 3146 11.1.2. Centrally Administered 3148 In general a DetNet network is not expected to be "plug and play" - 3149 it is expected that there is some centralized network configuration 3150 and control system. Such a system may be in a single central 3151 location, or it maybe distributed across multiple control entities 3152 that function together as a unified control system for the network. 3153 However, the ability to "hot swap" components (e.g. due to 3154 malfunction) is similar enough to "plug and play" that this kind of 3155 behavior may be expected in DetNet networks, depending on the 3156 implementation. 3158 11.1.3. Standardized Data Flow Information Models 3160 Data Flow Information Models to be used with DetNet networks are to 3161 be specified by DetNet. 3163 11.1.4. L2 and L3 Integration 3165 A DetNet network is intended to integrate between Layer 2 (bridged) 3166 network(s) (e.g. AVB/TSN LAN) and Layer 3 (routed) network(s) (e.g. 3167 using IP-based protocols). One example of this is "making AVB/TSN- 3168 type deterministic performance available from Layer 3 applications, 3169 e.g. using RTP". Another example is "connecting two AVB/TSN LANs 3170 ("islands") together through a standard router". 3172 11.1.5. Consideration for IPv4 3174 This Use Cases draft explicitly does not specify any particular 3175 implementation or protocol, however it has been observed that various 3176 of the use cases described (and their associated industries) are 3177 explicitly based on IPv4 (as opposed to IPv6) and it is not 3178 considered practical to expect them to migrate to IPv6 in order to 3179 use DetNet. Thus the expectation is that even if not every feature 3180 of DetNet is available in an IPv4 context, at least some of the 3181 significant benefits (such as guaranteed end-to-end delivery and low 3182 latency) are expected to be available. 3184 11.1.6. Guaranteed End-to-End Delivery 3186 Packets sent over DetNet are guaranteed not to be dropped by the 3187 network due to congestion. However, the network may drop packets for 3188 intended reasons, e.g. per security measures. Also note that this 3189 guarantee applies to the actions of DetNet protocol software, and 3190 does not provide any guarantee against lower level errors such as 3191 media errors or checksum errors. 3193 11.1.7. Replacement for Multiple Proprietary Deterministic Networks 3195 There are many proprietary non-interoperable deterministic Ethernet- 3196 based networks currently available; DetNet is intended to provide an 3197 open-standards-based alternative to such networks. 3199 11.1.8. Mix of Deterministic and Best-Effort Traffic 3201 DetNet is intended to support coexistance of time-sensitive 3202 operational (OT) traffic and information (IT) traffic on the same 3203 ("unified") network. 3205 11.1.9. Unused Reserved BW to be Available to Best Effort Traffic 3207 If bandwidth reservations are made for a stream but the associated 3208 bandwidth is not used at any point in time, that bandwidth is made 3209 available on the network for best-effort traffic. If the owner of 3210 the reserved stream then starts transmitting again, the bandwidth is 3211 no longer available for best-effort traffic, on a moment-to-moment 3212 basis. Note that such "temporarily available" bandwidth is not 3213 available for time-sensitive traffic, which must have its own 3214 reservation. 3216 11.1.10. Lower Cost, Multi-Vendor Solutions 3218 The DetNet network specifications are intended to enable an ecosystem 3219 in which multiple vendors can create interoperable products, thus 3220 promoting device diversity and potentially higher numbers of each 3221 device manufactured, promoting cost reduction and cost competition 3222 among vendors. The intent is that DetNet networks should be able to 3223 be created at lower cost and with greater diversity of available 3224 devices than existing proprietary networks. 3226 11.2. Scalable Size 3228 DetNet networks range in size from very small, e.g. inside a single 3229 industrial machine, to very large, for example a Utility Grid network 3230 spanning a whole country, and involving many "hops" over various 3231 kinds of links for example radio repeaters, microwave linkes, fiber 3232 optic links, etc.. However recall that the scope of DetNet is 3233 confined to networks that are centrally administered, and explicitly 3234 excludes unbounded decentralized networks such as the Internet. 3236 11.3. Scalable Timing Parameters and Accuracy 3238 11.3.1. Bounded Latency 3240 The DetNet Data Flow Information Model is expected to provide means 3241 to configure the network that include parameters for querying network 3242 path latency, requesting bounded latency for a given stream, 3243 requesting worst case maximum and/or minimum latency for a given path 3244 or stream, and so on. It is an expected case that the network may 3245 not be able to provide a given requested service level, and if so the 3246 network control system should reply that the requested services is 3247 not available (as opposed to accepting the parameter but then not 3248 delivering the desired behavior). 3250 11.3.2. Low Latency 3252 Applications may require "extremely low latency" however depending on 3253 the application these may mean very different latency values; for 3254 example "low latency" across a Utility grid network is on a different 3255 time scale than "low latency" in a motor control loop in a small 3256 machine. The intent is that the mechanisms for specifying desired 3257 latency include wide ranges, and that architecturally there is 3258 nothing to prevent arbirtrarily low latencies from being implemented 3259 in a given network. 3261 11.3.3. Symmetrical Path Delays 3263 Some applications would like to specify that the transit delay time 3264 values be equal for both the transmit and return paths. 3266 11.4. High Reliability and Availability 3268 Reliablity is of critical importance to many DetNet applications, in 3269 which consequences of failure can be extraordinarily high in terms of 3270 cost and even human life. DetNet based systems are expected to be 3271 implemented with essentially arbitrarily high availability (for 3272 example 99.9999% up time, or even 12 nines). The intent is that the 3273 DetNet designs should not make any assumptions about the level of 3274 reliability and availability that may be required of a given system, 3275 and should define parameters for communicating these kinds of metrics 3276 within the network. 3278 A strategy used by DetNet for providing such extraordinarily high 3279 levels of reliability is to provide redundant paths that can be 3280 seamlessly switched between, while maintaining the required 3281 performance of that system. 3283 11.5. Security 3285 Security is of critical importance to many DetNet applications. A 3286 DetNet network must be able to be made secure against devices 3287 failures, attackers, misbehaving devices, and so on. In a DetNet 3288 network the data traffic is expected to be be time-sensitive, thus in 3289 addition to arriving with the data content as intended, the data must 3290 also arrive at the expected time. This may present "new" security 3291 challenges to implementers, and must be addressed accordingly. There 3292 are other security implications, including (but not limited to) the 3293 change in attack surface presented by packet replication and 3294 elimination. 3296 11.6. Deterministic Flows 3298 Reserved bandwidth data flows must be isolated from each other and 3299 from best-effort traffic, so that even if the network is saturated 3300 with best-effort (and/or reserved bandwidth) traffic, the configured 3301 flows are not adversely affected. 3303 12. Use Cases Explicitly Out of Scope for DetNet 3305 This section contains use case text that has been determined to be 3306 outside of the scope of the present DetNet work. 3308 12.1. DetNet Scope Limitations 3310 The scope of DetNet is deliberately limited to specific use cases 3311 that are consistent with the WG charter, subject to the 3312 interpretation of the WG. At the time the DetNet Use Cases were 3313 solicited and provided by the authors the scope of DetNet was not 3314 clearly defined, and as that clarity has emerged, certain of the use 3315 cases have been determined to be outside the scope of the present 3316 DetNet work. Such text has been moved into this section to clarify 3317 that these use cases will not be supported by the DetNet work. 3319 The text in this section was moved here based on the following 3320 "exclusion" principles. Or, as an alternative to moving all such 3321 text to this section, some draft text has been modified in situ to 3322 reflect these same principles. 3324 The following principles have been established to clarify the scope 3325 of the present DetNet work. 3327 o The scope of network addressed by DetNet is limited to networks 3328 that can be centrally controlled, i.e. an "enterprise" aka 3329 "corporate" network. This explicitly excludes "the open 3330 Internet". 3332 o Maintaining synchronized time across a DetNet network is crucial 3333 to its operation, however DetNet assumes that time is to be 3334 maintained using other means, for example (but not limited to) 3335 Precision Time Protocol ([IEEE1588]). A use case may state the 3336 accuracy and reliability that it expects from the DetNet network 3337 as part of a whole system, however it is understood that such 3338 timing properties are not guaranteed by DetNet itself. It is 3339 currently an open question as to whether DetNet protocols will 3340 include a way for an application to communicate such timing 3341 expectations to the network, and if so whether they would be 3342 expected to materially affect the performance they would receive 3343 from the network as a result. 3345 12.2. Internet-based Applications 3347 There are many applications that communicate over the open Internet 3348 that could benefit from guaranteed delivery and bounded latency. 3349 However as noted above, all such applications when run over the open 3350 Internet are out of scope for DetNet. These same applications may be 3351 in-scope when run in constrained environments, i.e. within a 3352 centrally controlled DetNet network. The following are some examples 3353 of such applications. 3355 12.2.1. Use Case Description 3357 12.2.1.1. Media Content Delivery 3359 Media content delivery continues to be an important use of the 3360 Internet, yet users often experience poor quality audio and video due 3361 to the delay and jitter inherent in today's Internet. 3363 12.2.1.2. Online Gaming 3365 Online gaming is a significant part of the gaming market, however 3366 latency can degrade the end user experience. For example "First 3367 Person Shooter" games are highly delay-sensitive. 3369 12.2.1.3. Virtual Reality 3371 Virtual reality has many commercial applications including real 3372 estate presentations, remote medical procedures, and so on. Low 3373 latency is critical to interacting with the virtual world because 3374 perceptual delays can cause motion sickness. 3376 12.2.2. Internet-Based Applications Today 3378 Internet service today is by definition "best effort", with no 3379 guarantees on delivery or bandwidth. 3381 12.2.3. Internet-Based Applications Future 3383 We imagine an Internet from which we will be able to play a video 3384 without glitches and play games without lag. 3386 For online gaming, the maximum round-trip delay can be 100ms and 3387 stricter for FPS gaming which can be 10-50ms. Transport delay is the 3388 dominate part with a 5-20ms budget. 3390 For VR, 1-10ms maximum delay is needed and total network budget is 3391 1-5ms if doing remote VR. 3393 Flow identification can be used for gaming and VR, i.e. it can 3394 recognize a critical flow and provide appropriate latency bounds. 3396 12.2.4. Internet-Based Applications Asks 3398 o Unified control and management protocols to handle time-critical 3399 data flow 3401 o Application-aware flow filtering mechanism to recognize the timing 3402 critical flow without doing 5-tuple matching 3404 o Unified control plane to provide low latency service on Layer-3 3405 without changing the data plane 3407 o OAM system and protocols which can help to provide E2E-delay 3408 sensitive service provisioning 3410 12.3. Pro Audio and Video - Digital Rights Management (DRM) 3412 This section was moved here because this is considered a Link layer 3413 topic, not direct responsibility of DetNet. 3415 Digital Rights Management (DRM) is very important to the audio and 3416 video industries. Any time protected content is introduced into a 3417 network there are DRM concerns that must be maintained (see 3418 [CONTENT_PROTECTION]). Many aspects of DRM are outside the scope of 3419 network technology, however there are cases when a secure link 3420 supporting authentication and encryption is required by content 3421 owners to carry their audio or video content when it is outside their 3422 own secure environment (for example see [DCI]). 3424 As an example, two techniques are Digital Transmission Content 3425 Protection (DTCP) and High-Bandwidth Digital Content Protection 3426 (HDCP). HDCP content is not approved for retransmission within any 3427 other type of DRM, while DTCP may be retransmitted under HDCP. 3428 Therefore if the source of a stream is outside of the network and it 3429 uses HDCP protection it is only allowed to be placed on the network 3430 with that same HDCP protection. 3432 12.4. Pro Audio and Video - Link Aggregation 3434 Note: The term "Link Aggregation" is used here as defined by the text 3435 in the following paragraph, i.e. not following a more common Network 3436 Industry definition. Current WG consensus is that this item won't be 3437 directly supported by the DetNet architecture, for example because it 3438 implies guarantee of in-order delivery of packets which conflicts 3439 with the core goal of achieving the lowest possible latency. 3441 For transmitting streams that require more bandwidth than a single 3442 link in the target network can support, link aggregation is a 3443 technique for combining (aggregating) the bandwidth available on 3444 multiple physical links to create a single logical link of the 3445 required bandwidth. However, if aggregation is to be used, the 3446 network controller (or equivalent) must be able to determine the 3447 maximum latency of any path through the aggregate link. 3449 13. Contributors 3451 RFC7322 limits the number of authors listed on the front page of a 3452 draft to a maximum of 5, far fewer than the 20 individuals below who 3453 made important contributions to this draft. The editor wishes to 3454 thank and acknowledge each of the following authors for contributing 3455 text to this draft. See also Section 14. 3457 Craig Gunther (Harman International) 3458 10653 South River Front Parkway, South Jordan,UT 84095 3459 phone +1 801 568-7675, email craig.gunther@harman.com 3461 Pascal Thubert (Cisco Systems, Inc) 3462 Building D, 45 Allee des Ormes - BP1200, MOUGINS 3463 Sophia Antipolis 06254 FRANCE 3464 phone +33 497 23 26 34, email pthubert@cisco.com 3466 Patrick Wetterwald (Cisco Systems) 3467 45 Allees des Ormes, Mougins, 06250 FRANCE 3468 phone +33 4 97 23 26 36, email pwetterw@cisco.com 3470 Jean Raymond (Hydro-Quebec) 3471 1500 University, Montreal, H3A3S7, Canada 3472 phone +1 514 840 3000, email raymond.jean@hydro.qc.ca 3474 Jouni Korhonen (Broadcom Corporation) 3475 3151 Zanker Road, San Jose, 95134, CA, USA 3476 email jouni.nospam@gmail.com 3478 Yu Kaneko (Toshiba) 3479 1 Komukai-Toshiba-cho, Saiwai-ku, Kasasaki-shi, Kanagawa, Japan 3480 email yu1.kaneko@toshiba.co.jp 3482 Subir Das (Vencore Labs) 3483 150 Mount Airy Road, Basking Ridge, New Jersey, 07920, USA 3484 email sdas@appcomsci.com 3486 Balazs Varga (Ericsson) 3487 Konyves Kalman krt. 11/B, Budapest, Hungary, 1097 3488 email balazs.a.varga@ericsson.com 3489 Janos Farkas (Ericsson) 3490 Konyves Kalman krt. 11/B, Budapest, Hungary, 1097 3491 email janos.farkas@ericsson.com 3493 Franz-Josef Goetz (Siemens) 3494 Gleiwitzerstr. 555, Nurnberg, Germany, 90475 3495 email franz-josef.goetz@siemens.com 3497 Juergen Schmitt (Siemens) 3498 Gleiwitzerstr. 555, Nurnberg, Germany, 90475 3499 email juergen.jues.schmitt@siemens.com 3501 Xavier Vilajosana (Worldsensing) 3502 483 Arago, Barcelona, Catalonia, 08013, Spain 3503 email xvilajosana@worldsensing.com 3505 Toktam Mahmoodi (King's College London) 3506 Strand, London WC2R 2LS, United Kingdom 3507 email toktam.mahmoodi@kcl.ac.uk 3509 Spiros Spirou (Intracom Telecom) 3510 19.7 km Markopoulou Ave., Peania, Attiki, 19002, Greece 3511 email spiros.spirou@gmail.com 3513 Petra Vizarreta (Technical University of Munich) 3514 Maxvorstadt, ArcisstraBe 21, Munich, 80333, Germany 3515 email petra.stojsavljevic@tum.de 3517 Daniel Huang (ZTE Corporation, Inc.) 3518 No. 50 Software Avenue, Nanjing, Jiangsu, 210012, P.R. China 3519 email huang.guangping@zte.com.cn 3521 Xuesong Geng (Huawei Technologies) 3522 email gengxuesong@huawei.com 3524 Diego Dujovne (Universidad Diego Portales) 3525 email diego.dujovne@mail.udp.cl 3527 Maik Seewald (Cisco Systems) 3528 email maseewal@cisco.com 3530 14. Acknowledgments 3532 14.1. Pro Audio 3534 This section was derived from draft-gunther-detnet-proaudio-req-01. 3536 The editors would like to acknowledge the help of the following 3537 individuals and the companies they represent: 3539 Jeff Koftinoff, Meyer Sound 3541 Jouni Korhonen, Associate Technical Director, Broadcom 3543 Pascal Thubert, CTAO, Cisco 3545 Kieran Tyrrell, Sienda New Media Technologies GmbH 3547 14.2. Utility Telecom 3549 This section was derived from draft-wetterwald-detnet-utilities-reqs- 3550 02. 3552 Faramarz Maghsoodlou, Ph. D. IoT Connected Industries and Energy 3553 Practice Cisco 3555 Pascal Thubert, CTAO Cisco 3557 14.3. Building Automation Systems 3559 This section was derived from draft-bas-usecase-detnet-00. 3561 14.4. Wireless for Industrial 3563 This section was derived from draft-thubert-6tisch-4detnet-01. 3565 This specification derives from the 6TiSCH architecture, which is the 3566 result of multiple interactions, in particular during the 6TiSCH 3567 (bi)Weekly Interim call, relayed through the 6TiSCH mailing list at 3568 the IETF. 3570 The authors wish to thank: Kris Pister, Thomas Watteyne, Xavier 3571 Vilajosana, Qin Wang, Tom Phinney, Robert Assimiti, Michael 3572 Richardson, Zhuo Chen, Malisa Vucinic, Alfredo Grieco, Martin Turon, 3573 Dominique Barthel, Elvis Vogli, Guillaume Gaillard, Herman Storey, 3574 Maria Rita Palattella, Nicola Accettura, Patrick Wetterwald, Pouria 3575 Zand, Raghuram Sudhaakar, and Shitanshu Shah for their participation 3576 and various contributions. 3578 14.5. Cellular Radio 3580 This section was derived from draft-korhonen-detnet-telreq-00. 3582 14.6. Industrial M2M 3584 The authors would like to thank Feng Chen and Marcel Kiessling for 3585 their comments and suggestions. 3587 14.7. Internet Applications and CoMP 3589 This section was derived from draft-zha-detnet-use-case-00 by Yiyong 3590 Zha. 3592 This document has benefited from reviews, suggestions, comments and 3593 proposed text provided by the following members, listed in 3594 alphabetical order: Jing Huang, Junru Lin, Lehong Niu and Oilver 3595 Huang. 3597 14.8. Electrical Utilities 3599 The wind power generation use case has been extracted from the study 3600 of Wind Farms conducted within the 5GPPP Virtuwind Project. The 3601 project is funded by the European Union's Horizon 2020 research and 3602 innovation programme under grant agreement No 671648 (VirtuWind). 3604 14.9. Network Slicing 3606 This section was written by Xuesong Geng, who would like to 3607 acknowledge Norm Finn and Mach Chen for their useful comments. 3609 14.10. Mining 3611 This section was written by Diego Dujovne in conjunction with Xavier 3612 Vilasojana. 3614 14.11. Private Blockchain 3616 This section was written by Daniel Huang. 3618 15. IANA Considerations 3620 This memo includes no requests from IANA. 3622 16. Informative References 3624 [Ahm14] Ahmed, M. and R. Kim, "Communication network architectures 3625 for smart-wind power farms.", Energies, p. 3900-3921. , 3626 June 2014. 3628 [bacnetip] 3629 ASHRAE, "Annex J to ANSI/ASHRAE 135-1995 - BACnet/IP", 3630 January 1999. 3632 [CoMP] NGMN Alliance, "RAN EVOLUTION PROJECT COMP EVALUATION AND 3633 ENHANCEMENT", NGMN Alliance NGMN_RANEV_D3_CoMP_Evaluation_ 3634 and_Enhancement_v2.0, March 2015, 3635 . 3638 [CONTENT_PROTECTION] 3639 Olsen, D., "1722a Content Protection", 2012, 3640 . 3643 [CPRI] CPRI Cooperation, "Common Public Radio Interface (CPRI); 3644 Interface Specification", CPRI Specification V6.1, July 3645 2014, . 3648 [DCI] Digital Cinema Initiatives, LLC, "DCI Specification, 3649 Version 1.2", 2012, . 3651 [eCPRI] IEEE Standards Association, "Common Public Radio 3652 Interface, "Common Public Radio Interface: eCPRI Interface 3653 Specification V1.0", 2017, . 3655 [ESPN_DC2] 3656 Daley, D., "ESPN's DC2 Scales AVB Large", 2014, 3657 . 3660 [flnet] Japan Electrical Manufacturers Association, "JEMA 1479 - 3661 English Edition", September 2012. 3663 [Fronthaul] 3664 Chen, D. and T. Mustala, "Ethernet Fronthaul 3665 Considerations", IEEE 1904.3, February 2015, 3666 . 3669 [I-D.ietf-6tisch-6top-interface] 3670 Wang, Q. and X. Vilajosana, "6TiSCH Operation Sublayer 3671 (6top) Interface", draft-ietf-6tisch-6top-interface-04 3672 (work in progress), July 2015. 3674 [I-D.ietf-6tisch-architecture] 3675 Thubert, P., "An Architecture for IPv6 over the TSCH mode 3676 of IEEE 802.15.4", draft-ietf-6tisch-architecture-14 (work 3677 in progress), April 2018. 3679 [I-D.ietf-6tisch-coap] 3680 Sudhaakar, R. and P. Zand, "6TiSCH Resource Management and 3681 Interaction using CoAP", draft-ietf-6tisch-coap-03 (work 3682 in progress), March 2015. 3684 [I-D.ietf-detnet-architecture] 3685 Finn, N., Thubert, P., Varga, B., and J. Farkas, 3686 "Deterministic Networking Architecture", draft-ietf- 3687 detnet-architecture-07 (work in progress), August 2018. 3689 [I-D.ietf-detnet-problem-statement] 3690 Finn, N. and P. Thubert, "Deterministic Networking Problem 3691 Statement", draft-ietf-detnet-problem-statement-06 (work 3692 in progress), July 2018. 3694 [I-D.ietf-tictoc-1588overmpls] 3695 Davari, S., Oren, A., Bhatia, M., Roberts, P., and L. 3696 Montini, "Transporting Timing messages over MPLS 3697 Networks", draft-ietf-tictoc-1588overmpls-07 (work in 3698 progress), October 2015. 3700 [I-D.kh-spring-ip-ran-use-case] 3701 Khasnabish, B., hu, f., and L. Contreras, "Segment Routing 3702 in IP RAN use case", draft-kh-spring-ip-ran-use-case-02 3703 (work in progress), November 2014. 3705 [I-D.svshah-tsvwg-deterministic-forwarding] 3706 Shah, S. and P. Thubert, "Deterministic Forwarding PHB", 3707 draft-svshah-tsvwg-deterministic-forwarding-04 (work in 3708 progress), August 2015. 3710 [I-D.wang-6tisch-6top-sublayer] 3711 Wang, Q. and X. Vilajosana, "6TiSCH Operation Sublayer 3712 (6top)", draft-wang-6tisch-6top-sublayer-04 (work in 3713 progress), November 2015. 3715 [IEC-60870-5-104] 3716 International Electrotechnical Commission, "International 3717 Standard IEC 60870-5-104: Network access for IEC 3718 60870-5-101 using standard transport profiles", June 2006. 3720 [IEC61400] 3721 "International standard 61400-25: Communications for 3722 monitoring and control of wind power plants", June 2013. 3724 [IEEE1588] 3725 IEEE, "IEEE Standard for a Precision Clock Synchronization 3726 Protocol for Networked Measurement and Control Systems", 3727 IEEE Std 1588-2008, 2008, 3728 . 3731 [IEEE1646] 3732 "Communication Delivery Time Performance Requirements for 3733 Electric Power Substation Automation", IEEE Standard 3734 1646-2004 , Apr 2004. 3736 [IEEE1722] 3737 IEEE, "1722-2011 - IEEE Standard for Layer 2 Transport 3738 Protocol for Time Sensitive Applications in a Bridged 3739 Local Area Network", IEEE Std 1722-2011, 2011, 3740 . 3743 [IEEE19143] 3744 IEEE Standards Association, "P1914.3/D3.1 Draft Standard 3745 for Radio over Ethernet Encapsulations and Mappings", 3746 IEEE 1914.3, 2018, 3747 . 3749 [IEEE802.1TSNTG] 3750 IEEE Standards Association, "IEEE 802.1 Time-Sensitive 3751 Networks Task Group", March 2013, 3752 . 3754 [IEEE802154] 3755 IEEE standard for Information Technology, "IEEE std. 3756 802.15.4, Part. 15.4: Wireless Medium Access Control (MAC) 3757 and Physical Layer (PHY) Specifications for Low-Rate 3758 Wireless Personal Area Networks". 3760 [IEEE802154e] 3761 IEEE standard for Information Technology, "IEEE standard 3762 for Information Technology, IEEE std. 802.15.4, Part. 3763 15.4: Wireless Medium Access Control (MAC) and Physical 3764 Layer (PHY) Specifications for Low-Rate Wireless Personal 3765 Area Networks, June 2011 as amended by IEEE std. 3766 802.15.4e, Part. 15.4: Low-Rate Wireless Personal Area 3767 Networks (LR-WPANs) Amendment 1: MAC sublayer", April 3768 2012. 3770 [IEEE8021AS] 3771 IEEE, "Timing and Synchronizations (IEEE 802.1AS-2011)", 3772 IEEE 802.1AS-2001, 2011, 3773 . 3776 [IEEE8021CM] 3777 Farkas, J., "Time-Sensitive Networking for Fronthaul", 3778 Unapproved PAR, PAR for a New IEEE Standard; 3779 IEEE P802.1CM, April 2015, 3780 . 3783 [ISA100] ISA/ANSI, "ISA100, Wireless Systems for Automation", 3784 . 3786 [knx] KNX Association, "ISO/IEC 14543-3 - KNX", November 2006. 3788 [lontalk] ECHELON, "LonTalk(R) Protocol Specification Version 3.0", 3789 1994. 3791 [MEF22.1.1] 3792 MEF, "Mobile Backhaul Phase 2 Amendment 1 -- Small Cells", 3793 MEF 22.1.1, July 2014, 3794 . 3797 [MEF8] MEF, "Implementation Agreement for the Emulation of PDH 3798 Circuits over Metro Ethernet Networks", MEF 8, October 3799 2004, 3800 . 3803 [METIS] METIS, "Scenarios, requirements and KPIs for 5G mobile and 3804 wireless system", ICT-317669-METIS/D1.1 ICT- 3805 317669-METIS/D1.1, April 2013, . 3808 [modbus] Modbus Organization, "MODBUS APPLICATION PROTOCOL 3809 SPECIFICATION V1.1b", December 2006. 3811 [MODBUS] Modbus Organization, Inc., "MODBUS Application Protocol 3812 Specification", Apr 2012. 3814 [NGMN] NGMN Alliance, "5G White Paper", NGMN 5G White Paper v1.0, 3815 February 2015, . 3818 [NGMN-fronth] 3819 NGMN Alliance, "Fronthaul Requirements for C-RAN", March 3820 2015, . 3823 [OPCXML] OPC Foundation, "OPC XML-Data Access Specification", Dec 3824 2004. 3826 [PCE] IETF, "Path Computation Element", 3827 . 3829 [profibus] 3830 IEC, "IEC 61158 Type 3 - Profibus DP", January 2001. 3832 [RFC3031] Rosen, E., Viswanathan, A., and R. Callon, "Multiprotocol 3833 Label Switching Architecture", RFC 3031, 3834 DOI 10.17487/RFC3031, January 2001, 3835 . 3837 [RFC3411] Harrington, D., Presuhn, R., and B. Wijnen, "An 3838 Architecture for Describing Simple Network Management 3839 Protocol (SNMP) Management Frameworks", STD 62, RFC 3411, 3840 DOI 10.17487/RFC3411, December 2002, 3841 . 3843 [RFC3985] Bryant, S., Ed. and P. Pate, Ed., "Pseudo Wire Emulation 3844 Edge-to-Edge (PWE3) Architecture", RFC 3985, 3845 DOI 10.17487/RFC3985, March 2005, 3846 . 3848 [RFC4553] Vainshtein, A., Ed. and YJ. Stein, Ed., "Structure- 3849 Agnostic Time Division Multiplexing (TDM) over Packet 3850 (SAToP)", RFC 4553, DOI 10.17487/RFC4553, June 2006, 3851 . 3853 [RFC5086] Vainshtein, A., Ed., Sasson, I., Metz, E., Frost, T., and 3854 P. Pate, "Structure-Aware Time Division Multiplexed (TDM) 3855 Circuit Emulation Service over Packet Switched Network 3856 (CESoPSN)", RFC 5086, DOI 10.17487/RFC5086, December 2007, 3857 . 3859 [RFC5087] Stein, Y(J)., Shashoua, R., Insler, R., and M. Anavi, 3860 "Time Division Multiplexing over IP (TDMoIP)", RFC 5087, 3861 DOI 10.17487/RFC5087, December 2007, 3862 . 3864 [RFC6550] Winter, T., Ed., Thubert, P., Ed., Brandt, A., Hui, J., 3865 Kelsey, R., Levis, P., Pister, K., Struik, R., Vasseur, 3866 JP., and R. Alexander, "RPL: IPv6 Routing Protocol for 3867 Low-Power and Lossy Networks", RFC 6550, 3868 DOI 10.17487/RFC6550, March 2012, 3869 . 3871 [RFC6551] Vasseur, JP., Ed., Kim, M., Ed., Pister, K., Dejean, N., 3872 and D. Barthel, "Routing Metrics Used for Path Calculation 3873 in Low-Power and Lossy Networks", RFC 6551, 3874 DOI 10.17487/RFC6551, March 2012, 3875 . 3877 [RFC7554] Watteyne, T., Ed., Palattella, M., and L. Grieco, "Using 3878 IEEE 802.15.4e Time-Slotted Channel Hopping (TSCH) in the 3879 Internet of Things (IoT): Problem Statement", RFC 7554, 3880 DOI 10.17487/RFC7554, May 2015, 3881 . 3883 [RFC8169] Mirsky, G., Ruffini, S., Gray, E., Drake, J., Bryant, S., 3884 and A. Vainshtein, "Residence Time Measurement in MPLS 3885 Networks", RFC 8169, DOI 10.17487/RFC8169, May 2017, 3886 . 3888 [Spe09] Sperotto, A., Sadre, R., Vliet, F., and A. Pras, "A First 3889 Look into SCADA Network Traffic", IP Operations and 3890 Management, p. 518-521. , June 2009. 3892 [SRP_LATENCY] 3893 Gunther, C., "Specifying SRP Latency", 2014, 3894 . 3897 [SyncE] ITU-T, "G.8261 : Timing and synchronization aspects in 3898 packet networks", Recommendation G.8261, August 2013, 3899 . 3901 [TR38801] IEEE Standards Association, "3GPP TR 38.801, Technical 3902 Specification Group Radio Access Network; Study on new 3903 radio access technology: Radio access architecture and 3904 interfaces (Release 14)", 2017, 3905 . 3908 [TS23401] 3GPP, "General Packet Radio Service (GPRS) enhancements 3909 for Evolved Universal Terrestrial Radio Access Network 3910 (E-UTRAN) access", 3GPP TS 23.401 10.10.0, March 2013. 3912 [TS25104] 3GPP, "Base Station (BS) radio transmission and reception 3913 (FDD)", 3GPP TS 25.104 3.14.0, March 2007. 3915 [TS36104] 3GPP, "Evolved Universal Terrestrial Radio Access 3916 (E-UTRA); Base Station (BS) radio transmission and 3917 reception", 3GPP TS 36.104 10.11.0, July 2013. 3919 [TS36133] 3GPP, "Evolved Universal Terrestrial Radio Access 3920 (E-UTRA); Requirements for support of radio resource 3921 management", 3GPP TS 36.133 12.7.0, April 2015. 3923 [TS36211] 3GPP, "Evolved Universal Terrestrial Radio Access 3924 (E-UTRA); Physical channels and modulation", 3GPP 3925 TS 36.211 10.7.0, March 2013. 3927 [TS36300] 3GPP, "Evolved Universal Terrestrial Radio Access (E-UTRA) 3928 and Evolved Universal Terrestrial Radio Access Network 3929 (E-UTRAN); Overall description; Stage 2", 3GPP TS 36.300 3930 10.11.0, September 2013. 3932 [TSNTG] IEEE Standards Association, "IEEE 802.1 Time-Sensitive 3933 Networks Task Group", 2013, 3934 . 3936 [WirelessHART] 3937 www.hartcomm.org, "Industrial Communication Networks - 3938 Wireless Communication Network and Communication Profiles 3939 - WirelessHART - IEC 62591", 2010. 3941 Author's Address 3942 Ethan Grossman (editor) 3943 Dolby Laboratories, Inc. 3944 1275 Market Street 3945 San Francisco, CA 94103 3946 USA 3948 Phone: +1 415 645 4726 3949 Email: ethan.grossman@dolby.com 3950 URI: http://www.dolby.com