idnits 2.17.1 draft-song-ntf-00.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 : ---------------------------------------------------------------------------- ** The document seems to lack an Introduction section. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year == The document doesn't use any RFC 2119 keywords, yet seems to have RFC 2119 boilerplate text. -- The document date (March 1, 2018) is 2245 days in the past. Is this intentional? Checking references for intended status: Informational ---------------------------------------------------------------------------- ** Obsolete normative reference: RFC 7540 (Obsoleted by RFC 9113) == Outdated reference: A later version (-05) exists of draft-ietf-netconf-udp-pub-channel-01 == Outdated reference: A later version (-25) exists of draft-ietf-netconf-yang-push-14 == Outdated reference: A later version (-01) exists of draft-openconfig-rtgwg-gnmi-spec-00 == Outdated reference: A later version (-10) exists of draft-zhou-netconf-multi-stream-originators-01 Summary: 2 errors (**), 0 flaws (~~), 6 warnings (==), 1 comment (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Network Working Group H. Song, Ed. 3 Internet-Draft T. Zhou 4 Intended status: Informational Z. Li 5 Expires: September 2, 2018 Huawei 6 March 1, 2018 8 Toward a Network Telemetry Framework 9 draft-song-ntf-00 11 Abstract 13 This document suggests the necessity of a framework of network 14 telemetry and articulates the categories and components of such a 15 framework. The requirement, challenges, existing solutions, and 16 future directions are discussed for each category of the framework. 17 The framework for network telemetry helps to set a common ground for 18 the collection of related works and put future developments into 19 perspective. 21 Requirements Language 23 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 24 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 25 document are to be interpreted as described in RFC 2119 [RFC2119]. 27 Status of This Memo 29 This Internet-Draft is submitted in full conformance with the 30 provisions of BCP 78 and BCP 79. 32 Internet-Drafts are working documents of the Internet Engineering 33 Task Force (IETF). Note that other groups may also distribute 34 working documents as Internet-Drafts. The list of current Internet- 35 Drafts is at https://datatracker.ietf.org/drafts/current/. 37 Internet-Drafts are draft documents valid for a maximum of six months 38 and may be updated, replaced, or obsoleted by other documents at any 39 time. It is inappropriate to use Internet-Drafts as reference 40 material or to cite them other than as "work in progress." 42 This Internet-Draft will expire on September 2, 2018. 44 Copyright Notice 46 Copyright (c) 2018 IETF Trust and the persons identified as the 47 document authors. All rights reserved. 49 This document is subject to BCP 78 and the IETF Trust's Legal 50 Provisions Relating to IETF Documents 51 (https://trustee.ietf.org/license-info) in effect on the date of 52 publication of this document. Please review these documents 53 carefully, as they describe your rights and restrictions with respect 54 to this document. Code Components extracted from this document must 55 include Simplified BSD License text as described in Section 4.e of 56 the Trust Legal Provisions and are provided without warranty as 57 described in the Simplified BSD License. 59 Table of Contents 61 1. Motivation . . . . . . . . . . . . . . . . . . . . . . . . . 2 62 1.1. Use Cases . . . . . . . . . . . . . . . . . . . . . . . . 3 63 1.2. Terminology and Abbreviations . . . . . . . . . . . . . . 4 64 1.3. Network Telemetry . . . . . . . . . . . . . . . . . . . . 5 65 1.4. The Necessity of a Network Telemetry Framework . . . . . 6 66 2. Network Telemetry Framework . . . . . . . . . . . . . . . . . 7 67 2.1. Existing Works Mapped in the Framework . . . . . . . . . 9 68 2.2. Management Plane Telemetry . . . . . . . . . . . . . . . 10 69 2.2.1. Requirements and Challenges . . . . . . . . . . . . . 10 70 2.2.2. Push Extensions for NETCONF . . . . . . . . . . . . . 11 71 2.2.3. gRPC Network Management Interface . . . . . . . . . . 11 72 2.3. Control Plane Telemetry . . . . . . . . . . . . . . . . . 12 73 2.3.1. Requirements and Challenges . . . . . . . . . . . . . 12 74 2.3.2. BGP Monitoring Protocol . . . . . . . . . . . . . . . 12 75 2.4. Data Plane Telemetry . . . . . . . . . . . . . . . . . . 12 76 2.4.1. Requirements and Challenges . . . . . . . . . . . . . 12 77 2.4.2. Dynamic Network Probe . . . . . . . . . . . . . . . . 13 78 2.4.3. IP Flow Information Export (IPFIX) protocol . . . . . 13 79 2.4.4. In-Situ OAM . . . . . . . . . . . . . . . . . . . . . 13 80 3. Security Considerations . . . . . . . . . . . . . . . . . . . 14 81 4. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 14 82 5. Contributors . . . . . . . . . . . . . . . . . . . . . . . . 14 83 6. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 14 84 7. References . . . . . . . . . . . . . . . . . . . . . . . . . 14 85 7.1. Normative References . . . . . . . . . . . . . . . . . . 14 86 7.2. Informative References . . . . . . . . . . . . . . . . . 15 87 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 16 89 1. Motivation 91 Intent-based automatic network is the logical next step of network 92 evolution, aiming to reduce human labor, make the most efficient use 93 of network resources, and provide better services. Tools based on 94 machine learning technologies and big data analytics are powerful for 95 faults, anomaly, pattern, and policy violation detection. Some tools 96 can even predict future events based on history data. The 97 observation and inference from the network data can guide the network 98 policy updates for planning, intrusion prevention, optimization, and 99 self-healing. A closed control loop is therefore achieved. 101 Network OAM provides necessary visibilities to a network. It plays 102 an important role in Intend-based Networks (IBN). 104 1.1. Use Cases 106 Specifically, we have identified a few key network OAM use cases that 107 service providers need the most. All these use cases involves the 108 data extracted from the network data plane and sometimes from the 109 network control plane and management plane: 111 Policy Compliance: Network policies are the rules that constraint 112 the services for network access. For example, a service function 113 chain is a policy that requires the selected flows to pass through 114 a set of network functions in order. While a policy is enforced, 115 the compliance needs to be monitored continuously. 117 SLA Compliance: A service-level agreement defines the level of 118 service a user expects from a service provider, which include the 119 metrics for the service measurement and remedy/penalty procedures 120 when the service level misses the agreement. Users need to check 121 if they get the service as promised and service providers need to 122 evaluate how they can deliver the services that can meet the 123 Service Level Agreement (SLA). 125 Root Cause Analysis: Network failure often involves a sequence of 126 chain events and the source of the failure is not straightforward 127 to identify, especially when t:0he failure is sporadic. While 128 machine learning or other data analytics technologies can be used 129 for root cause analysis, it up to the network to provide all the 130 relevant data for analysis. 132 Load Balancing and Traffic Engineering: Service providers are 133 motivated to optimize their network utilization for better ROI or 134 lower CAPEX. The first step is to know the real-time network 135 condition before applying policies to steer the user traffic or 136 adjust the load balancing algorithm. In some cases the network 137 micro-bursts need to be detected in a very short time-frame so 138 does the fine grained traffic control can be applied to avoid the 139 possible network congestion. 141 Packet Drop Detection: Sporadic packet drops in networks are 142 notoriously hard to locate and debug. Network operators are 143 plagued by the lack of tools that can identify the packet drop 144 locations and reasons. Both active and passive measurements are 145 not very effective in solving this problem. 147 These use cases show that the conventional OAM techniques are not 148 enough for the following reasons: 150 o Most use cases need to continuously monitor the network and 151 dynamically refine the data collection in real-time. The poll- 152 based low-frequency data collection is ill-suited for these 153 applications. Streaming data directly pushed from the data source 154 is preferred. 156 o Various data are needed from any place ranging from the packet 157 processing engine to the QoS traffic manager. Traditional data 158 plane devices cannot provide the necessary probes. An open and 159 programmable data plane is therefore needed. 161 o Many application scenarios need to correlate data from multiple 162 sources (e.g., from distributed nodes or from different network 163 plane). A piecemeal solution is often lack of the capability to 164 consolidate the data from multiple sources. The composition of a 165 complete solution can be guided by a comprehensive framework. 167 o The passive measurement techniques can either consume too much 168 network resources and render too much redundant data, or lead to 169 inaccurate results. The active measurement techniques are 170 indirect, and they can interfere with the user traffic. We need 171 techniques that can collect direct and on-demand data from user 172 traffic. 174 1.2. Terminology and Abbreviations 176 AI: Artificial Intelligence. Use machine-learning based 177 technologies to automate network operation. 179 BMP: BGP Monitoring Protocol 181 DNP: Dynamic Network Probe 183 gNMI: gPRC Network Managment Interface 185 gRPC: gRPC Remote Procesure Call 187 IBN: Intent-Based Network 189 IPFIX: IP Flow Information Export Protocol 191 IPFPM: IP Flow Performance Measurement 192 IOAM: In-situ OAM 194 NETCONF: Network Configuration Protocol 196 Network Telemetry: A general term for techniques to gain network 197 visibility, through network data collection for analysis and 198 measurement. 200 NMS: Network Management System 202 OAM: Operations, Administration, and Maintenance. A group of 203 network management functions that provide network fault 204 indication, fault localization, performance information, and data 205 and diagnosis functions. 207 SNMP: Simple Network Managment Protocol 209 YANG: A data modeling language for NETCONF 211 YANG FSM: A YANG model to define device side finite state machine 213 YANG PUSH: A method to subscribe pushed data from remote YANG 214 datastore 216 1.3. Network Telemetry 218 For a long time, network OAM applications rely on protocols such as 219 SNMP [RFC1157] to monitor the networks. SNMP can only provide 220 limited information about the network. Since SNMP is poll-based, it 221 incurs low data rate and high processing overhead. Such drawbacks 222 make SNMP unsuitable for today's automatic network applications. 224 Network telemetry has emerged as a mainstream technical term to refer 225 to the newer technologies of data collection and consumption in the 226 IBN paradigm, distinguishing itself form the convention technologies 227 for network OAM. It is expected that the network telemetry can 228 provide the necessary network visibility for automated network OAM, 229 address the shortcomings of the conventional technologies, and allow 230 the emergence of new technologies. 232 Although the network telemetry technologies continue evolving, 233 several defining characteristics of network telemetry have been well 234 accepted: 236 o Instead of polling data from the network devices, the telemetry 237 collector subscribes the streaming data pushed from the data 238 source in network devices. 240 o The data is normalized and encoded efficiently for export. 242 o The data is model-based which allows applications to configure and 243 consume data with ease. 245 In addition, we believe the ideal network telemetry should also 246 support the following features: 248 o The data can be customized at runtime to cater the specific need 249 of applications. This needs the support of a programmable data 250 plane which allows probes to be deployed at flexible locations. 252 o The data for a single application can come from multiple data 253 sources (e.g., cross domain, cross device, and cross layer) and 254 need to be correlated to take effect. 256 1.4. The Necessity of a Network Telemetry Framework 258 Big data analytics and machine-learning based AI technologies are 259 applied for network OAM, relying on abundant data from networks. The 260 single-sourced and static data acquisition cannot meet the data 261 requirements. It is desired to have a framework that integrates 262 multiple telemetry approaches from different layers and angels, and 263 allows flexible combinations for different applications. The 264 framework will benefit the application development for the following 265 reasons. 267 o Network visibility presents multiple viewpoints. For example, the 268 device viewpoint takes the network infrastructure as the 269 monitoring object from which the network topology and device 270 status can be acquired; the traffic viewpoint takes the flows or 271 packets as the monitoring object from which the traffic quality 272 and path can be acquired. An application may need to switch its 273 viewpoint during operation. It may also need to correlate a 274 service and its network experience to acquire the comprehensive 275 information. 277 o Applications require the network telemetry to be elastic in order 278 to efficiently use the network resource and reduce the performance 279 impact. The routine network monitoring covers the entire network 280 with low data sampling rate. When issues arise or trends emerge, 281 the telemetry data source can be refocused and the data rate can 282 be boosted. 284 o Efficient data fusion is critical for applications to reduce the 285 overall quantity of data and improve the accuracy of analysis. 287 So far, some telemetry related works have been done within IETF. 288 However, these works are fragmented and scattered in different 289 working groups. The lack of coherence makes it difficult to assemble 290 a comprehensive network telemetry system and causes repetitive and 291 redundant works. 293 A formal network telemetry framework is needed for constructing a 294 working system. The framework should cover the concepts and 295 components from the standardization perspective. This document 296 clarifies the layers on which the telemetry is exerted and decomposes 297 the telemetry system into a set of distinct components that the 298 existing and future works can easily map to. 300 2. Network Telemetry Framework 302 The telemetry can be applied on the data plane, the control plane, 303 and the management plane in a network, as shown in Figure 1. 305 +------------------------------+ 306 | | 307 | OAM Applications | 308 | | 309 +------------------------------+ 310 ^ ^ ^ 311 | | | 312 V | V 313 +-----------|---+--------------+ 314 | | | | 315 | Control Pl|ane| | 316 | Telemetry | <---> | 317 | | | | 318 | ^ V | Management | 319 +------|--------+ Plane | 320 | V | Telemetry | 321 | | | 322 | Data Plane <---> | 323 | Telemetry | | 324 | | | 325 +---------------+--------------+ 327 Figure 1: Layer Category of the Network Telemetry Framework 329 Note that the interaction with OAM applications can be indirect. For 330 example, in the management plane telemetry, the management plane may 331 need to acquire data from the data plane. On the other hand, an OAM 332 application may involve more than one plane simultaneously. For 333 example, an SLA compliance application may require both the data 334 plane telemetry and the control plane telemetry. 336 At each plane, the telemetry can be further partitioned into five 337 distinct components: 339 Data Source: Determine where the original data is acquired. The 340 data source usually just provide raw data which needs further 341 processing. A data source can be considered a probe. A probe can 342 be statically installed or dynamically installed. 344 Data Subscription: Determine the protocol and channel for 345 applications to acquire desired data. Data subscription is also 346 responsible to define the desired data that might not directly 347 available form data sources. The subscribe data can be described 348 by a model. The model can be statically installed or dynamically 349 installed. 351 Data Generation: The original data needs to be processed, encoded, 352 and formatted in network devices to meet application subscription 353 requirements. This may involve in-network computing and 354 processing on either the fast path or the slow path in network 355 devices. 357 Data Export: Determine how the ready data are delivered to 358 applications. 360 Data Analysis: In this final step, data is consumed by applications. 361 Data analysis can be interactive. It may initiate further data 362 subscription. 364 +------------------------------+ 365 | | 366 | Data Analysis | 367 | | 368 +------------------------------+ 369 | ^ 370 | | 371 V | 372 +---------------+--------------+ 373 | | | 374 | Data | Data | 375 | Subscription | Export | 376 | | | 377 +---------------+--------------| 378 | | 379 | Data Generation | 380 | | 381 +------------------------------| 382 | | 383 | Data Source | 384 | | 385 +------------------------------+ 387 Figure 2: Components in the Network Telemetry Framework 389 Since most existing standard-related works belong to the first four 390 components, in the remaining of the document, we focus on these 391 components only. 393 2.1. Existing Works Mapped in the Framework 395 The following table provides a non-exhaustive list of existing works 396 (mainly published in IETF and with the emphasis on the latest new 397 technologies) and shows their positions in the framework. 399 +-----------+--------------+---------------+--------------+ 400 | | Management | Control | Data | 401 | | Plane | Plane | Plane | 402 +-----------+--------------+---------------+--------------+ 403 | | YANG Data | Control Proto.| Flow/Packet | 404 | Data | Store | Network State | Statistics | 405 | Source | | | States | 406 | | | | | 407 +-----------+--------------+---------------+--------------+ 408 | | gPRC | NETCONF/YANG | NETCONF/YANG | 409 | Data | YANG PUSH | BGP | YANG FSM | 410 | Subscribe | | | | 411 | | | | | 412 +-----------+--------------+---------------+--------------+ 413 | | Soft DNP | Soft DNP | In-situ OAM | 414 | Data | | | IPFPM | 415 | Generation| | | Hard DNP | 416 | | | | | 417 +-----------+--------------+---------------+--------------+ 418 | | gRPC | BMP | IPFIX | 419 | Data | YANG PUSH | | UDP | 420 | Export | UDP | | | 421 | | | | | 422 +-----------+--------------+---------------+--------------+ 424 Figure 3: Existing Works 426 2.2. Management Plane Telemetry 428 2.2.1. Requirements and Challenges 430 The management plane of the network element interacts with the 431 Network Management System (NMS), and provides information such as 432 performance data, network logging data, network warning and defects 433 data, and network statistics and state data. Some legacy protocols 434 are widely used for the management plane, such as SNMP and Syslog, 435 but these protocols do not meet the requirements of the automatic 436 network OAM applications. 438 New management plane telemetry protocols should consider the 439 following requirements: 441 Convenient Data Subscription: An application should have the freedom 442 to choose the data export means such as the data types and the 443 export frequency. 445 Structured Data: For automatic network OAM, machine will replace 446 human for network data comprehension. The schema languages such 447 as YANG can efficiently describe structured data and normalize 448 data encoding and transformation. 450 High Speed Data Transport: In order to retain the information, a 451 server need to send a large amount of data at high frequency. 452 Compact encoding format is needed to compress the data and improve 453 the data transport efficiency. The push mode, by replacing the 454 poll mode, can also reduce the interactions between clients and 455 servers, which help to improve the server's efficiency. 457 2.2.2. Push Extensions for NETCONF 459 NETCONF [RFC6241] is one popular network management protocol, which 460 is also recommended by IETF. Although it can be used for data 461 collection, NETCONF is good at configurations. YANG Push 462 [I-D.ietf-netconf-yang-push] extends NETCONF and enables subscriber 463 applications to request a continuous, customized stream of updates 464 from a YANG datastore. Providing such visibility into changes made 465 upon YANG configuration and operational objects enables new 466 capabilities based on the remote mirroring of configuration and 467 operational state. Moreover, distributed data collection mechanism 468 [I-D.zhou-netconf-multi-stream-originators] via UDP based publication 469 channel [I-D.ietf-netconf-udp-pub-channel] provides enhanced 470 efficiency for the NETCONF based telemetry. 472 2.2.3. gRPC Network Management Interface 474 gRPC Network Management Interface (gNMI) 475 [I-D.openconfig-rtgwg-gnmi-spec] is a network management protocol 476 based on the gRPC [I-D.kumar-rtgwg-grpc-protocol] RPC (Remote 477 Procedure Call) framework. With a single gRPC service definition, 478 both configuration and telemetry can be covered. gRPC is an HTTP/2 479 [RFC7540] based open source micro service communication framework. 480 It provides a number of capabilities that makes it well-suited for 481 network telemetry, including: 483 o Full-duplex streaming transporting model combined with a binary 484 encoding mechanism provided further improved telemetry efficiency. 486 o gRPC provides higher-level features consistency across platforms 487 that common HTTP/2 libraries typically do not. This 488 characteristic is especially valuable for the fact that telemetry 489 data collectors are normally resides on a large variety of 490 platforms. 492 o The build in load balancing and failover mechanism. 494 2.3. Control Plane Telemetry 496 2.3.1. Requirements and Challenges 498 The control plane runs the routing protocol (e.g., BGP, OSPF, and IS- 499 IS) to calculate the routing table for a network device. The control 500 plane telemetry monitors the routing protocols to ensure they are 501 working properly. 503 2.3.2. BGP Monitoring Protocol 505 BGP Monitoring Protocol (BMP) [RFC7854] is used to monitor BGP 506 sessions and intended to provide a convenient interface for obtaining 507 route views. The data is collected from the Adjacency-RIB-In routing 508 tables, which are the pre-policy tables, meaning that the routes in 509 these tables have not been filtered or modified by routing policies. 510 So the monitoring station can receive all routes, not just the active 511 routes. 513 2.4. Data Plane Telemetry 515 2.4.1. Requirements and Challenges 517 An effective data plane telemetry system relies on the data that the 518 network device can expose. The data's quality, quantity, and 519 timeliness must meet some stringent requirements. This raises some 520 challenges to the network data plane devices where the first hand 521 data originate. 523 o A data plane device's main function is user traffic processing and 524 forwarding. While supporting network visibility is important, the 525 telemetry is just an auxiliary function and it should not impede 526 normal traffic processing and forwarding (i.e., the performance is 527 not lowered and the behavior is not altered due to the telemetry 528 functions). 530 o The network OAM applications requires end-to-end visibility from 531 various sources, which results in a huge volume of data. However, 532 the sheer data quantity should not stress the network bandwidth, 533 regardless of the data delivery approach (i.e., through in-band or 534 out-of-band channels). 536 o The data plane devices must provide the data in a timely manner 537 with the minimum possible delay. Long processing, transport, 538 storage, and analysis delay can impact the effectiveness of the 539 control loop and even render the data useless. 541 o The data should be structured and labeled, and easy for 542 applications to parse and consume. At the same time, the data 543 types needed by applications can vary significantly. The data 544 plane devices need to provide enough flexibility and 545 programmability to support the precise data provision for 546 applications. 548 o The data plane telemetry should support incremental deployment and 549 work even though some devices are unaware of the system. This 550 challenge is highly relevant to the standards and legacy networks. 552 2.4.2. Dynamic Network Probe 554 Hardware based Dynamic Network Probe (DNP) [I-D.song-opsawg-dnp4iq] 555 provides a programmable means to customize the data that an 556 application collects from the data plane. A direct benefit of DNP is 557 the reduction of the exported data. A full DNP solution covers 558 several components including data source, data subscription, and data 559 generation. The data subscription needs to define the custom data 560 which can be composed and derived from the raw data sources. The 561 data generation takes advantage of the moderate in-network computing 562 to produce the desired data. 564 While DNP can introduce unforeseeable flexibility to the data plane 565 telemetry, it also faces some challenges. It requires a flexible 566 data plane that can be dynamically reprogrammed at runtime. The 567 programming API is yet to be defined. 569 2.4.3. IP Flow Information Export (IPFIX) protocol 571 Traffic on a network can be seen as a set of flows passing through 572 network elements. IP Flow Information Export (IPFIX) [RFC7011] 573 provides a means of transmitting traffic flow information for 574 administrative or other purposes. A typical IPFIX enabled system 575 includes a pool of Metering Processes collects data packets at one or 576 more Observation Points, optionally filters them and aggregates 577 information about these packets. An Exporter then gathers each of 578 the Observation Points together into an Observation Domain and sends 579 this information via the IPFIX protocol to a Collector. 581 2.4.4. In-Situ OAM 583 Traditional passive and active monitoring and measurement techniques 584 are either inaccurate or resource-consuming. It is preferable to 585 directly acquire data associated with a flow's packets when the 586 packets pass through a network. In-situ OAM (iOAM) 587 [I-D.brockners-inband-oam-requirements], a data generation technique, 588 embeds a new instruction header to user packets and the instruction 589 directs the network nodes to add the requested data to the packets. 590 Thus, at the path end the packet's experience on the entire 591 forwarding path can be collected. Such firsthand data is invaluable 592 to many network OAM applications. 594 However, iOAM also faces some challenges. The issues on performance 595 impact, security, scalability and overhead limits, encapsulation 596 difficulties in some protocols, and cross-domain deployment need to 597 be addressed. 599 3. Security Considerations 601 TBD 603 4. IANA Considerations 605 This document includes no request to IANA. 607 5. Contributors 609 The other contributors of this document are listed as follows. 611 o Yunan Gu, Huawei 613 6. Acknowledgments 615 TBD. 617 7. References 619 7.1. Normative References 621 [RFC1157] Case, J., Fedor, M., Schoffstall, M., and J. Davin, 622 "Simple Network Management Protocol (SNMP)", RFC 1157, 623 DOI 10.17487/RFC1157, May 1990, 624 . 626 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 627 Requirement Levels", BCP 14, RFC 2119, 628 DOI 10.17487/RFC2119, March 1997, 629 . 631 [RFC6241] Enns, R., Ed., Bjorklund, M., Ed., Schoenwaelder, J., Ed., 632 and A. Bierman, Ed., "Network Configuration Protocol 633 (NETCONF)", RFC 6241, DOI 10.17487/RFC6241, June 2011, 634 . 636 [RFC7011] Claise, B., Ed., Trammell, B., Ed., and P. Aitken, 637 "Specification of the IP Flow Information Export (IPFIX) 638 Protocol for the Exchange of Flow Information", STD 77, 639 RFC 7011, DOI 10.17487/RFC7011, September 2013, 640 . 642 [RFC7540] Belshe, M., Peon, R., and M. Thomson, Ed., "Hypertext 643 Transfer Protocol Version 2 (HTTP/2)", RFC 7540, 644 DOI 10.17487/RFC7540, May 2015, 645 . 647 [RFC7854] Scudder, J., Ed., Fernando, R., and S. Stuart, "BGP 648 Monitoring Protocol (BMP)", RFC 7854, 649 DOI 10.17487/RFC7854, June 2016, 650 . 652 7.2. Informative References 654 [I-D.brockners-inband-oam-requirements] 655 Brockners, F., Bhandari, S., Dara, S., Pignataro, C., 656 Gredler, H., Leddy, J., Youell, S., Mozes, D., Mizrahi, 657 T., <>, P., and r. remy@barefootnetworks.com, 658 "Requirements for In-situ OAM", draft-brockners-inband- 659 oam-requirements-03 (work in progress), March 2017. 661 [I-D.ietf-netconf-udp-pub-channel] 662 Zheng, G., Zhou, T., and A. Clemm, "UDP based Publication 663 Channel for Streaming Telemetry", draft-ietf-netconf-udp- 664 pub-channel-01 (work in progress), November 2017. 666 [I-D.ietf-netconf-yang-push] 667 Clemm, A., Voit, E., Prieto, A., Tripathy, A., Nilsen- 668 Nygaard, E., Bierman, A., and B. Lengyel, "YANG Datastore 669 Subscription", draft-ietf-netconf-yang-push-14 (work in 670 progress), February 2018. 672 [I-D.kumar-rtgwg-grpc-protocol] 673 Kumar, A., Kolhe, J., Ghemawat, S., and L. Ryan, "gRPC 674 Protocol", draft-kumar-rtgwg-grpc-protocol-00 (work in 675 progress), July 2016. 677 [I-D.openconfig-rtgwg-gnmi-spec] 678 Shakir, R., Shaikh, A., Borman, P., Hines, M., Lebsack, 679 C., and C. Morrow, "gRPC Network Management Interface 680 (gNMI)", draft-openconfig-rtgwg-gnmi-spec-00 (work in 681 progress), March 2017. 683 [I-D.song-opsawg-dnp4iq] 684 Song, H. and J. Gong, "Requirements for Interactive Query 685 with Dynamic Network Probes", draft-song-opsawg-dnp4iq-01 686 (work in progress), June 2017. 688 [I-D.zhou-netconf-multi-stream-originators] 689 Zhou, T., Zheng, G., Voit, E., Clemm, A., and A. Bierman, 690 "Subscription to Multiple Stream Originators", draft-zhou- 691 netconf-multi-stream-originators-01 (work in progress), 692 November 2017. 694 Authors' Addresses 696 Haoyu Song (editor) 697 Huawei 698 2330 Central Expressway 699 Santa Clara 700 USA 702 Email: haoyu.song@huawei.com 704 Tianran Zhou 705 Huawei 706 156 Beiqing Road 707 Beijing, 100095 708 P.R. China 710 Email: zhoutianran@huawei.com 712 Zhenbin Li 713 Huawei 714 156 Beiqing Road 715 Beijing, 100095 716 P.R. China 718 Email: lizhenbin@huawei.com