idnits 2.17.1 draft-ietf-roll-dao-projection-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 : ---------------------------------------------------------------------------- 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 (December 07, 2016) is 2669 days in the past. Is this intentional? Checking references for intended status: Proposed Standard ---------------------------------------------------------------------------- (See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) ** Downref: Normative reference to an Informational RFC: RFC 7102 == Outdated reference: A later version (-30) exists of draft-ietf-6tisch-architecture-10 Summary: 1 error (**), 0 flaws (~~), 2 warnings (==), 1 comment (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 ROLL P. Thubert, Ed. 3 Internet-Draft J. Pylakutty 4 Intended status: Standards Track Cisco 5 Expires: June 10, 2017 December 07, 2016 7 Root initiated routing state in RPL 8 draft-ietf-roll-dao-projection-00 10 Abstract 12 This document proposes a protocol extension to RPL that enables to 13 install a limited amount of centrally-computed routes in a RPL graph, 14 enabling loose source routing down a non-storing mode DODAG, or 15 transversal routes inside the DODAG. As opposed to the classical 16 route injection by DAO messages, this draft projects the routes from 17 the root of the DODAG. 19 Status of This Memo 21 This Internet-Draft is submitted in full conformance with the 22 provisions of BCP 78 and BCP 79. 24 Internet-Drafts are working documents of the Internet Engineering 25 Task Force (IETF). Note that other groups may also distribute 26 working documents as Internet-Drafts. The list of current Internet- 27 Drafts is at http://datatracker.ietf.org/drafts/current/. 29 Internet-Drafts are draft documents valid for a maximum of six months 30 and may be updated, replaced, or obsoleted by other documents at any 31 time. It is inappropriate to use Internet-Drafts as reference 32 material or to cite them other than as "work in progress." 34 This Internet-Draft will expire on April 27, 2017. 36 Copyright Notice 38 Copyright (c) 2016 IETF Trust and the persons identified as the 39 document authors. All rights reserved. 41 This document is subject to BCP 78 and the IETF Trust's Legal 42 Provisions Relating to IETF Documents 43 (http://trustee.ietf.org/license-info) in effect on the date of 44 publication of this document. Please review these documents 45 carefully, as they describe your rights and restrictions with respect 46 to this document. Code Components extracted from this document must 47 include Simplified BSD License text as described in Section 4.e of 48 the Trust Legal Provisions and are provided without warranty as 49 described in the Simplified BSD License. 51 Table of Contents 53 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 2 54 2. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . 3 55 3. New RPL Control Message Options . . . . . . . . . . . . . . . 4 56 3.1. Via Information . . . . . . . . . . . . . . . . . . . . . 4 57 4. Loose Source Routing in Non-storing Mode . . . . . . . . . . 5 58 5. Centralized Computation of Optimized Peer-to-Peer Routes . . 9 59 6. Security Considerations . . . . . . . . . . . . . . . . . . . 12 60 7. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 12 61 8. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 12 62 9. References . . . . . . . . . . . . . . . . . . . . . . . . . 12 63 9.1. Normative References . . . . . . . . . . . . . . . . . . 13 64 9.2. Informative References . . . . . . . . . . . . . . . . . 13 65 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 14 67 1. Introduction 69 The Routing Protocol for Low Power and Lossy Networks [RFC6550] (LLN) 70 (RPL) specification defines a generic Distance Vector protocol that 71 is designed for very low energy consumption and adapted to a variety 72 of LLNs. RPL forms Destination Oriented Directed Acyclic Graphs 73 (DODAGs) which root often acts as the Border Router to connect the 74 RPL domain to the Internet. The root is responsible to select the 75 RPL Instance that is used to forward a packet coming from the 76 Internet into the RPL domain and set the related RPL information in 77 the packets. 79 In the non-storing mode (NSM) of operation (MOP), the root also 80 computes routes down the DODAG towards the end device and leverages 81 source routing to get there, while the default route via the root is 82 used for routing upwards within the LLN and to the Internet at large. 83 NSM is the dominant MOP because because networks may get arbitrary 84 large and in Storing Mode, the amount of memory in nodes close to the 85 root may unexpectedly require memory beyond a node's capabilities. 87 But as a network gets deep, the size of the source routing header 88 that the root must add to all the downward packets may also become an 89 issue for far away target devices. In some use cases, a RPL network 90 forms long lines and a limited amount of well-targeted routing state 91 would allow to make the source routing operation loose as opposed to 92 strict, and save packet size. Limiting the packet size is directly 93 beneficial to the energy budget, but, mostly, it reduces the chances 94 of frame loss and/or packet fragmentation, which is highly 95 detrimental to the LLN operation. Because the capability to store a 96 routing state in every node is limited, the decision of which route 97 is installed where can only be optimized with a global knowledge of 98 the system, a knowledge that the root has in non-storing mode. 100 Additionally, RPL storing mode is optimized or Point-to-Multipoint 101 (P2MP), root to leaves and Multipoint-to-Point (MP2P) leaves to root 102 operations, whereby routes are always installed along the RPL DODAG. 103 Transversal Peer to Peer (P2P) routes in a RPL network will generally 104 suffer from some stretch since routing between 2 peers always happens 105 via a common parent. In NSM, all peer-to-peer routes travel all the 106 way to the root, which adds a source routing header and forwards the 107 packet down to the destination, resulting in the longest stretch and 108 overload of the radio bandwidth near the root. A controller, for 109 instance collocated with the RPL root, with enough topological 110 awareness of the connectivity between nodes, would be able to compute 111 more direct routes, avoiding the vicinity of the root whenever 112 possible. 114 The 6TiSCH architecture [I-D.ietf-6tisch-architecture] leverages the 115 Deterministic Networking Architecture [I-D.finn-detnet-architecture] 116 as one possible model whereby the device resources and capabilities 117 are exposed to an external controller which installs routing states 118 into the network based on some objective functions that reside in 119 that external entity. 121 Based on heuristics of usage, path length, and knowledge of device 122 capacity and available resources such as battery levels and 123 reservable buffers, a Path Computation Element ([PCE]) with a global 124 visibility on the system could install additional P2P routes that are 125 more optimized for the current needs as expressed by the objective 126 function. 128 This draft enables a RPL root, with optionally the assistance of a 129 PCE, to install and maintain additional storing mode routes within 130 the RPL domain, along a selected set of nodes and for a selected 131 duration, thus providing routes from suitable than those obtained 132 from the distributed operation of RPL in either storing and non- 133 storing modes. 135 2. Terminology 137 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 138 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 139 document are to be interpreted as described in [RFC2119]. 141 The Terminology used in this document is consistent with and 142 incorporates that described in `Terminology in Low power And Lossy 143 Networks' [RFC7102] and [RFC6550]. 145 3. New RPL Control Message Options 147 Section 6.7 of [RFC6550] specifies Control Message Options (CMO) to 148 be placed in RPL messages such as the DAO message. The RPL Target 149 Option and the Transit Information Option (TIO) are such options; the 150 former indicates a node to be reached and the latter specifies a 151 parent that can be used to reach that node. Options may be 152 factorized; one or more contiguous TIOs apply to the one or more 153 contiguous Target options that immediately precede the TIOs in the 154 RPL message. 156 This specification introduces a new Control Message Option, the Via 157 Information option (VIO). Like the TIO, the VIO MUST be preceded by 158 one or more RPL Target options to which it applies. Unlike the TIO, 159 the VIO are not factorized: multiple contiguous Via options indicate 160 an ordered sequence of hops to reach the target(s), presented in the 161 same order as they would appear in a routing header. 163 3.1. Via Information 165 The Via Information option MAY be present in DAO messages, and its 166 format is as follows: 168 0 1 2 3 169 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 170 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 171 | Type = 0x0A | Option Length | Path Sequence | Path Lifetime | 172 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 173 | | 174 + + 175 . . 176 . Next-Hop Address . 177 . . 178 + + 179 | | 180 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 182 Figure 1: Eliding the RPLInstanceID 184 Option Type: 0x0A (to be confirmed by IANA) 186 Option Length: Variable, depending on whether or not Parent Address 187 is present. 189 Path Sequence: 8-bit unsigned integer. When a RPL Target option is 190 issued by the root of the DODAG (i.e. in a DAO message), that 191 root sets the Path Sequence and increments the Path Sequence 192 each time it issues a RPL Target option with updated 193 information. The indicated sequence deprecates any state for a 194 given Target that was learned from a previous sequence and adds 195 to any state that was learned for that sequence. 197 Path Lifetime: 8-bit unsigned integer. The length of time in 198 Lifetime Units (obtained from the Configuration option) that 199 the prefix is valid for route determination. The period starts 200 when a new Path Sequence is seen. A value of all one bits 201 (0xFF) represents infinity. A value of all zero bits (0x00) 202 indicates a loss of reachability. A DAO message that contains 203 a Via Information option with a Path Lifetime of 0x00 for a 204 Target is referred as a No-Path (for that Target) in this 205 document. 207 Next-Hop Address: 8 or 16 bytes. IPv6 Address of the next hop 208 towards the destination(s) indicated in the target option that 209 immediately precede the VIO. The /64 prefix can be elided if 210 it is the same as that of (all of) the target(s). In that 211 case, the Next-Hop Address is expressed as the 8-bytes suffix 212 only, otherwise it is expressed as 16 bytes. 214 4. Loose Source Routing in Non-storing Mode 216 A classical RPL implementation in a very constrained LLN uses the 217 non-storing mode of operation whereby a RPL node indicates a parent- 218 child relationship to the root, using a Destination Advertisement 219 Object (DAO) that is unicast from the node directly to the root, and 220 the root builds a path to a destination down the DODAG by 221 concatenating this information. 223 ------+--------- 224 | Internet 225 | 226 +-----+ 227 | | Border Router 228 | | (RPL Root) 229 +-----+ ^ | | 230 | | DAO | ACK | 231 o o o o | | | Strict 232 o o o o o o o o o | | | Source 233 o o o o o o o o o o | | | Route 234 o o o o o o o o o | | | 235 o o o o o o o o | v v 236 o o o o 237 LLN 239 Figure 2: RPL non-storing operation 241 Nodes are not expected to store downward routing state via their 242 children, and the routing operates in strict source routing mode as 243 detailed in An IPv6 Routing Header for Source Routes with RPL 244 [RFC6554] 246 This draft proposes an addition whereby the root projects a route 247 through an extended DAO to an arbitrary node down the DODAG, 248 indicating a child or a direct sequence of children via which a 249 certain destination (target) may be reached. The root is expected to 250 use the mechanism optimally and with required parsimony to fit within 251 the device resources, but how the root figures the amount of 252 resources that are available is out of scope. 254 ------+--------- 255 | Internet 256 | 257 +-----+ 258 | | Border Router 259 | | (RPL Root) 260 +-----+ | ^ | 261 | | DAO | ACK | 262 o o o o | | | Loose 263 o o o o o o o o o | ^ | Source 264 o o o o o o o o o o | | DAO | Route 265 o o o o o o o o o | ^ | 266 o o o o o o o o v | DAO v 267 o o o o 268 LLN 270 Figure 2: Non-Storing with Projected routes 272 When a RPL domain operates in non-storing Mode of Operation (NS-MOP), 273 only the root possesses routing information about the whole network. 274 A packet that is generated within the domain first reaches the root, 275 which can then apply a source routing information to reach the 276 destination. Similarly, a packet coming from the outside of the 277 domain for a destination that is expected to be in a RPL domain 278 reaches the root. 280 In NS-MOP, the root, or some associated centralized computation 281 engine, can thus determine the amount of packets that reach a 282 destination in the RPL domain, and thus the amount of energy and 283 bandwidth that is wasted for transmission, between itself and the 284 destination, as well as the risk of fragmentation, any potential 285 delays because of a paths longer than necessary (shorter paths exist 286 that would not traverse the root). 288 Additionally, the DAG root knows the whole DAG topology, so when the 289 source of a packet is also in the RPL domain, the root can determine 290 the common parent that would have been used in storing mode, and thus 291 the list of nodes in the path between the common parent and the 292 destination. For instance in the below diagram, if the source is 41 293 and the destination 52, the common parent is the node 22. 295 ------+--------- 296 | Internet 297 | 298 +-----+ 299 | | Border Router 300 | | (RPL Root) 301 +-----+ 302 | \ \____ 303 / \ \ 304 o 11 o 12 o 13 305 / | / \ 306 o 22 o 23 o 24 o 25 307 / \ | \ \ 308 o 31 o 32 o o o 35 309 / / | \ | \ 310 o 41 o 42 o o o 45 o 46 311 | | | | \ | 312 o 51 o 52 o 53 o o 55 o 56 313 LLN 315 Figure 3: Non-Storing with Projected routes 317 With this draft, the root can install routing states along a segment 318 that is either itself to the destination, or from one or more common 319 parents for a particular source/destination pair towards that 320 destination (in our example, this would be the segment made of nodes 321 22, 32, 42). 323 The draft expects that the root has enough information about the 324 capability for each node to store a number of routes, which can be 325 discovered for instance using a Network Management System (NMS) and/ 326 or the RPL routing extensions specified in Routing for Path 327 Calculation in LLNs [RFC6551]. Based on that information, the root 328 computes which segment should be routed and which relevant state 329 should be installed in which nodes. The algorithm is out of scope 330 but it is envisaged that the root could compute the ratio between the 331 optimal path (existing path not traversing the root, and the current 332 path), the application SLA for specific flows that could benefit from 333 shorter paths, the energy wasted in the network, local congestion on 334 various links that would benefit from having flows routed along other 335 paths. 337 This draft introduces a new mode of operation for loose source 338 routing in the LLN, the Non-Storing with Projected routes MOP. With 339 this new MOP, the root sends a unicast DAO message to the last node 340 of the routing segment that must be installed. The DAO message 341 contains the ordered list of hops along the segment as a list of Via 342 Information options that are preceded by one or more RPL Target 343 options to which they relate. Each Via Information option contains a 344 lifetime for which state is to be maintained. 346 The root sends the DAO directly to the last node in the segment, 347 which is expected to be able to route to the targets on its own. 349 The last node in the segment may have another information to reach 350 the target(s), such as a connected route or an already installed 351 projected route. If it does not have such a route then the node 352 should lookup the address on the relevant interfaces. If one of the 353 targets cannot be located, the node MUST answer to the root with a 354 negative DAO-ACK listing the target(s) that could not be located 355 (suggested status 10), and continue the process for those targets 356 that could be located if any. 358 For the targets that could be located, last node in the segment 359 generates a DAO to its loose predecessor in the segment as indicated 360 in the list of Via Information options. 362 The node strips the last Via Information option which corresponds to 363 self, and uses it as source address for the DAO to the predecessor. 364 The address of the predecessor to be used as destination for the DAO 365 message is found in the now last Via Information option. The 366 predecessor is expected to have a route to the address used as 367 source, either connected, installed previously as another DAO, or 368 from other means. 370 The predecessor is expected to have a route to the address used as 371 source and that is his successor. If it does not and cannot locate 372 the successor, the predecessor node MUST answer to the root with a 373 negative DAO-ACK indicating the successor that could not be located. 374 The DAO-ACK contains the list of targets that could not be routed to 375 (suggested status 11). 377 If the predecessor can route to the successor node, then it installs 378 a route to the targets via the successor. If that route is not 379 connected then a recursive lookup will take place to reach the 380 target(s). From there, the node strips the last Via Information 381 option and either answers to the root with a positive DAO-ACK that 382 contains the list of targets that could be routed to, or propagates 383 the DAO to its own predecessor. 385 A NULL lifetime in the Via Information option along the segment is 386 used to clean up the state. 388 In the example below, say that there is a lot of traffic to nodes 55 389 and 56 and the root decides to reduce the size of routing headers to 390 those destinations. The root can first send a DAO to node 45 391 indicating target 55 and a Via segment (35, 45), as well as another 392 DAO to node 46 indicating target 56 and a Via segment (35, 46). This 393 will save one entry in the routing header on both sides. The root 394 may then send a DAO to node 35 indicating targets 55 and 56 a Via 395 segment (13, 24, 35) to fully optimize that path. 397 Alternatively, the root may send a DAO to node 45 indicating target 398 55 and a Via segment (13, 24, 35, 45) and then a DAO to node 46 399 indicating target 56 and a Via segment (13, 24, 35, 46), indicating 400 the same DAO Sequence. 402 5. Centralized Computation of Optimized Peer-to-Peer Routes 404 With the initial specifications of RPL [RFC6550], the P2P path from a 405 source to a destination is often stretched, as illustrated in 406 [RFC6550]: 408 - in non-storing mode, all packets routed within the DODAG flow 409 all the way up to the root of the DODAG. If the destination is in 410 the same DODAG, the root must encapsulate the packet to place a 411 Routing Header that has the strict source route information down 412 the DODAG to the destination. This will be the case even if the 413 destination is relatively close to the source and the root is 414 relatively far off. 416 - in storing mode, unless the destination is a child of the 417 source, the packets will follow the default route up the DODAG as 418 well. If the destination is in the same DODAG, they will 419 eventually reach a common parent that has a DAO route to the 420 destination; at worse, the common parent may also be the root. 421 From that common parent, the packet will follow a path down the 422 DODAG that is optimized for the Objective Function that was used 423 to build the DODAG. 425 It results that it is often beneficial to enable additional P2P 426 routes, either if the RPL route present a stretch from shortest path, 427 or if the new route is engineered with a different objective. 429 ------+--------- 430 | Internet 431 | 432 +-----+ 433 | | Border Router 434 | | (RPL Root) 435 +-----+ 436 X 437 ^ v o o 438 ^ o o v o o o o o 439 ^ o o o v o o o o o 440 ^ o o v o o o o o 441 S o o o D o o o 442 o o o o 443 LLN 445 Figure 4: Routing Stretch 447 For that reason, earlier work at the IETF introduced the Reactive 448 Discovery of Point-to-Point Routes in Low Power and Lossy Networks 449 [RFC6997], which specifies a distributed method for establishing 450 optimized P2P routes. This draft proposes an alternate based on a 451 centralized route computation. 453 It must be noted that RPL has a concept of instance but does not have 454 a concept of an administrative distance, which exists in certain 455 proprietary implementations to sort out conflicts between multiple 456 sources. This draft conforms the instance model as follows: 458 - if the PCE needs to influence a particular instance to add 459 better routes in conformance with the routing objectives in that 460 instance, it may do so. When the PCE modifies an existing 461 instance then the added routes must not create a loop in that 462 instance. This is achieved by always preferring a route obtained 463 from the PCE over a route that is learned via RPL. 465 - If the PCE installs a more specific (Traffic Engineering) route 466 between a particular pair of nodes then it should use a Local 467 Instance from the ingress node of that path. Only packets 468 associated with that instance will be routed along that path. 470 In all cases, the path is indicated by VIA options, and the flow is 471 similar to the flow used to obtain loose source routing. 473 The root sends the DAO with the target option and the Via Option to 474 the lest router in the path; the last router removes the last Via 475 Option and passes the DAO to the previous hop. 477 ------+--------- 478 | Internet 479 | 480 +-----+ 481 | | Border Router 482 | | (RPL Root) 483 +-----+ 484 | Projected DAO message to C 485 o | o o 486 o o o | o o o o o 487 o o o | o o o o o o 488 o o V o o o o o o 489 S A B C D o o o 490 o o o o 491 LLN 493 Figure 5: Projected DAO from root 495 The process recurses till the destination which sends a DAO-ACK to 496 the root. In the example above, for target D, the list of via 497 options is S, A, B and C. The projected DAO is sent by the root to 499 ------+--------- 500 | Internet 501 | 502 +-----+ 503 | | Border Router 504 | | (RPL Root) 505 +-----+ 506 ^ Projected DAO-ACK from S 507 / o o o 508 / o o o o o o o 509 | o o o o o o o o o 510 | o o o o o o o o 511 S A B C D o o o 512 o o o o 513 LLN 515 Figure 6: Projected DAO-ACK to root 517 The process recurses till the destination which sends a DAO-ACK to 518 the root. In the example above, for target D, the list of via 519 options is S, A, B and C. The projected DAO is sent by the root to 520 ------+--------- 521 | Internet 522 | 523 +-----+ 524 | | Border Router 525 | | (RPL Root) 526 +-----+ 527 | 528 o o o o 529 o o o o o o o o o 530 o o o o o o o o o o 531 o o o o o o o o o 532 S>>A>>>B>>C>>>D o o o 533 o o o o 534 LLN 536 Figure 7: Projected Transversal Route 538 6. Security Considerations 540 This draft uses messages that are already present in [RFC6550] with 541 optional secured versions. The same secured versions may be used 542 with this draft, and whatever security is deployed for a given 543 network also applies to the flows in this draft. 545 7. IANA Considerations 547 This document updates the IANA registry for the Mode of Operation 548 (MOP) 550 4: Non-Storing with Projected routes [this] 552 This document updates IANA registry for the RPL Control Message 553 Options 555 0x0A: Via descriptor [this] 557 8. Acknowledgments 559 The authors wish to acknowledge JP Vasseur and Patrick Wetterwald for 560 their contributions to the ideas developed here. 562 9. References 563 9.1. Normative References 565 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 566 Requirement Levels", BCP 14, RFC 2119, 567 DOI 10.17487/RFC2119, March 1997, 568 . 570 [RFC6550] Winter, T., Ed., Thubert, P., Ed., Brandt, A., Hui, J., 571 Kelsey, R., Levis, P., Pister, K., Struik, R., Vasseur, 572 JP., and R. Alexander, "RPL: IPv6 Routing Protocol for 573 Low-Power and Lossy Networks", RFC 6550, 574 DOI 10.17487/RFC6550, March 2012, 575 . 577 [RFC6551] Vasseur, JP., Ed., Kim, M., Ed., Pister, K., Dejean, N., 578 and D. Barthel, "Routing Metrics Used for Path Calculation 579 in Low-Power and Lossy Networks", RFC 6551, 580 DOI 10.17487/RFC6551, March 2012, 581 . 583 [RFC6554] Hui, J., Vasseur, JP., Culler, D., and V. Manral, "An IPv6 584 Routing Header for Source Routes with the Routing Protocol 585 for Low-Power and Lossy Networks (RPL)", RFC 6554, 586 DOI 10.17487/RFC6554, March 2012, 587 . 589 [RFC7102] Vasseur, JP., "Terms Used in Routing for Low-Power and 590 Lossy Networks", RFC 7102, DOI 10.17487/RFC7102, January 591 2014, . 593 9.2. Informative References 595 [I-D.finn-detnet-architecture] 596 Finn, N. and P. Thubert, "Deterministic Networking 597 Architecture", draft-finn-detnet-architecture-08 (work in 598 progress), August 2016. 600 [I-D.ietf-6tisch-architecture] 601 Thubert, P., "An Architecture for IPv6 over the TSCH mode 602 of IEEE 802.15.4", draft-ietf-6tisch-architecture-10 (work 603 in progress), June 2016. 605 [PCE] IETF, "Path Computation Element", 606 . 608 [RFC6997] Goyal, M., Ed., Baccelli, E., Philipp, M., Brandt, A., and 609 J. Martocci, "Reactive Discovery of Point-to-Point Routes 610 in Low-Power and Lossy Networks", RFC 6997, 611 DOI 10.17487/RFC6997, August 2013, 612 . 614 Authors' Addresses 616 Pascal Thubert (editor) 617 Cisco Systems 618 Village d'Entreprises Green Side 619 400, Avenue de Roumanille 620 Batiment T3 621 Biot - Sophia Antipolis 06410 622 FRANCE 624 Phone: +33 4 97 23 26 34 625 Email: pthubert@cisco.com 627 James Pylakutty 628 Cisco Systems 629 Cessna Business Park 630 Kadubeesanahalli 631 Marathalli ORR 632 Bangalore, Karnataka 560087 633 INDIA 635 Phone: +91 80 4426 4140 636 Email: mundenma@cisco.com