idnits 2.17.00 (12 Aug 2021) /tmp/idnits53953/draft-thubert-6lowpan-simple-fragment-recovery-04.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- ** The document seems to lack a License Notice according IETF Trust Provisions of 28 Dec 2009, Section 6.b.ii or Provisions of 12 Sep 2009 Section 6.b -- however, there's a paragraph with a matching beginning. Boilerplate error? (You're using the IETF Trust Provisions' Section 6.b License Notice from 12 Feb 2009 rather than one of the newer Notices. See https://trustee.ietf.org/license-info/.) 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 seems to lack a disclaimer for pre-RFC5378 work, but may have content which was first submitted before 10 November 2008. If you have contacted all the original authors and they are all willing to grant the BCP78 rights to the IETF Trust, then this is fine, and you can ignore this comment. If not, you may need to add the pre-RFC5378 disclaimer. (See the Legal Provisions document at https://trustee.ietf.org/license-info for more information.) -- The document date (March 28, 2009) is 4802 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) == Unused Reference: 'I-D.mathis-frag-harmful' is defined on line 468, but no explicit reference was found in the text == Unused Reference: 'RFC2309' is defined on line 476, but no explicit reference was found in the text == Unused Reference: 'RFC2581' is defined on line 483, but no explicit reference was found in the text == Unused Reference: 'RFC2914' is defined on line 486, but no explicit reference was found in the text == Unused Reference: 'RFC3168' is defined on line 489, but no explicit reference was found in the text ** Obsolete normative reference: RFC 2988 (Obsoleted by RFC 6298) == Outdated reference: draft-ietf-tsvwg-udp-guidelines has been published as RFC 5405 -- Obsolete informational reference (is this intentional?): RFC 2309 (Obsoleted by RFC 7567) -- Obsolete informational reference (is this intentional?): RFC 2581 (Obsoleted by RFC 5681) Summary: 2 errors (**), 0 flaws (~~), 7 warnings (==), 4 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 6LoWPAN P. Thubert, Ed. 3 Internet-Draft Cisco 4 Intended status: Standards Track J. Hui 5 Expires: September 29, 2009 Arch Rock Corporation 6 March 28, 2009 8 LoWPAN simple fragment Recovery 9 draft-thubert-6lowpan-simple-fragment-recovery-04 11 Status of this Memo 13 This Internet-Draft is submitted to IETF in full conformance with the 14 provisions of BCP 78 and BCP 79. 16 Internet-Drafts are working documents of the Internet Engineering 17 Task Force (IETF), its areas, and its working groups. Note that 18 other groups may also distribute working documents as Internet- 19 Drafts. 21 Internet-Drafts are draft documents valid for a maximum of six months 22 and may be updated, replaced, or obsoleted by other documents at any 23 time. It is inappropriate to use Internet-Drafts as reference 24 material or to cite them other than as "work in progress." 26 The list of current Internet-Drafts can be accessed at 27 http://www.ietf.org/ietf/1id-abstracts.txt. 29 The list of Internet-Draft Shadow Directories can be accessed at 30 http://www.ietf.org/shadow.html. 32 This Internet-Draft will expire on September 29, 2009. 34 Copyright Notice 36 Copyright (c) 2009 IETF Trust and the persons identified as the 37 document authors. All rights reserved. 39 This document is subject to BCP 78 and the IETF Trust's Legal 40 Provisions Relating to IETF Documents in effect on the date of 41 publication of this document (http://trustee.ietf.org/license-info). 42 Please review these documents carefully, as they describe your rights 43 and restrictions with respect to this document. 45 Abstract 47 Considering that 6LoWPAN packets can be as large as 2K bytes and that 48 an 802.15.4 frame with security will carry in the order of 80 bytes 49 of effective payload, a packet might end up fragmented into as many 50 as 25 fragments at the 6LoWPAN shim layer. If a single one of those 51 fragments is lost in transmission, all fragments must be resent, 52 further contributing to the congestion that might have caused the 53 initial packet loss. This draft introduces a simple protocol to 54 recover individual fragments that might be lost over multiple hops 55 between 6LoWPAN endpoints. 57 Table of Contents 59 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 3 60 2. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . 4 61 3. Rationale . . . . . . . . . . . . . . . . . . . . . . . . . . 4 62 4. Requirements . . . . . . . . . . . . . . . . . . . . . . . . . 6 63 5. Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 64 6. New Dispatch types and headers . . . . . . . . . . . . . . . . 7 65 6.1. Recoverable Fragment Dispatch type and Header . . . . . . 8 66 6.2. Fragment Acknowledgement Dispatch type and Header . . . . 8 67 7. Fragments Recovery . . . . . . . . . . . . . . . . . . . . . . 9 68 8. Security Considerations . . . . . . . . . . . . . . . . . . . 10 69 9. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 10 70 10. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 10 71 11. References . . . . . . . . . . . . . . . . . . . . . . . . . . 11 72 11.1. Normative References . . . . . . . . . . . . . . . . . . . 11 73 11.2. Informative References . . . . . . . . . . . . . . . . . . 11 74 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . . 12 76 1. Introduction 78 In many 6LoWPAN applications, the majority of traffic is spent 79 sending small chunks of data (order few bytes to few tens of bytes) 80 per packet. Given that an 802.15.4 frame can carry on the order of 81 80 bytes in the worst case, fragmentation is often not needed for 82 most application traffic. However, many applications also require 83 occasional bulk data transfer capabilities to support firmware 84 upgrades of 6LoWPAN devices or extraction of logs from 6LoWPAN 85 devices. In the former case, bulk data is transferred to the 6LoWPAN 86 device, and in the latter, bulk data flows away from the 6LoWPAN 87 device. In both cases, the bulk data size is often on the order of 88 10K bytes or more and end-to-end reliable transport is required. 90 Mechanisms such as TCP or application-layer segmentation will be used 91 to support end-to-end reliable transport. One option to support bulk 92 data transfer over 6LoWPAN links is to set the Maximum Segment Size 93 to fit within the 802.15.4 MTU. Doing so, however, can add 94 significant header overhead to each 802.15.4 frame. This causes the 95 end-to-end transport to be aware of the delivery properties of 96 6LoWPAN networks, which is a layer violation. 98 An alternative mechanism combines the use of 6LoWPAN fragmentation in 99 addition to transport or application-layer segmentation. Increasing 100 the Maximum Segment Size reduces header overhead by the end-to-end 101 transport protocol. It also encourages the transport protocol to 102 reduce the number of outstanding datagrams, ideally to a single 103 datagram, thus reducing the need to support out-of-order delivery 104 common to 6LoWPAN networks. 106 [RFC4944] defines a datagram fragmentation mechanism for 6LoWPAN 107 networks. However, because [RFC4944] does not define a mechanism for 108 recovering fragments that are lost, datagram forwarding fails if even 109 one fragment is not delivered properly to the next IP hop. End-to- 110 end transport mechanisms will require retransmission of all 111 fragments, wasting resources in an already resource-constrained 112 network. 114 Past experience with fragmentation has shown that missassociated or 115 lost fragments can lead to poor network behavior and, eventually, 116 trouble at application layer. The reader is encouraged to read 117 [RFC4963] and follow the references for more information. That 118 experience led to the definition of the Path MTU discovery [RFC1191] 119 protocol that limits fragmentation over the Internet. 121 For one-hop communications, a number of media propose a local 122 acknowledgement mechanism that is enough to protect the fragments. 123 In a multihop environment, an end-to-end fragment recovery mechanism 124 might be a good complement to a hop-by-hop MAC level recovery. This 125 draft introduces a simple protocol to recover individual fragments 126 between 6LoWPAN endpoints. Specifically in the case of UDP, valuable 127 additional information can be found in UDP Usage Guidelines for 128 Application Designers [I-D.ietf-tsvwg-udp-guidelines]. 130 2. Terminology 132 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 133 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 134 document are to be interpreted as described in [RFC2119]. 136 Readers are expected to be familiar with all the terms and concepts 137 that are discussed in "IPv6 over Low-Power Wireless Personal Area 138 Networks (6LoWPANs): Overview, Assumptions, Problem Statement, and 139 Goals" [RFC4919] and "Transmission of IPv6 Packets over IEEE 802.15.4 140 Networks" [RFC4944]. 142 ERP 144 Error Recovery Procedure. 146 LoWPAN endpoints 148 The LoWPAN nodes in charge of generating or expanding a 6LoWPAN 149 header from/to a full IPv6 packet. The LoWPAN endpoints are the 150 points where fragmentation and reassembly take place. 152 3. Rationale 154 There are a number of usages for large packets in Wireless Sensor 155 Networks. Such usages may not be the most typical or represent the 156 largest amount of traffic over the LoWPAN; however, the associated 157 functionality can be critical enough to justify extra care for 158 ensuring effective transport of large packets across the LoWPAN. 160 The list of those usages includes: 162 Towards the LoWPAN node: 164 Packages of Commands: A number of commands or a full 165 configuration can by packaged as a single message to ensure 166 consistency and enable atomic execution or complete roll back. 167 Until such commands are fully received and interpreted, the 168 intended operation will not take effect. 170 Firmware update: For example, a new version of the LoWPAN node 171 software is downloaded from a system manager over unicast or 172 multicast services. Such a reflashing operation typically 173 involves updating a large number of similar 6LoWPAN nodes over 174 a relatively short period of time. 176 From the LoWPAN node: 178 Waveform captures: A number of consecutive samples are measured 179 at a high rate for a short time and then transferred from a 180 sensor to a gateway or an edge server as a single large report. 182 Data logs: 6LoWPAN nodes may generate large logs of sampled data 183 for later extraction. 6LoWPAN nodes may also generate system 184 logs to assist in diagnosing problems on the node or network. 186 Large data packets: Rich data types might require more than one 187 fragment. 189 Uncontrolled firmware download or waveform upload can easily result 190 in a massive increase of the traffic and saturate the network. 192 When a fragment is lost in transmission, all fragments are resent, 193 further contributing to the congestion that caused the initial loss, 194 and potentially leading to congestion collapse. 196 This saturation may lead to excessive radio interference, or random 197 early discard (leaky bucket) in relaying nodes. Additional queuing 198 and memory congestion may result while waiting for a low power next 199 hop to emerge from its sleeping state. 201 To demonstrate the severity of the problem, consider a fairly 202 reliable 802.15.4 frame delivery rate of 99.9% over a single 802.15.4 203 hop. The expected delivery rate of a 5-fragment datagram would be 204 about 99.5% over a single 802.15.4 hop. However, the expected 205 delivery rate would drop to 95.1% over 10 hops, a reasonable network 206 diameter for 6LoWPAN applications. The expected delivery rate for a 207 1280-byte datagram is 98.4% over a single hop and 85.2% over 10 hops. 209 Considering that 6LoWPAN packets can be as large as 2K bytes and that 210 a 802.15.4 frame with security will carry in the order of 80 bytes of 211 effective payload, a packet might be fragmented into about 25 212 fragments at the 6LoWPAN shim layer. This level of fragmentation is 213 much higher than that traditionally experienced over the Internet 214 with IPv4 fragments. At the same time, the use of radios increases 215 the probability of transmission loss and Mesh-Under techniques 216 compound that risk over multiple hops. 218 4. Requirements 220 This paper proposes a method to recover individual fragments between 221 LoWPAN endpoints. The method is designed to fit the following 222 requirements of a LoWPAN (with or without a Mesh-Under routing 223 protocol): 225 Number of fragments 227 The recovery mechanism must support highly fragmented packets, 228 with a maximum of 32 fragments per packet. 230 Minimum acknowledgement overhead 232 Because the radio is half duplex, and because of silent time spent 233 in the various medium access mechanisms, an acknowledgment 234 consumes roughly as many resources as data fragment. 236 The recovery mechanism should be able to acknowledge multiple 237 fragments in a single message and not require an acknowledgement 238 at all if fragments are already protected at a lower layer. 240 Controlled latency 242 The recovery mechanism must succeed or give up within the time 243 boundary imposed by the recovery process of the Upper Layer 244 Protocols. 246 Support for out-of-order fragment delivery 248 A Mesh-Under load balancing mechanism such as the ISA100 Data Link 249 Layer can introduce out-of-sequence packets. 251 The recovery mechanism must account for packets that appear lost 252 but are actually only delayed over a different path. 254 Optional congestion control 256 The aggregation of multiple concurrent flows may lead to the 257 saturation of the radio network and congestion collapse. 259 The recovery mechanism should provide means for controlling the 260 number of fragments in transit over the LoWPAN. 262 5. Overview 264 Considering that a multi-hop LoWPAN can be a very sensitive 265 environment due to the limited queuing capabilities of a large 266 population of its nodes, this draft recommends a simple and 267 conservative approach to congestion control, based on TCP congestion 268 avoidance. 270 From the standpoint of a source LoWPAN endpoint, an outstanding 271 fragment is a fragment that was sent but for which no explicit 272 acknowledgment was received yet. This means that the fragment might 273 be on the way, received but not yet acknowledged, or the 274 acknowledgment might be on the way back. It is also possible that 275 either the fragment or the acknowledgment was lost on the way. 277 Because a meshed LoWPAN might deliver frames out of order, it is 278 virtually impossible to differentiate these situations. In other 279 words, from the sender standpoint, all outstanding fragments might 280 still be in the network and contribute to its congestion. There is 281 an assumption, though, that after a certain amount of time, a frame 282 is either received or lost, so it is not causing congestion anymore. 283 This amount of time can be estimated based on the round trip delay 284 between the LoWPAN endpoints. The method detailed in [RFC2988] is 285 recommended for that computation. 287 6. New Dispatch types and headers 289 This specification extends "Transmission of IPv6 Packets over IEEE 290 802.15.4 Networks" [RFC4944] with 4 new dispatch types, for 291 Recoverable Fragments (RFRAG) headers with or without Acknowledgment 292 Request, and for the Acknowledgment back. 294 Pattern Header Type 295 +------------+-----------------------------------------------+ 296 | 11 101000 | RFRAG - Recoverable Fragment | 297 | 11 101001 | RFRAG-AR - RFRAG with Ack Request | 298 | 11 10101x | RFRAG-ACK - RFRAG Acknowledgment | 299 +------------+-----------------------------------------------+ 301 Figure 1: Additional Dispatch Value Bit Patterns 303 In the following sections, the semantics of "datagram_tag," 304 "datagram_offset" and "datagram_size" and the reassembly process are 305 unchanged from [RFC4944] Section 5.3. "Fragmentation Type and 306 Header." 308 6.1. Recoverable Fragment Dispatch type and Header 310 1 2 3 311 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 1 312 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 313 |1 1 1 0 1 0 0 X|datagram_offset| datagram_tag | 314 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 315 |Sequence | datagram_size | 316 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 317 X set == Ack Requested 319 Figure 2: Recoverable Fragment Dispatch type and Header 321 X bit 323 When set, the sender requires an Acknowledgment from the receiver 325 Sequence 327 The sequence number of the fragment. Fragments are numbered 328 [0..N] where N is in [0..31]. 330 6.2. Fragment Acknowledgement Dispatch type and Header 332 1 2 3 333 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 1 334 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 335 |1 1 1 0 1 0 1 Y| datagram_tag | 336 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 337 | Acknowledgment Bitmap | 338 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 339 ^ ^ 340 | | Y is reserved 341 | | 342 | | bitmap indicating whether 343 | +-----Fragment with sequence 10 was received 344 +-------------------------Fragment with sequence 00 was received 346 Figure 3: Fragment Acknowledgement Dispatch type and Header 348 Y bit 350 Reserved. 352 Acknowledgement Bitmap 354 Each bit in the Bitmap refers to a particular fragment: bit n set 355 indicates that fragment with sequence n was received, for n in 356 [0..31]. 358 A NULL bitmap (All zeroes) means that the fragment was dropped . 360 7. Fragments Recovery 362 The node that fragments the packets at 6LoWPAN level (the sender) 363 controls the Fragment Acknowledgements. If may do that at any 364 fragment to implement its own policy or perform congestion control 365 which is out of scope for this document. When the sender of the 366 fragment knows that an underlying mechanism protects the Fragments 367 already it MAY refrain from using the Acknowledgement mechanism, and 368 never set the Ack Requested bit. The node that recomposes the 369 packets at 6LoWPAN level (the receiver) MUST acknowledge the 370 fragments it has received when asked to, and MAY slightly defer that 371 acknowledgement. 373 The sender transfers a controlled number of fragments and MAY flag 374 the last fragment of a series with an acknowledgment request. The 375 received MUST acknowledge a fragment with the acknowledgment request 376 bit set. If any fragment immediately preceding an acknowledgment 377 request is still missing, the receiver MAY intentionally delay its 378 acknowledgment to allow in-transit fragments to arrive. delaying the 379 acknowledgement might defeat the round trip delay computation so it 380 should be configurable and not enabled by default. 382 The receiver interacts with the sender using an Acknowledgment 383 message with a bitmap that indicates which fragments were actually 384 received. The bitmap is a 32bit SWORD, which accommodates up to 32 385 fragments and is sufficient for the 6LoWPAN MTU. For all n in 386 [0..31], bit n is set to 1 in the bitmap to indicate that fragment 387 with sequence n was received, otherwise the bit is set to 0. All 388 zeroes is a NULL bitmap that indicates that the fragmentation process 389 was cancelled by the receiver for that datagram. 391 The receiver MAY issue unsolicited acknowledgments. An unsolicited 392 acknowledgment enables the sender endpoint to resume sending if it 393 had reached its maximum number of outstanding fragments or indicate 394 that the receiver has cancelled the process of an individual 395 datagram. Note that acknowledgments might consume precious resources 396 so the use of unsolicited acknowledgments should be configurable and 397 not enabled by default. 399 The sender arms a retry timer to cover the fragment that carries the 400 Acknowledgment request. Upon time out, the sender assumes that all 401 the fragments on the way are received or lost. The process must 402 completed within an acceptable time that is within the boundaries of 403 upper layer retries. The method detailed in [RFC2988] is recommended 404 for the computation of the retry timer. It is expected that the 405 upper layer retries obey the same or friendly rules in which case a 406 single round of fragment recovery should fit within the upper layer 407 recovery timers. 409 Fragments are sent in a round robin fashion: the sender sends all the 410 fragments for a first time before it retries any lost fragment; lost 411 fragments are retried in sequence, oldest first. This mechanism 412 enables the receiver to acknowledge fragments that were delayed in 413 the network before they are actually retried. 415 When the sender decides that a packet should be dropped and the 416 fragmentation process canceled, it sends a pseudo fragment with the 417 datagram_offset, sequence and datagram_size all set to zero, and no 418 data. Upon reception of this message, the receiver should clean up 419 all resources for the packet associated to the datagram_tag. If an 420 acknowledment is requested, the receiver responds with a NULL bitmap. 422 The receiver might need to cancel the process of a fragmented packet 423 for internal reasons, for instance if it is out of recomposition 424 buffers, or considers that this packet is already fully recomposed 425 and passed to the upper layer. In that case, the receiver SHOULD 426 indicate so to the sender with a NULL bitmap. Upon an 427 acknowledgement with a NULL bitmap, the sender MUST drop the 428 datagram. 430 8. Security Considerations 432 The process of recovering fragments does not appear to create any 433 opening for new threat compared to "Transmission of IPv6 Packets over 434 IEEE 802.15.4 Networks" [RFC4944]. 436 9. IANA Considerations 438 Need extensions for formats defined in "Transmission of IPv6 Packets 439 over IEEE 802.15.4 Networks" [RFC4944]. 441 10. Acknowledgments 443 The author wishes to thank Jay Werb, Christos Polyzois, Soumitri 444 Kolavennu and Harry Courtice for their contribution and review. 446 11. References 448 11.1. Normative References 450 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 451 Requirement Levels", BCP 14, RFC 2119, March 1997. 453 [RFC2988] Paxson, V. and M. Allman, "Computing TCP's Retransmission 454 Timer", RFC 2988, November 2000. 456 [RFC4944] Montenegro, G., Kushalnagar, N., Hui, J., and D. Culler, 457 "Transmission of IPv6 Packets over IEEE 802.15.4 458 Networks", RFC 4944, September 2007. 460 11.2. Informative References 462 [I-D.ietf-tsvwg-udp-guidelines] 463 Eggert, L. and G. Fairhurst, "Unicast UDP Usage Guidelines 464 for Application Designers", 465 draft-ietf-tsvwg-udp-guidelines-11 (work in progress), 466 October 2008. 468 [I-D.mathis-frag-harmful] 469 Mathis, M., "Fragmentation Considered Very Harmful", 470 draft-mathis-frag-harmful-00 (work in progress), 471 July 2004. 473 [RFC1191] Mogul, J. and S. Deering, "Path MTU discovery", RFC 1191, 474 November 1990. 476 [RFC2309] Braden, B., Clark, D., Crowcroft, J., Davie, B., Deering, 477 S., Estrin, D., Floyd, S., Jacobson, V., Minshall, G., 478 Partridge, C., Peterson, L., Ramakrishnan, K., Shenker, 479 S., Wroclawski, J., and L. Zhang, "Recommendations on 480 Queue Management and Congestion Avoidance in the 481 Internet", RFC 2309, April 1998. 483 [RFC2581] Allman, M., Paxson, V., and W. Stevens, "TCP Congestion 484 Control", RFC 2581, April 1999. 486 [RFC2914] Floyd, S., "Congestion Control Principles", BCP 41, 487 RFC 2914, September 2000. 489 [RFC3168] Ramakrishnan, K., Floyd, S., and D. Black, "The Addition 490 of Explicit Congestion Notification (ECN) to IP", 491 RFC 3168, September 2001. 493 [RFC4919] Kushalnagar, N., Montenegro, G., and C. Schumacher, "IPv6 494 over Low-Power Wireless Personal Area Networks (6LoWPANs): 495 Overview, Assumptions, Problem Statement, and Goals", 496 RFC 4919, August 2007. 498 [RFC4963] Heffner, J., Mathis, M., and B. Chandler, "IPv4 Reassembly 499 Errors at High Data Rates", RFC 4963, July 2007. 501 Authors' Addresses 503 Pascal Thubert (editor) 504 Cisco Systems 505 Village d'Entreprises Green Side 506 400, Avenue de Roumanille 507 Batiment T3 508 Biot - Sophia Antipolis 06410 509 FRANCE 511 Phone: +33 4 97 23 26 34 512 Email: pthubert@cisco.com 514 Jonathan W. Hui 515 Arch Rock Corporation 516 501 2nd St. Ste. 410 517 San Francisco, California 94107 518 USA 520 Phone: +415 692 0828 521 Email: jhui@archrock.com