idnits 2.17.1 draft-ietf-mboned-64-multicast-address-format-06.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 24, 2014) is 3499 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(See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) == Outdated reference: A later version (-18) exists of draft-ietf-softwire-dslite-multicast-08 == Outdated reference: A later version (-25) exists of draft-ietf-softwire-mesh-multicast-07 == Outdated reference: A later version (-15) exists of draft-ietf-softwire-multicast-prefix-option-07 -- Obsolete informational reference (is this intentional?): RFC 4566 (Obsoleted by RFC 8866) Summary: 0 errors (**), 0 flaws (~~), 4 warnings (==), 2 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Network Working Group M. Boucadair, Ed. 3 Internet-Draft France Telecom 4 Updates: 7371 (if approved) J. Qin 5 Intended status: Standards Track Cisco 6 Expires: March 28, 2015 Y. Lee 7 Comcast 8 S. Venaas 9 Cisco Systems 10 X. Li 11 CERNET Center/Tsinghua University 12 M. Xu 13 Tsinghua University 14 September 24, 2014 16 IPv6 Multicast Address With Embedded IPv4 Multicast Address 17 draft-ietf-mboned-64-multicast-address-format-06 19 Abstract 21 This document reserves one bit (M-bit) of the unicast prefix-based 22 multicast IPv6 address for ASM and an IPv6 multicast prefix for SSM 23 mode to be used in the context of IPv4-IPv6 interconnection. 25 The document specifies an algorithmic translation of an IPv6 26 multicast address to a corresponding IPv4 multicast address, and vice 27 versa. This algorithmic translation can be used in both IPv4-IPv6 28 translation or encapsulation schemes. 30 This document updates RFC 7371. 32 Requirements Language 34 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 35 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 36 document are to be interpreted as described in RFC 2119 [RFC2119]. 38 Status of This Memo 40 This Internet-Draft is submitted in full conformance with the 41 provisions of BCP 78 and BCP 79. 43 Internet-Drafts are working documents of the Internet Engineering 44 Task Force (IETF). Note that other groups may also distribute 45 working documents as Internet-Drafts. The list of current Internet- 46 Drafts is at http://datatracker.ietf.org/drafts/current/. 48 Internet-Drafts are draft documents valid for a maximum of six months 49 and may be updated, replaced, or obsoleted by other documents at any 50 time. It is inappropriate to use Internet-Drafts as reference 51 material or to cite them other than as "work in progress." 53 This Internet-Draft will expire on March 28, 2015. 55 Copyright Notice 57 Copyright (c) 2014 IETF Trust and the persons identified as the 58 document authors. All rights reserved. 60 This document is subject to BCP 78 and the IETF Trust's Legal 61 Provisions Relating to IETF Documents 62 (http://trustee.ietf.org/license-info) in effect on the date of 63 publication of this document. Please review these documents 64 carefully, as they describe your rights and restrictions with respect 65 to this document. Code Components extracted from this document must 66 include Simplified BSD License text as described in Section 4.e of 67 the Trust Legal Provisions and are provided without warranty as 68 described in the Simplified BSD License. 70 Table of Contents 72 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 3 73 2. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . 3 74 3. IPv4-Embedded IPv6 Multicast Prefix & Address . . . . . . . . 4 75 3.1. ASM Mode . . . . . . . . . . . . . . . . . . . . . . . . 4 76 3.2. SSM Mode . . . . . . . . . . . . . . . . . . . . . . . . 5 77 3.3. IPv4-Embedded IPv6 Multicast Address . . . . . . . . . . 5 78 3.4. Address Translation Algorithm . . . . . . . . . . . . . . 6 79 3.5. Textual Representation . . . . . . . . . . . . . . . . . 6 80 3.6. Source IPv4 Address in the IPv6 Realm . . . . . . . . . . 6 81 4. Examples . . . . . . . . . . . . . . . . . . . . . . . . . . 6 82 5. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 7 83 6. Security Considerations . . . . . . . . . . . . . . . . . . . 7 84 7. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . 7 85 8. References . . . . . . . . . . . . . . . . . . . . . . . . . 7 86 8.1. Normative References . . . . . . . . . . . . . . . . . . 7 87 8.2. Informative References . . . . . . . . . . . . . . . . . 8 88 Appendix A. Motivations . . . . . . . . . . . . . . . . . . . . 9 89 A.1. Why an Address Format is Needed for Multicast IPv4-IPv6 90 Interconnection? . . . . . . . . . . . . . . . . . . . . 9 91 A.2. Why Identifying an IPv4-Embedded IPv6 Multicast Address 92 is Required? . . . . . . . . . . . . . . . . . . . . . . 9 93 A.3. Location of the IPv4 Address . . . . . . . . . . . . . . 10 94 Appendix B. Design Considerations . . . . . . . . . . . . . . . 10 95 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 11 97 1. Introduction 99 Various solutions (e.g., [I-D.ietf-softwire-mesh-multicast], 100 [I-D.ietf-softwire-dslite-multicast]) have been proposed to allow 101 access to IPv4 multicast content from hosts attached to IPv6-enabled 102 domains. Even if these solutions have distinct applicability scopes 103 (translation vs. encapsulation) and target different use cases, they 104 all make use of specific IPv6 multicast addresses to embed an IPv4 105 multicast address. Particularly, the IPv4-Embedded IPv6 Multicast 106 Address is used as a destination IPv6 address of multicast flows 107 received from an IPv4-enabled domain and injected by the IPv4-IPv6 108 Interconnection Function into an IPv6-enabled domain. It is also 109 used to build an IPv6 multicast state (*, G6) or (S6, G6) 110 corresponding to their (*, G4) or (S4, G4) IPv4 counter parts by the 111 IPv4-IPv6 Interconnection Function. [I-D.ietf-mboned-v4v6-mcast-ps] 112 provides more discussion about issues related to IPv4/IPv6 multicast. 114 This document reserves one bit of the unicast prefix-based multicast 115 IPv6 address ([RFC7371]) for Any-Source Multicast (ASM) mode and an 116 IPv6 multicast prefix for Source-Specific Multicast (SSM) mode to be 117 used in the context of IPv4-IPv6 interconnection. This document also 118 defines how IPv4-Embedded IPv6 Multicast Addresses are constructed. 119 Both IPv4-IPv6 translation and encapsulation schemes can make use of 120 this specification. 122 This specification can be used in conjunction with other extensions 123 such as embedding the rendezvous point [RFC3956]. Unicast prefix- 124 based and embedded-RP techniques are important tools to simplify IPv6 125 multicast deployments. Indeed, unicast prefix-based IPv6 addressing 126 is used in many current IPv6 multicast deployments, and has also been 127 defined for IPv4, and is seen as a very useful technique. Also 128 embedded-RP is used in existing deployments. 130 This document is a companion document to [RFC6052] which focuses 131 exclusively on IPv4-embedded IPv6 unicast addresses. 133 2. Terminology 135 This document makes use of the following terms: 137 o IPv4-Embedded IPv6 Multicast Address: denotes a multicast IPv6 138 address which includes in 32 bits an IPv4 address. 140 o Multicast Prefix64 (or MPREFIX64 for short) refers to an IPv6 141 multicast prefix to be used to construct IPv4-Embedded IPv6 142 Multicast Addresses. This prefix is used to build an 143 IPv4-Embedded IPv6 Multicast Address as defined in Section 3.4. 145 Section 3.4 specifies also how to extract an IPv4 address from an 146 IPv4-Embedded IPv6 Multicast Address. 148 o ASM_MPREFIX64: denotes a multicast Prefix64 used in Any Source 149 Multicast (ASM) mode. 151 o SSM_MPREFIX64: denotes a multicast Prefix64 used in Source 152 Specific Multicast (SSM) mode. 154 o IPv4-IPv6 Interconnection Function: refers to a function which is 155 enabled in a node interconnecting an IPv4-enabled domain with an 156 IPv6-enabled one. It can be located in various places of the 157 multicast network. Particularly, in terms of multicast control 158 messages, it can be an IGMP/MLD Interworking Function or an 159 IPv4-IPv6 PIM Interworking Function. An IPv4-IPv6 Interconnection 160 Function is configured with one or two MPREFIX64s. 162 3. IPv4-Embedded IPv6 Multicast Prefix & Address 164 3.1. ASM Mode 166 The format specified in Figure 1 uses some bits defined in [RFC7371]: 167 M-bit (20th bit position, where bit 1 is the most significant bit) 168 now has a meaning. 170 Details on design considerations are discussed in Appendix B. 172 | 8 | 4 | 4 | 4 | 76 | 32 | 173 +--------+----+----+----+------------------------------+----------+ 174 |11111111|ff1 |scop|ff2 | sub-group-id |v4 address| 175 +--------+----+----+----+-----------------------------------------+ 176 +-+-+-+-+ 177 ff2 (flag field 2) is a set of 4 flags: |r|r|r|M| 178 +-+-+-+-+ 180 "r" bits MUST each be set to 0. 182 Figure 1: IPv4-Embedded IPv6 Multicast Address Format: ASM Mode 184 The description of the fields is as follows: 186 o ff1 field is defined in [RFC7371]. 187 o "scop" field is defined in [RFC3956]. 188 o M (20th bit position): When this bit is set to 1, it indicates 189 that a multicast IPv4 address is embedded in the low-order 32 bits 190 of the multicast IPv6 address. 192 o sub-group-id: This field is configurable according to local 193 policies (e.g., enable embedded-RP) of the entity managing the 194 IPv4-IPv6 Interconnection Function. This field MUST follow the 195 recommendations specified in [RFC3306] if unicast-based prefix is 196 used or the recommendations specified in [RFC3956] if embedded-RP 197 is used. The default value is all zeros. 198 o The low-order 32 bits MUST include an IPv4 multicast address when 199 the M-bit is set to 1. The enclosed IPv4 multicast address SHOULD 200 NOT be in 232/8 range. 202 3.2. SSM Mode 204 For SSM mode, and given what is discussed in Appendix B, the 205 following IPv6 prefix to embed IPv4 multicast addresses is reserved: 207 o ff3x:0:8000::/96 ('x' is any valid scope). 209 3.3. IPv4-Embedded IPv6 Multicast Address 211 For the delivery of the IPv4-IPv6 multicast interconnection services, 212 a dedicated multicast prefix denoted as MPREFIX64 should be 213 provisioned (e.g., using NETCONF or 214 [I-D.ietf-softwire-multicast-prefix-option]) to any function 215 requiring to build an IPv4-Embedded IPv6 Multicast Address based on 216 an IPv4 multicast address. MPREFIX64 can be of ASM or SSM type. 217 When both modes are used, two prefixes are required to be 218 provisioned. 220 The length of MPREFIX64 MUST be /96. For SSM, MPREFIX64 MUST be 221 equal to ff3x:0:8000::/96. For the ASM mode, MPREFIX64 MUST have the 222 M-bit set to 1. Furthermore, the format of the ASM_MPREFIX64 should 223 follow what is specified in [RFC3306] and [RFC3956] if corresponding 224 mechanisms are used. If not, bits 21-96 can be set to any value. 226 Figure 2 shows how to build an IPv4-Embedded IPv6 Multicast Address 227 using a configured MPREFIX64 and an IPv4 multicast address. The low- 228 order 32 bits MUST include an IPv4 multicast address. The enclosed 229 IPv4 multicast address SHOULD NOT be in 232/8 range if an 230 ASM_PREFIX64 is configured. The enclosed IPv4 multicast address 231 SHOULD be in 232/8 range if an SSM_PREFIX64 is configured. 233 Embedding an IPv4 multicast address in the last 32 bits does not 234 conflict with the Group IDs assigned by IANA (i.e., 0x00000001 to 235 0x3FFFFFFF [RFC3307]). 237 When several MPREFIX64 are available, it is RECOMMENDED to use the 238 MPREFIX64 which preserve the scope of the IPv4 multicast address. 240 | 96 | 32 | 241 +------------------------------------------------------+----------+ 242 | MPREFIX64 |v4 address| 243 +------------------------------------------------------+----------+ 245 Figure 2: IPv4-Embedded IPv6 Multicast Address Format 247 3.4. Address Translation Algorithm 249 IPv4-Embedded IPv6 Multicast Addresses are composed according to the 250 following algorithm: 252 o Concatenate the MPREFIX64 and the 32 bits of the IPv4 address to 253 obtain a 128-bit address. 255 The IPv4 multicast addresses are extracted from the IPv4-Embedded 256 IPv6 Multicast Addresses according to the following algorithm: 258 o If the multicast address has the 20th bit set to 1 or it matches 259 ff3x:0:8000::/96 or a preconfigured MPREFIX64, extract the last 32 260 bits of the IPv6 multicast address. 262 3.5. Textual Representation 264 The embedded IPv4 address in an IPv6 multicast address is included in 265 the last 32 bits; therefore dotted decimal notation can be used. 267 3.6. Source IPv4 Address in the IPv6 Realm 269 An IPv4 source is represented in the IPv6 realm with its 270 IPv4-converted IPv6 address [RFC6052]. 272 4. Examples 274 Figure 3 provides some examples of ASM IPv4-Embedded IPv6 Address 275 while Figure 4 provides an example of SSM IPv4-Embedded IPv6 Address. 277 IPv4 multicast addresses used in the examples are derived from the 278 IPv4 multicast block reserved for documentation in [RFC6676]. 280 +----------------------+--------------+-----------------------------+ 281 | MPREFIX64 | IPv4 address | IPv4-Embedded IPv6 Address | 282 +----------------------+--------------+-----------------------------+ 283 | ff3x:z000:0:abc::/96 | 233.252.0.1 |ff3x:z000:0:abc::233.252.0.1 | 284 | ff7x:z000:0:abc::/96 | 233.252.0.2 |ff7x:z000:0:abc::233.252.0.2 | 285 +----------------------+--------------+-----------------------------+ 286 where: 287 "x" is any valid scope 288 "z" is any 4 bits where the last bit is set (e.g., 1, 3, 7, ...) 290 Figure 3: Example of ASM IPv4-embedded IPv6 address 292 +---------------------+--------------+----------------------------+ 293 | MPREFIX64 | IPv4 address | IPv4-Embedded IPv6 Address | 294 +---------------------+--------------+----------------------------+ 295 | ff3x:0:8000::/96 | 233.252.0.5 | ff3x:0:8000::233.252.0.5 | 296 +---------------------+--------------+----------------------------+ 298 Figure 4: Example of SSM IPv4-embedded IPv6 address 300 5. IANA Considerations 302 This document requests IANA to reserve: 304 o ff3x:0:8000::/96 SSM range to embed an IPv4 multicast address in 305 the last 32 bits. 307 6. Security Considerations 309 This document defines an algorithmic translation of an IPv6 multicast 310 address into an IPv4 multicast address, and vice versa. The security 311 considerations discussed in [RFC6052] are to be taken into 312 consideration. 314 7. Acknowledgements 316 Many thanks to R. Bonica, B. Sarikaya, P. Savola, T. Tsou, C. 317 Bormann, T. Chown, P. Koch, B. Haberman, and B. Hinden for their 318 comments and review. 320 8. References 322 8.1. Normative References 324 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 325 Requirement Levels", BCP 14, RFC 2119, March 1997. 327 [RFC3306] Haberman, B. and D. Thaler, "Unicast-Prefix-based IPv6 328 Multicast Addresses", RFC 3306, August 2002. 330 [RFC3307] Haberman, B., "Allocation Guidelines for IPv6 Multicast 331 Addresses", RFC 3307, August 2002. 333 [RFC3956] Savola, P. and B. Haberman, "Embedding the Rendezvous 334 Point (RP) Address in an IPv6 Multicast Address", RFC 335 3956, November 2004. 337 [RFC4607] Holbrook, H. and B. Cain, "Source-Specific Multicast for 338 IP", RFC 4607, August 2006. 340 [RFC6052] Bao, C., Huitema, C., Bagnulo, M., Boucadair, M., and X. 341 Li, "IPv6 Addressing of IPv4/IPv6 Translators", RFC 6052, 342 October 2010. 344 [RFC7371] Boucadair, M. and S. Venaas, "Updates to the IPv6 345 Multicast Addressing Architecture", RFC 7371, September 346 2014. 348 8.2. Informative References 350 [I-D.ietf-mboned-v4v6-mcast-ps] 351 Jacquenet, C., Boucadair, M., Lee, Y., Qin, J., Tsou, T., 352 and Q. Qiong, "IPv4-IPv6 Multicast: Problem Statement and 353 Use Cases", draft-ietf-mboned-v4v6-mcast-ps-04 (work in 354 progress), September 2013. 356 [I-D.ietf-softwire-dslite-multicast] 357 Qin, J., Boucadair, M., Jacquenet, C., Lee, Y., and Q. 358 Wang, "Delivery of IPv4 Multicast Services to IPv4 Clients 359 over an IPv6 Multicast Network", draft-ietf-softwire- 360 dslite-multicast-08 (work in progress), September 2014. 362 [I-D.ietf-softwire-mesh-multicast] 363 Xu, M., Cui, Y., Wu, J., Yang, S., Metz, C., and G. 364 Shepherd, "Softwire Mesh Multicast", draft-ietf-softwire- 365 mesh-multicast-07 (work in progress), July 2014. 367 [I-D.ietf-softwire-multicast-prefix-option] 368 Boucadair, M., Qin, J., Tsou, T., and X. Deng, "DHCPv6 369 Option for IPv4-Embedded Multicast and Unicast IPv6 370 Prefixes", draft-ietf-softwire-multicast-prefix-option-07 371 (work in progress), September 2014. 373 [RFC4566] Handley, M., Jacobson, V., and C. Perkins, "SDP: Session 374 Description Protocol", RFC 4566, July 2006. 376 [RFC6676] Venaas, S., Parekh, R., Van de Velde, G., Chown, T., and 377 M. Eubanks, "Multicast Addresses for Documentation", RFC 378 6676, August 2012. 380 [RFC7051] Korhonen, J. and T. Savolainen, "Analysis of Solution 381 Proposals for Hosts to Learn NAT64 Prefix", RFC 7051, 382 November 2013. 384 Appendix A. Motivations 386 A.1. Why an Address Format is Needed for Multicast IPv4-IPv6 387 Interconnection? 389 Arguments why an IPv6 address format is needed to embed multicast 390 IPv4 address are quite similar to those of [RFC6052]. Concretely, 391 the definition of a multicast address format embedding a multicast 392 IPv4 address allows: 394 o Stateless IPv4-IPv6 header translation of multicast flows; 396 o Stateless IPv4-IPv6 PIM interworking function; 398 o Stateless IGMP-MLD interworking function (e.g., required for an 399 IPv4 receiver to access to IPv4 multicast content via an IPv6 400 network); 402 o Stateless (local) synthesis of IPv6 address when IPv4 multicast 403 address are embedded in application payload (e.g., SDP [RFC4566]); 405 o Except the provisioning of the same MPREFIX64, no coordination is 406 required between IPv4-IPv6 PIM interworking function, IGMP-MLD 407 interworking function, IPv4-IPv6 Interconnection Function and any 408 ALG (Application Level Gateway) in the path; 410 o Minimal operational constraints on the multicast address 411 management: IPv6 multicast addresses can be constructed using what 412 has been deployed for IPv4 delivery mode. 414 A.2. Why Identifying an IPv4-Embedded IPv6 Multicast Address is 415 Required? 417 Reserving a dedicated multicast prefix for IPv4-IPv6 interconnection 418 purposes is a means to guide the address selection process at the 419 receiver side; in particular it assists the receiver to select the 420 appropriate IP multicast address while avoiding to involve an 421 IPv4-IPv6 interconnection function in the path. 423 Two use cases to illustrate this behavior are provided below: 425 1. An ALG is required to help an IPv6 receiver to select the 426 appropriate IP address when only the IPv4 address is advertised 427 (e.g., using SDP); otherwise the access to the IPv4 multicast 428 content can not be offered to the IPv6 receiver. The ALG may be 429 located downstream the receiver. As such, the ALG does not know 430 in advance whether the receiver is dual-stack or IPv6-only. The 431 ALG may be tuned to insert both the original IPv4 address and 432 corresponding IPv6 multicast address. If a dedicated prefix is 433 not used, a dual-stack receiver may prefer to use the IPv6 434 address to receive the multicast content. This address selection 435 would require multicast flows to cross an IPv4-IPv6 436 interconnection function. 438 2. In order to avoid involving an ALG in the path, an IPv4-only 439 source can advertise both its IPv4 address and IPv4-Embedded IPv6 440 Multicast Address. If a dedicated prefix is not reserved, a 441 dual-stack receiver may prefer to use the IPv6 address to receive 442 the multicast content. This address selection would require 443 multicast flows to cross an IPv4-IPv6 interconnection function. 445 Reserving dedicated IPv6 multicast prefixes for IPv4-IPv6 446 interconnection purposes mitigates the issues discussed in [RFC7051] 447 in a multicast context. 449 A.3. Location of the IPv4 Address 451 There is no strong argument to allow for flexible options to encode 452 the IPv4 address inside the multicast IPv6 address. The option 453 retained by the authors is to encode the multicast IPv4 address in 454 the low-order 32 bits of the IPv6 address. 456 This choice is also motivated by the need to be compliant with 457 [RFC3306] and [RFC3956]. 459 Appendix B. Design Considerations 461 The following constraints should be met when reserving dedicated 462 prefix(es) to be used for IPv4/IPv6 multicast interconnection: 464 1: Belong to SSM prefix range (preferably ff3x::/32) and be 465 compatible with unicast-based prefix [RFC3306] for SSM. Note 466 that [RFC3306] suggests to set "plen" to 0 and "network-prefix" 467 to 0. As such, any prefix in the 33-96 range can be convenient. 468 Given [RFC4607] indicates future specifications may allow a non- 469 zero network prefix field, a /33 would allow for future 470 extensions but it has the drawback of reserving a large block. A 471 /96 would be adequate for the use cases already identified in 473 [I-D.ietf-mboned-v4v6-mcast-ps]. In the event of any concrete 474 extension, reserving additional prefixes may be considered. 476 2: Be compatible with embedded-RP [RFC3956] and unicast-based prefix 477 [RFC3306] for ASM. This results in associating a meaning with 478 one of the reserved bits in [RFC7371]. Defining the 17-20 bits 479 range to have a meaning and be used for IPv4/IPv6 transition has 480 the advantage of allowing for future extensions but it may be 481 seen as a waste of the multicast address space. Consequently, 482 using one of the reserved bits (in the range 17-20) from the 483 unicast-based IPv6 multicast address format [RFC3306] is 484 preferred. 486 Meeting (1) and (2) with the same reserved bit is not feasible 487 without modifying embedded-RP and unicast-based prefix 488 specifications; this option is avoided. 490 As a consequence, this document proposes to reserve a multicast 491 prefix for SSM and define one bit of the unicast prefix-based 492 multicast IPv6 address for ASM when embedding IPv4 multicast address 493 in an IPv6 multicast address. 495 Authors' Addresses 497 Mohamed Boucadair (editor) 498 France Telecom 499 Rennes 35000 500 France 502 Email: mohamed.boucadair@orange.com 504 Jacni Qin 505 Cisco 506 China 508 Email: jacni@jacni.com 510 Yiu L. Lee 511 Comcast 512 U.S.A 514 Email: yiu_lee@cable.comcast.com 515 Stig Venaas 516 Cisco Systems 517 Tasman Drive 518 San Jose, CA 95134 519 USA 521 Email: stig@cisco.com 523 Xing Li 524 CERNET Center/Tsinghua University 525 Room 225, Main Building, Tsinghua University 526 Beijing 100084 527 P.R. China 529 Phone: +86 10-62785983 530 Email: xing@cernet.edu.cn 532 Mingwei Xu 533 Tsinghua University 534 Department of Computer Science, Tsinghua University 535 Beijing 100084 536 P.R.China 538 Phone: +86-10-6278-5822 539 Email: xmw@cernet.edu.cn