idnits 2.17.1 draft-ietf-nfsv4-rfc5667bis-01.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 (June 30, 2016) is 2857 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(See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) == Outdated reference: A later version (-11) exists of draft-ietf-nfsv4-rfc5666bis-07 == Outdated reference: A later version (-08) exists of draft-ietf-nfsv4-rpcrdma-bidirection-05 ** Obsolete normative reference: RFC 5661 (Obsoleted by RFC 8881) -- Obsolete informational reference (is this intentional?): RFC 5667 (Obsoleted by RFC 8267) Summary: 1 error (**), 0 flaws (~~), 3 warnings (==), 2 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Network File System Version 4 C. Lever, Ed. 3 Internet-Draft Oracle 4 Obsoletes: 5667 (if approved) June 30, 2016 5 Intended status: Standards Track 6 Expires: January 1, 2017 8 Network File System (NFS) Upper Layer Binding To RPC-Over-RDMA 9 draft-ietf-nfsv4-rfc5667bis-01 11 Abstract 13 This document specifies the Upper Layer Bindings of Network File 14 System (NFS) protocol versions to RPC-over-RDMA transports. Such 15 Upper Layer Bindings are required to enable RPC-based protocols to 16 use direct data placement when conveying large data payloads on RPC- 17 over-RDMA transports. This document obsoletes RFC 5667. 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 January 1, 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 1.1. Requirements Language . . . . . . . . . . . . . . . . . . 3 55 1.2. Changes Since RFC 5667 . . . . . . . . . . . . . . . . . 3 56 1.3. Planned Changes To This Document . . . . . . . . . . . . 4 57 2. Conveying NFS Operations On RPC-Over-RDMA Transports . . . . 4 58 2.1. Use Of The Read List . . . . . . . . . . . . . . . . . . 4 59 2.2. Use Of The Write List . . . . . . . . . . . . . . . . . . 5 60 2.3. Construction Of Individual Chunks . . . . . . . . . . . . 5 61 2.4. Use Of Long Calls And Replies . . . . . . . . . . . . . . 5 62 3. NFS Versions 2 And 3 Upper Layer Binding . . . . . . . . . . 5 63 4. NFS Version 4 Upper Layer Binding . . . . . . . . . . . . . . 6 64 4.1. NFS Version 4 COMPOUND Considerations . . . . . . . . . . 7 65 4.2. NFS Version 4 Callbacks . . . . . . . . . . . . . . . . . 8 66 5. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 8 67 6. Security Considerations . . . . . . . . . . . . . . . . . . . 9 68 7. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 9 69 8. References . . . . . . . . . . . . . . . . . . . . . . . . . 9 70 8.1. Normative References . . . . . . . . . . . . . . . . . . 9 71 8.2. Informative References . . . . . . . . . . . . . . . . . 10 72 Author's Address . . . . . . . . . . . . . . . . . . . . . . . . 11 74 1. Introduction 76 Remote Direct Memory Access Transport for Remote Procedure Call, 77 Version One [I-D.ietf-nfsv4-rfc5666bis] (RPC-over-RDMA) enables the 78 use of direct data placement to accelerate the transmission of large 79 data payloads associated with RPC transactions. 81 Each RPC-over-RDMA transport header can convey lists of memory 82 locations involved in direct transfers of data payloads. These 83 memory locations correspond to XDR data items defined in an Upper 84 Layer Protocol (such as NFS). 86 To facilitate interoperation, RPC client and server implementations 87 must agree on what XDR data items in which RPC procedures are 88 eligible for direct data placement (DDP). 90 This document specifies the set of XDR data items in each of the 91 following NFS protocol versions that are eligible for DDP. It also 92 contains additional material required of Upper Layer Bindings as 93 specified in [I-D.ietf-nfsv4-rfc5666bis]. 95 o NFS Version 2 [RFC1094] 96 o NFS Version 3 [RFC1813] 98 o NFS Version 4.0 [RFC7530] 100 o NFS Version 4.1 [RFC5661] 102 o NFS Version 4.2 [I-D.ietf-nfsv4-minorversion2] 104 The Upper Layer Binding specified in this document can be extended to 105 cover the addition of new DDP-eligible XDR data items defined by 106 versions of the NFS version 4 protocol specified after this document 107 has been ratified. 109 1.1. Requirements Language 111 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 112 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 113 document are to be interpreted as described in [RFC2119]. 115 1.2. Changes Since RFC 5667 117 Corrections and updates made necessary by new language in 118 [I-D.ietf-nfsv4-rfc5666bis] has been introduced. For example, 119 references to deprecated features of RPC-over-RDMA Version One, such 120 as RDMA_MSGP, and the use of the Read list for handling RPC replies, 121 has been removed. The term "mapping" has been replaced with the term 122 "binding" or "Upper Layer Binding" throughout the document. Material 123 that duplicates what is in [I-D.ietf-nfsv4-rfc5666bis] has been 124 deleted. 126 Material required by [I-D.ietf-nfsv4-rfc5666bis] for Upper Layer 127 Bindings that was not present in [RFC5667] has been added, including 128 discussion of how each NFS version properly estimates the maximum 129 size of RPC replies. 131 The following changes have been made, relative to [RFC5667]: 133 o Ambiguous or erroneous uses of RFC2119 terms have been corrected. 135 o References to specific data movement mechanisms have been made 136 generic or removed. 138 o References to obsolete RFCs have been replaced. 140 o Technical corrections have been made. For example, the mention of 141 12KB and 36KB inline thresholds have been removed. The reference 142 to a non-existant NFS version 4 SYMLINK operation has been 143 replaced with NFS version 4 CREATE(NF4LNK). 145 o An IANA Considerations Section has replaced the "Port Usage 146 Considerations" Section. 148 o Code excerpts have been removed, and figures have been modernized. 150 o Language inconsistent with or contradictory to 151 [I-D.ietf-nfsv4-rfc5666bis] has been removed from Sections 2 and 152 3, and both Sections have been combined into Section 2 in the 153 present document. 155 o An explicit discussion of NFSv4.0 and NFSv4.1 backchannel 156 operation will replace the previous treatment of callback 157 operations. No NFSv4.x callback operation is DDP-eligible. 159 o The binding for NFSv4.1 has been completed. No additional DDP- 160 eligible operations exist in NFSv4.1. 162 o A binding for NFSv4.2 has been added that includes discussion of 163 new data-bearing operations like READ_PLUS. 165 1.3. Planned Changes To This Document 167 The following changes are planned, relative to [RFC5667]: 169 o The discussion of NFS version 4 COMPOUND handling will be 170 completed. 172 o Remarks about handling DDP-eligibility violations will be 173 introduced. 175 o A discussion of how the NFS binding to RPC-over-RDMA is extended 176 by standards action will be added. 178 2. Conveying NFS Operations On RPC-Over-RDMA Transports 180 Definitions of terminology and a general discussion of how RPC-over- 181 RDMA is used to convey RPC transactions can be found in 182 [I-D.ietf-nfsv4-rfc5666bis]. In this section, these general 183 principals are applied to the specifics of the NFS protocol. 185 2.1. Use Of The Read List 187 The Read list in each RPC-over-RDMA transport header represents a set 188 of memory regions containing DDP-eligible NFS argument data. Large 189 data items, such as the file data payload of an NFS WRITE request, 190 are referenced by the Read list and placed directly into server 191 memory. 193 XDR unmarshaling code on the NFS server identifies the correspondence 194 between Read chunks and particular NFS arguments via the chunk 195 Position value encoded in each Read chunk. 197 2.2. Use Of The Write List 199 The Write list in each RPC-over-RDMA transport header represents a 200 set of memory regions that can receive DDP-eligible NFS result data. 201 Large data items such as the payload of an NFS READ request are 202 referenced by the Write list and placed directly into client memory. 204 Each Write chunk corresponds to a specific XDR data item in an NFS 205 reply. This document specifies how NFS client and server 206 implementations identify the correspondence between Write chunks and 207 each XDR result. 209 2.3. Construction Of Individual Chunks 211 Each Read chunk is represented as a list of segments at the same XDR 212 Position, and each Write chunk is represented as an array of 213 segments. An NFS client thus has the flexibility to advertise a set 214 of discontiguous memory regions in which to send or receive a single 215 DDP-eligible data item. 217 2.4. Use Of Long Calls And Replies 219 Small RPC messages are conveyed using RDMA Send operations which are 220 of limited size. If an NFS request is too large to be conveyed via 221 an RDMA Send, and there are no DDP-eligible data items that can be 222 removed, an NFS client must send the request using a Long Call. The 223 entire NFS request is sent in a special Read chunk. 225 If a client expects that an NFS reply will be too large to be 226 conveyed via an RDMA Send, it provides a Reply chunk in the RPC-over- 227 RDMA transport header conveying the NFS request. The server can 228 place the entire NFS reply in the Reply chunk. 230 These are described in more detail in [I-D.ietf-nfsv4-rfc5666bis]. 232 3. NFS Versions 2 And 3 Upper Layer Binding 234 An NFS client MAY send a single Read chunk to supply opaque file data 235 for an NFS WRITE procedure, or the pathname for an NFS SYMLINK 236 procedure. For all other NFS procedures, the server MUST ignore Read 237 chunks that have a non-zero value in their Position fields, and Read 238 chunks beyond the first in the Read list. 240 Similarly, an NFS client MAY provide a single Write chunk to receive 241 either opaque file data from an NFS READ procedure, or the pathname 242 from an NFS READLINK procedure. The server MUST ignore the Write 243 list for any other NFS procedure, and any Write chunks beyond the 244 first in the Write list. 246 There are no NFS version 2 or 3 procedures that have DDP-eligible 247 data items in both their Call and Reply. However, if an NFS client 248 is sending a Long Call or Reply, it MAY provide a combination of Read 249 list, Write list, and/or a Reply chunk in the same transaction. 251 NFS clients already successfully estimate the maximum reply size of 252 each operation in order to provide an adequate set of buffers to 253 receive each NFS reply. An NFS client provides a Reply chunk when 254 the maximum possible reply size is larger than the client's responder 255 inline threshold. 257 How does the server respond if the client has not provided enough 258 Write list resources to handle an NFS WRITE or READLINK reply? How 259 does the server respond if the client has not provided enough Reply 260 chunk resources to handle an NFS reply? 262 4. NFS Version 4 Upper Layer Binding 264 This specification applies to NFS Version 4.0 [RFC7530], NFS Version 265 4.1 [RFC5661], and NFS Version 4.2 [I-D.ietf-nfsv4-minorversion2]. 266 It also applies to the callback protocols associated with each of 267 these minor versions. 269 An NFS client MAY send a Read chunk to supply opaque file data for a 270 WRITE operation or the pathname for a CREATE(NF4LNK) operation in an 271 NFS version 4 COMPOUND procedure. An NFS client MUST NOT send a Read 272 chunk that corresponds with any other XDR data item in any other NFS 273 version 4 operation. 275 Similarly, an NFS client MAY provide a Write chunk to receive either 276 opaque file data from a READ operation, NFS4_CONTENT_DATA from a 277 READ_PLUS operation, or the pathname from a READLINK operation in an 278 NFS version 4 COMPOUND procedure. An NFS client MUST NOT provide a 279 Write chunk that corresponds with any other XDR data item in any 280 other NFS version 4 operation. 282 There is no prohibition against an NFS version 4 COMPOUND procedure 283 constructed with both a READ and WRITE operation, say. Thus it is 284 possible for NFS version 4 COMPOUND procedures to use both the Read 285 list and Write list simultaneously. An NFS client MAY provide a Read 286 list and a Write list in the same transaction if it is sending a Long 287 Call or Reply. 289 Some remarks need to be made about how NFS version 4 clients estimate 290 reply size, and how DDP-eligibility violations are reported. 292 4.1. NFS Version 4 COMPOUND Considerations 294 An NFS version 4 COMPOUND procedure supplies arguments for a sequence 295 of operations, and returns results from that sequence. A client MAY 296 construct an NFS version 4 COMPOUND procedure that uses more than one 297 chunk in either the Read list or Write list. The NFS client provides 298 XDR Position values in each Read chunk to disambiguate which chunk is 299 associated with which XDR data item. 301 However NFS server and client implementations must agree in advance 302 on how to pair Write chunks with returned result data items. The 303 mechanism specified in [I-D.ietf-nfsv4-rfc5666bis]) is applied here: 305 o The first chunk in the Write list MUST be used by the first READ 306 or READLINK operation in an NFS version 4 COMPOUND procedure. The 307 next Write chunk is used by the next READ or READLINK, and so on. 309 o If there are more READ or READLINK operations than Write chunks, 310 then any remaining operations MUST return their results inline. 312 o If an NFS client presents a Write chunk, then the corresponding 313 READ or READLINK operation MUST return its data by placing data 314 into that chunk. 316 o If the Write chunk has zero RDMA segments, or if the total size of 317 the segments is zero, then the corresponding READ or READLINK 318 operation MUST return its result inline. 320 The following example shows a Write list with three Write chunks, A, 321 B, and C. The server consumes the provided Write chunks by writing 322 the results of the designated operations in the compound request, 323 READ and READLINK, back to each chunk. 325 Write list: 327 A --> B --> C 329 NFS version 4 COMPOUND request: 331 PUTFH LOOKUP READ PUTFH LOOKUP READLINK PUTFH LOOKUP READ 332 | | | 333 v v v 334 A B C 336 If the client does not want to have the READLINK result returned 337 directly, it provides a zero-length array of segment triplets for 338 buffer B or sets the values in the segment triplet for buffer B to 339 zeros to indicate that the READLINK result must be returned inline. 341 Unlike NFS versions 2 and 3, the maximum size of an NFS version 4 342 COMPOUND is not bounded. However, typical NFS version 4 clients 343 rarely issue such problematic requests. In practice, NFS version 4 344 clients behave in much more predictable ways. Rsize and wsize apply 345 to COMPOUND operations by capping the total amount of data payload 346 allowed in each COMPOUND. An extension to NFS version 4 supporting a 347 comprehensive exchange of upper-layer message size parameters is part 348 of [RFC5661]. 350 4.2. NFS Version 4 Callbacks 352 The NFS version 4 protocols support server-initiated callbacks to 353 notify clients of events such as recalled delegations. There are no 354 DDP-eligible data items in callback protocols associated with 355 NFSv4.0, NFSv4.1, or NFSv4.2. 357 In NFS version 4.1 and 4.2, callback operations may appear on the 358 same connection as one used for NFS version 4 client requests. To 359 operate on RPC-over-RDMA transports, NFS version 4 clients and 360 servers MUST use the mechanism described in 361 [I-D.ietf-nfsv4-rpcrdma-bidirection]. 363 5. IANA Considerations 365 NFS use of direct data placement introduces a need for an additional 366 NFS port number assignment for networks that share traditional UDP 367 and TCP port spaces with RDMA services. The iWARP [RFC5041] 368 [RFC5040] protocol is such an example (InfiniBand is not). 370 NFS servers for versions 2 and 3 [RFC1094] [RFC1813] traditionally 371 listen for clients on UDP and TCP port 2049, and additionally, they 372 register these with the portmapper and/or rpcbind [RFC1833] service. 373 However, [RFC7530] requires NFS servers for version 4 to listen on 374 TCP port 2049, and they are not required to register. 376 An NFS version 2 or version 3 server supporting RPC-over-RDMA on such 377 a network and registering itself with the RPC portmapper MAY choose 378 an arbitrary port, or MAY use the alternative well-known port number 379 for its RPC-over-RDMA service. The chosen port MAY be registered 380 with the RPC portmapper under the netid assigned by the requirement 381 in [I-D.ietf-nfsv4-rfc5666bis]. 383 An NFS version 4 server supporting RPC-over-RDMA on such a network 384 MUST use the alternative well-known port number for its RPC-over-RDMA 385 service. Clients SHOULD connect to this well-known port without 386 consulting the RPC portmapper (as for NFSv4/TCP). 388 The port number assigned to an NFS service over an RPC-over-RDMA 389 transport is available from the IANA port registry [RFC3232]. 391 6. Security Considerations 393 The RDMA transport for RPC [I-D.ietf-nfsv4-rfc5666bis] supports all 394 RPC [RFC5531] security models, including RPCSEC_GSS [RFC2203] 395 security and transport-level security. The choice of RDMA Read and 396 RDMA Write to convey RPC argument and results does not affect this, 397 since it only changes the method of data transfer. Specifically, the 398 requirements of [I-D.ietf-nfsv4-rfc5666bis] ensure that this choice 399 does not introduce new vulnerabilities. 401 Because this document defines only the binding of the NFS protocols 402 atop [I-D.ietf-nfsv4-rfc5666bis], all relevant security 403 considerations are therefore to be described at that layer. 405 7. Acknowledgments 407 The author gratefully acknowledges the work of Brent Callaghan and 408 Tom Talpey on the original NFS Direct Data Placement specification 409 [RFC5667]. The author also wishes to thank Bill Baker and Greg 410 Marsden for their support of this work. 412 Dave Noveck provided excellent review, constructive suggestions, and 413 consistent navigational guidance throughout the process of drafting 414 this document. 416 Special thanks go to nfsv4 Working Group Chair Spencer Shepler and 417 nfsv4 Working Group Secretary Thomas Haynes for their support. 419 8. References 421 8.1. Normative References 423 [I-D.ietf-nfsv4-minorversion2] 424 Haynes, T., "NFS Version 4 Minor Version 2", draft-ietf- 425 nfsv4-minorversion2-41 (work in progress), January 2016. 427 [I-D.ietf-nfsv4-rfc5666bis] 428 Lever, C., Simpson, W., and T. Talpey, "Remote Direct 429 Memory Access Transport for Remote Procedure Call, Version 430 One", draft-ietf-nfsv4-rfc5666bis-07 (work in progress), 431 May 2016. 433 [I-D.ietf-nfsv4-rpcrdma-bidirection] 434 Lever, C., "Bi-directional Remote Procedure Call On RPC- 435 over-RDMA Transports", draft-ietf-nfsv4-rpcrdma- 436 bidirection-05 (work in progress), June 2016. 438 [RFC1833] Srinivasan, R., "Binding Protocols for ONC RPC Version 2", 439 RFC 1833, DOI 10.17487/RFC1833, August 1995, 440 . 442 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 443 Requirement Levels", BCP 14, RFC 2119, 444 DOI 10.17487/RFC2119, March 1997, 445 . 447 [RFC2203] Eisler, M., Chiu, A., and L. Ling, "RPCSEC_GSS Protocol 448 Specification", RFC 2203, DOI 10.17487/RFC2203, September 449 1997, . 451 [RFC5531] Thurlow, R., "RPC: Remote Procedure Call Protocol 452 Specification Version 2", RFC 5531, DOI 10.17487/RFC5531, 453 May 2009, . 455 [RFC5661] Shepler, S., Ed., Eisler, M., Ed., and D. Noveck, Ed., 456 "Network File System (NFS) Version 4 Minor Version 1 457 Protocol", RFC 5661, DOI 10.17487/RFC5661, January 2010, 458 . 460 [RFC7530] Haynes, T., Ed. and D. Noveck, Ed., "Network File System 461 (NFS) Version 4 Protocol", RFC 7530, DOI 10.17487/RFC7530, 462 March 2015, . 464 8.2. Informative References 466 [RFC1094] Nowicki, B., "NFS: Network File System Protocol 467 specification", RFC 1094, DOI 10.17487/RFC1094, March 468 1989, . 470 [RFC1813] Callaghan, B., Pawlowski, B., and P. Staubach, "NFS 471 Version 3 Protocol Specification", RFC 1813, 472 DOI 10.17487/RFC1813, June 1995, 473 . 475 [RFC3232] Reynolds, J., Ed., "Assigned Numbers: RFC 1700 is Replaced 476 by an On-line Database", RFC 3232, DOI 10.17487/RFC3232, 477 January 2002, . 479 [RFC5040] Recio, R., Metzler, B., Culley, P., Hilland, J., and D. 480 Garcia, "A Remote Direct Memory Access Protocol 481 Specification", RFC 5040, DOI 10.17487/RFC5040, October 482 2007, . 484 [RFC5041] Shah, H., Pinkerton, J., Recio, R., and P. Culley, "Direct 485 Data Placement over Reliable Transports", RFC 5041, 486 DOI 10.17487/RFC5041, October 2007, 487 . 489 [RFC5667] Talpey, T. and B. Callaghan, "Network File System (NFS) 490 Direct Data Placement", RFC 5667, DOI 10.17487/RFC5667, 491 January 2010, . 493 Author's Address 495 Charles Lever (editor) 496 Oracle Corporation 497 1015 Granger Avenue 498 Ann Arbor, MI 48104 499 USA 501 Phone: +1 734 274 2396 502 Email: chuck.lever@oracle.com