idnits 2.17.1 draft-ietf-nfsv4-nfsdirect-01.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- ** It looks like you're using RFC 3978 boilerplate. You should update this to the boilerplate described in the IETF Trust License Policy document (see https://trustee.ietf.org/license-info), which is required now. -- Found old boilerplate from RFC 3979, Section 5, paragraph 3 on line 396. ** Found boilerplate matching RFC 3978, Section 5.4, paragraph 1 (on line 364), which is fine, but *also* found old RFC 2026, Section 10.4C, paragraph 1 text on line 34. ** The document seems to lack an RFC 3978 Section 5.1 IPR Disclosure Acknowledgement -- however, there's a paragraph with a matching beginning. Boilerplate error? ** This document has an original RFC 3978 Section 5.4 Copyright Line, instead of the newer IETF Trust Copyright according to RFC 4748. ** The document seems to lack an RFC 3978 Section 5.4 Reference to BCP 78. ** The document seems to lack an RFC 3978 Section 5.5 (updated by RFC 4748) Disclaimer -- however, there's a paragraph with a matching beginning. Boilerplate error? ** The document seems to lack an RFC 3979 Section 5, para. 1 IPR Disclosure Acknowledgement -- however, there's a paragraph with a matching beginning. Boilerplate error? ** The document seems to lack an RFC 3979 Section 5, para. 2 IPR Disclosure Acknowledgement -- however, there's a paragraph with a matching beginning. Boilerplate error? Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- ** The document seems to lack a 1id_guidelines paragraph about 6 months document validity -- however, there's a paragraph with a matching beginning. Boilerplate error? ** The document seems to lack a 1id_guidelines paragraph about the list of Shadow Directories. == No 'Intended status' indicated for this document; assuming Proposed Standard Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- No issues found here. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the RFC 3978 Section 5.4 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.) -- Couldn't find a document date in the document -- date freshness check skipped. -- Found something which looks like a code comment -- if you have code sections in the document, please surround them with '' and '' lines. 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: 'RFC1831' is defined on line 301, but no explicit reference was found in the text == Unused Reference: 'RFC1832' is defined on line 307, but no explicit reference was found in the text == Unused Reference: 'RFC1094' is defined on line 312, but no explicit reference was found in the text == Unused Reference: 'RFC1813' is defined on line 317, but no explicit reference was found in the text == Unused Reference: 'RFC3530' is defined on line 323, but no explicit reference was found in the text ** Obsolete normative reference: RFC 1831 (Obsoleted by RFC 5531) ** Obsolete normative reference: RFC 1832 (Obsoleted by RFC 4506) ** Downref: Normative reference to an Informational RFC: RFC 1094 ** Downref: Normative reference to an Informational RFC: RFC 1813 ** Obsolete normative reference: RFC 3530 (Obsoleted by RFC 7530) == Outdated reference: A later version (-02) exists of draft-ietf-nfsv4-sess-01 Summary: 15 errors (**), 0 flaws (~~), 8 warnings (==), 4 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Internet-Draft Brent Callaghan 3 Expires: August 2005 Tom Talpey 5 Document: draft-ietf-nfsv4-nfsdirect-01 February, 2005 7 NFS Direct Data Placement 9 Status of this Memo 11 By submitting this Internet-Draft, I certify that any applicable 12 patent or other IPR claims of which I am aware have been disclosed, 13 or will be disclosed, and any of which I become aware will be dis- 14 closed, in accordance with RFC 3668. 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 22 months and may be updated, replaced, or obsoleted by other docu- 23 ments at any time. It is inappropriate to use Internet-Drafts as 24 reference material or to cite them other than as "work in 25 progress." 27 The list of current Internet-Drafts can be accessed at 28 http://www.ietf.org/ietf/1id-abstracts.txt The list of Inter- 29 net-Draft Shadow Directories can be accessed at 30 http://www.ietf.org/shadow.html. 32 Copyright Notice 34 Copyright (C) The Internet Society (2005). All Rights Reserved. 36 Abstract 38 The RDMA transport for ONC RPC provides direct data placement for NFS 39 data. Direct data placement not only reduces the amount of data that 40 needs to be copied in an NFS call, but allows much of the data 41 movement over the network to be implemented in RDMA hardware. This 42 draft describes the use of direct data placement by means of server- 43 initiated RDMA operations into client-supplied buffers in a Chunk 44 list for implementations of NFS versions 2, 3, and 4 over an RDMA 45 transport. 47 Table of Contents 49 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 2 50 2. RDMA Read List . . . . . . . . . . . . . . . . . . . . . . . 2 51 3. RDMA Write List . . . . . . . . . . . . . . . . . . . . . . 3 52 4. NFS Versions 2 and 3 Mapping . . . . . . . . . . . . . . . . 4 53 5. NFS Version 4 Mapping . . . . . . . . . . . . . . . . . . . 5 54 6. Security . . . . . . . . . . . . . . . . . . . . . . . . . . 7 55 7. IANA Considerations . . . . . . . . . . . . . . . . . . . . 7 56 8. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . 7 57 9. Normative References . . . . . . . . . . . . . . . . . . . . 7 58 10. Informative References . . . . . . . . . . . . . . . . . . 8 59 11. Authors' Addresses . . . . . . . . . . . . . . . . . . . . 9 60 12. Full Copyright Statement . . . . . . . . . . . . . . . . . 9 61 Acknowledgement . . . . . . . . . . . . . . . . . . . . . . . 10 63 1. Introduction 65 The RDMA Transport for ONC RPC [RPCRDMA] allows an RPC client 66 application to post buffers in a Chunk list for specific arguments 67 and results from an RPC call. The RDMA transport header conveys this 68 list of client buffer addresses to the server where the application 69 can associate them with client data and use RDMA operations to 70 transfer the results directly to and from the posted buffers on the 71 client. The client and server must agree on a consistent mapping of 72 posted buffers to RPC. This document details the mapping for each 73 version of the NFS protocol. 75 2. RDMA Read List 77 The RDMA Read list, in the RDMA transport header, allows an RPC 78 client to marshall RPC call data selectively. Large chunks of data, 79 such as the file data of an NFS WRITE request, can be referenced from 80 the RDMA Read list and be moved efficiently and direct-placed by an 81 RDMA READ operation initiated by the server. 83 The process of identifying these chunks for the RDMA Read list can be 84 implemented entirely within the RPC layer. It is transparent to the 85 upper-level protocol, such as NFS. For instance, the file data 86 portion of an NFS WRITE request can be selected as an RDMA "chunk" 87 within the XDR marshalling code of RPC based on a size criterion, 88 independently of the NFS protocol layer. The XDR unmarshalling on the 89 receiving system can identify the correspondence between Read chunks 90 and protocol elements via the XDR position value encoded in the Read 91 chunk entry. 93 However, the implementation of the RDMA Write list requires some help 94 from the NFS protocol layer to identify the candidate chunks. Since 95 there is no simple XDR position value to unambiguously label Write 96 chunks, both client and server must agree on a mapping of write chunk 97 entries to protocol elements. 99 3. RDMA Write List 101 The RDMA Write list, in the RDMA transport header, allows the client 102 to post one or more buffers into which the server will RDMA Write 103 designated result chunks directly. If the client sends a null write 104 list, then results from the RPC call will be returned as either an 105 in-line reply, as chunks in an RDMA Read list of server-posted 106 buffers, or in a client-posted reply buffer. 108 Each posted buffer in a Write list is represented as an array of 109 memory segments. This allows the client some flexibility in 110 submitting discontiguous memory segments into which the server will 111 scatter the result. Each segment is described by a triplet 112 consisting of the segment handle or steering tag (STag), segment 113 length, and memory address or offset. 115 struct xdr_rdma_segment { 116 uint32 handle; /* Registered memory handle */ 117 uint32 length; /* Length of the chunk in bytes */ 118 uint64 offset; /* Chunk virtual address or offset */ 119 }; 121 struct xdr_write_chunk { 122 struct xdr_rdma_segment target<>; 123 }; 124 struct xdr_write_list { 125 struct xdr_write_chunk entry; 126 struct xdr_write_list *next; 127 }; 129 The sum of the segment lengths yields the total size of the buffer, 130 which must be large enough to accept the result. If the buffer is 131 too small, the server must return an XDR encode error. The server 132 must return the result data for a posted buffer by progressively 133 filling its segments, perhaps leaving some trailing segments unfilled 134 or partially full if the size of the result is less than the total 135 size of the buffer segments. 137 The server returns the RDMA Write list to the client with the segment 138 length fields overwritten to indicate the amount of data RDMA Written 139 to each segment. Results returned by direct placement must not be 140 returned by other methods, e.g. by read chunk list or in-line. 142 The RDMA Write list allows the client to provide multiple result 143 buffers - each buffer must map to a specific result in the reply. The 144 NFS client and server implementations must agree on the mapping of 145 results to buffers for each RPC procedure. The following sections 146 describe this mapping for versions of the NFS protocol. 148 Through the use of RDMA Write lists in NFS requests, it is not 149 necessary to employ the RDMA Read lists in the NFS replies, as 150 described in the RPC/RDMA protocol. This enables more efficient 151 operation, by avoiding the need for the server to expose buffers for 152 RDMA, and also avoiding "RDMA_DONE" messages. 154 4. NFS Versions 2 and 3 Mapping 156 A single RDMA Write list entry may be posted by the client to receive 157 either the opaque file data from a READ request or the pathname from 158 a READLINK request. The server will ignore a Write list for any 159 other NFS procedure, as well as any Write list entries beyond the 160 first in the list. 162 Similarly, a single RDMA Read list entry may be posted by the client 163 to supply the opaque file data for a WRITE request or the pathname 164 for a SYMLINK request. The server will ignore any Read list for 165 other NFS procedures, as well as additional Read list entries beyond 166 the first in the list. 168 Because there are no NFS version 2 or 3 requests that transfer bulk 169 data in both directions, it is not necessary to post requests 170 containing both Write and Read lists. Any unneeded Read or Write 171 lists are ignored by the server. 173 In the case where the outgoing request or expected incoming reply is 174 larger than the maximum size supported on the connection, it is 175 possible for the RPC layer to post the entire message or result in a 176 special "RDMA_NOMSG" message type which is transferred entirely by 177 RDMA. This is implemented in RPC, below NFS and therefore has no 178 effect on the message contents. 180 Non-RDMA (inline) WRITE transfers may optionally employ the 181 "RDMA_MSGP" padding method described in the RPC/RDMA protocol, if the 182 appropriate value for the server is known to the client. Padding 183 allows the opaque file data to arrive at the server in an aligned 184 fashion, which may improve server performance. 186 The NFS version 2 and 3 protocols are frequently limited in practice 187 to requests containing less than or equal to 8 kilobytes and 32 188 kilobytes of data, respectively. In these cases, it is often 189 practical to support basic operation without employing a 190 configuration exchange as discussed in [RPCRDMA]. The server can 191 post buffers large enough to receive the largest possible incoming 192 message (approximately 12KB/36KB would be vastly sufficient in the 193 above cases), and the client can post buffers large enough to receive 194 replies based on the "rsize" it is using to the server. Flow control 195 is handled dynamically by the RPC RDMA protocol, and write padding is 196 optional and therefore may remain unused. 198 Alternatively, if the server is administratively configured to values 199 appropriate for all its clients, the same assurance of 200 interoperability within the domain can be made. 202 The use of a configuration protocol with NFS v2 and v3 is therefore 203 optional. Employing a configuration exchange may allow some advantage 204 to server resource management through accurately sizing buffers, 205 enabling the server to know exactly how many RDMA Reads may be in 206 progress at once on the client connection, and enabling client write 207 padding which may be desirable for certain servers when RDMA Read is 208 impractical. 210 5. NFS Version 4 Mapping 212 This specification applies to the first minor version of NFS version 213 4 (NFSv4.0) and any subsequent minor versions that do not override 214 this mapping. 216 The Write list will be considered only for the COMPOUND procedure. 217 This procedure returns results from a sequence of operations. Only 218 the opaque file data from an NFS READ operation, and the pathname 219 from a READLINK operation will utilize entries from the Write list. 221 If there is no Write list, i.e. the list is null, then any READ or 222 READLINK operations in the COMPOUND must return their data either in- 223 line or via RDMA READ (using the Read list). 225 The first entry in the Write list must be used by the first READ or 226 READLINK in the request. The next Write list entry by the by the 227 next READ or READLINK, and so on. If there are more READ or READLINK 228 operations than Write list entries, then any remaining operations 229 must return their results in-line or via the Read list. 231 If a Write list entry is presented, then the corresponding READ or 232 READLINK must return its data via an RDMA WRITE to the buffer 233 indicated by the Write list entry. If the Write list entry has zero 234 RDMA segments, or if the total size of the segments is zero, then the 235 corresponding READ or READLINK operation must return its result in- 236 line or via Read list. 238 The following example shows an RDMA Write list with three posted 239 buffers A, B, and C. The designated operations in the compound 240 request, READ and READLINK, consume the posted buffers by writing 241 their results back to each buffer. 243 RDMA Write list: 245 A --> B --> C 247 Compound request: 249 PUTFH LOOKUP READ PUTFH LOOKUP READLINK PUTFH LOOKUP READ 250 | | | 251 v v v 252 A B C 254 If the client does not want to have the READLINK result returned 255 directly, then it provides a zero length array of segment triplets 256 for buffer B or sets the values in the segment triplet for buffer B 257 to zeros so that the READLINK result will be returned in-line. 259 The situation is similar for RDMA Read lists and applies to the 260 NFSv4.0 WRITE and SYMLINK procedures as for v3. Additionally, inline 261 segments too large to fit in posted buffers may be transferred in 262 special "RDMA_NOMSG" messages. 264 Non-RDMA (inline) WRITE transfers may optionally employ the 265 "RDMA_MSGP" padding method described in the RPC/RDMA protocol, if the 266 appropriate value for the server is known to the client. Padding 267 allows the opaque file data to arrive at the server in an aligned 268 fashion, which may improve server performance. In order to ensure 269 accurate alignment for all data, it is likely that the client will 270 restrict its use of optional padding to COMPOUND requests containing 271 only a single WRITE operation. 273 Unlike NFS versions 2 and 3, the maximum size of an NFS version 4 274 COMPOUND is unbounded, even when RDMA chunks are in use. While it 275 might appear that a configuration protocol exchange (as described in 276 [RPCRDMA]) would help, in fact the layering issues involved in 277 building COMPOUNDs by NFS make such a mechanism unworkable. Instead, 278 an extension to NFS version 4 supporting a more comprehensive 279 exchange of upper layer (NFSv4) parameters is proposed in 280 [NFSSESSIONS]. This proposal also addresses other use of the sizes, 281 such as in the server's response cache. 283 6. Security 285 The RDMA transport for ONC RPC supports RPCSEC_GSS security as well 286 as link-level security. The use of RDMA Write to return RPC results 287 does not affect ONC RPC security. 289 7. IANA Considerations 291 NFS use of direct data placement introduces no new IANA 292 considerations. 294 8. Acknowledgements 296 The authors would like to thank Dave Noveck and Chet Juszczak for 297 their contributions to this document. 299 9. Normative References 301 [RFC1831] 302 R. Srinivasan, "RPC: Remote Procedure Call Protocol Specification 303 Version 2", 304 Standards Track RFC, 305 http://www.ietf.org/rfc/rfc1831.txt 307 [RFC1832] 308 R. Srinivasan, "XDR: External Data Representation Standard", 309 Standards Track RFC, 310 http://www.ietf.org/rfc/rfc1832.txt 312 [RFC1094] 313 "NFS: Network File System Protocol Specification", 314 (NFS version 2) Informational RFC, 315 http://www.ietf.org/rfc/rfc1094.txt 317 [RFC1813] 318 B. Callaghan, B. Pawlowski, P. Staubach, "NFS Version 3 Protocol 319 Specification", 320 Informational RFC, 321 http://www.ietf.org/rfc/rfc1813.txt 323 [RFC3530] 324 S. Shepler, B. Callaghan, D. Robinson, R. Thurlow, C. Beame, M. 325 Eisler, D. Noveck, "NFS version 4 Protocol", 326 Standards Track RFC, 327 http://www.ietf.org/rfc/rfc3530.txt 329 10. Informative References 331 [RPCRDMA] 332 B. Callaghan, T. Talpey, "RDMA Transport for ONC RPC" 333 Internet Draft Work in Progress, 334 http://www.ietf.org/internet-drafts/ 336 [NFSSESSIONS] 337 T. Talpey, S. Shepler, J. Bauman, "NFSv4 Session Extensions" 338 Internet Draft Work in Progress, 339 http://www.ietf.org/internet-drafts/ 340 draft-ietf-nfsv4-sess-01.txt 342 11. Authors' Addresses 344 Brent Callaghan 345 1614 Montalto Dr. 346 Mountain View, California 94040 USA 348 Phone: +1 650 968 2333 349 EMail: brent.callaghan@gmail.com 351 Tom Talpey 352 Network Appliance, Inc. 353 375 Totten Pond Road 354 Waltham, MA 02451 USA 356 Phone: +1 781 768 5329 357 EMail: thomas.talpey@netapp.com 359 12. Full Copyright Statement 361 Copyright (C) The Internet Society (2005). This document is sub- 362 ject to the rights, licenses and restrictions contained in BCP 78 363 and except as set forth therein, the authors retain all their 364 rights. 366 This document and the information contained herein are provided on 367 an "AS IS" basis and THE CONTRIBUTOR, THE ORGANIZATION HE/SHE REP- 368 RESENTS OR IS SPONSORED BY (IF ANY), THE INTERNET SOCIETY AND THE 369 INTERNET ENGINEERING TASK FORCE DISCLAIM ALL WARRANTIES, EXPRESS OR 370 IMPLIED, INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF 371 THE INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED 372 WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. 374 Intellectual Property 376 The IETF takes no position regarding the validity or scope of any 377 Intellectual Property Rights or other rights that might be claimed 378 to pertain to the implementation or use of the technology described 379 in this document or the extent to which any license under such 380 rights might or might not be available; nor does it represent that 381 it has made any independent effort to identify any such rights. 382 Information on the procedures with respect to rights in RFC docu- 383 ments can be found in BCP 78 and BCP 79. 385 Copies of IPR disclosures made to the IETF Secretariat and any 386 assurances of licenses to be made available, or the result of an 387 attempt made to obtain a general license or permission for the use 388 of such proprietary rights by implementers or users of this speci- 389 fication can be obtained from the IETF on-line IPR repository at 390 http://www.ietf.org/ipr. 392 The IETF invites any interested party to bring to its attention any 393 copyrights, patents or patent applications, or other proprietary 394 rights that may cover technology that may be required to implement 395 this standard. Please address the information to the IETF at ietf- 396 ipr@ietf.org. 398 Acknowledgement 400 Funding for the RFC Editor function is currently provided by the 401 Internet Society.