S. Bailey (Sandburst) Internet-draft Expires: July 2002 The Architecture of Direct Data Placement (DDP) And Remote Direct Memory Access (RDMA) On Internet Protocols draft-bailey-roi-ddp-rdma-arch-00 Status of this Memo This document is an Internet-Draft and is in full conformance with all provisions of Section 10 of RFC2026. Internet-Drafts are working documents of the Internet Engineering Task Force (IETF), its areas, and its working groups. Note that other groups may also distribute working documents as Internet- Drafts. Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress." The list of current Internet-Drafts can be accessed at http://www.ietf.org/ietf/1id-abstracts.txt The list of Internet-Draft Shadow Directories can be accessed at http://www.ietf.org/shadow.html. Copyright Notice Copyright (C) The Internet Society (2001). All Rights Reserved. Abstract This document defines an abstract architecture for Direct Data Placement (DDP) and Remote Direct Memory Access (RDMA) protocols to run on Internet Protocol-suite transport protocols. This architecture does not necessarily reflect the proper way to implement such protocols, but is, rather, a descriptive tool for defining and understanding the protocols. Bailey Expires July 2002 [Page 1] Internet-Draft DDP & RDMA Architecture 4 February 2002 Table Of Contents 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . 2 2. Direct Data Placement (DDP) Architecture . . . . . . . . . 2 2.1. Transport Operations . . . . . . . . . . . . . . . . . . . 4 2.2. DDP Operations . . . . . . . . . . . . . . . . . . . . . . 5 2.3. Transport Characterstics In DDP . . . . . . . . . . . . . 8 3. Remote Direct Memory Access (RDMA) Protocol Architecture . 9 3.1. RDMA Operations . . . . . . . . . . . . . . . . . . . . . 10 3.2. Transport Characterstics In RDMA . . . . . . . . . . . . . 12 4. Security Considerations . . . . . . . . . . . . . . . . . 13 5. IANA Considerations . . . . . . . . . . . . . . . . . . . 13 Author's Address . . . . . . . . . . . . . . . . . . . . . 13 Full Copyright Statement . . . . . . . . . . . . . . . . . 14 1. Introduction This document defines an abstract architecture for Direct Data Placement (DDP) and Remote Direct Memory Access (RDMA) protocols to run on Internet Protocol-suite transport protocols. This architecture does not necessarily reflect the proper way to implement such protocols, but is, rather, a descriptive tool for defining and understanding the protocols. The first section describes the architecture of DDP protocols, including assumptions of the transports on which DDP is built. The second section describes the architecture of RDMA protocols layered on top of DDP. 2. Direct Data Placement (DDP) Architecture The central idea of general-purpose DDP is that a data sender will supplement the data it sends with placement information that allows the receiver's network interface (NI) to place the data directly at its final destination without any copying. DDP can be used to steer received data to its final destination for any ULP without requiring ULP-specific behavior in the NI for each different ULP. Data sent with DDP information is said to be `DDP-decorated'. The central component of the DDP architecture is the `buffer', which is an object with beginning and ending addresses, and a method (set()) to set the value of an octet at an address. In many cases, a buffer corresponds directly to a portion of host memory. However, DDP does not depend on this---a buffer could be a disk file, or anything else that can be viewed as an addressable collection of octets. Abstractly, a buffer provides the interface: Bailey Expires July 2002 [Page 2] Internet-Draft DDP & RDMA Architecture 4 February 2002 typedef struct { const address_t start; const address_t end; void set(address_t a, uint8_t v); } buffer_t; The protocol layering and in-line data flow of DDP is: Client Protocol (e.g. ULP or RDMA) | ^ undecorated messages | | undecorated messages DDP-decorated messages | | DDP-decorated message reception v | indications DDP ^ | transport messages v Transport (e.g. SCTP, DCP) ^ | IP datagrams v . . . In addition to in-line data flow, the client protocol registers buffers with DDP, and DDP performs buffer update (set()) operations as a result of receiving DDP-decorated messages. Undecorated messages correspond directly to messages of the underlying transport, but must still be distinguished from DDP- decorated messages in some way. DDP-decorated messages may be split into multiple, smaller DDP- decorated messages each in a separate transport message. However, if the transport is unreliable or unordered, DDP-decorated messages split across transport messages may or may not provide useful behavior, in the same way as splitting regular, undecorated messages across unreliable or unordered transport messages may or may not provide useful behavior. In other words, the same considerations apply to building client protocols on different types of transports with or without the use of DDP. A DDP-decorated message split across transport messages looks like: Bailey Expires July 2002 [Page 3] Internet-Draft DDP & RDMA Architecture 4 February 2002 DDP-decorated message: Transport messages: stag=s, offset=o, message 1: notify=y, id=i |type=ddp | message= |stag=s | |aabbccddee|-------. |offset=o | ~ ... ~----. \ |notify=n | |vvwwxxyyzz|-. \ \ |id=? | | \ `--->|aabbccddee| | \ ~ ... ~ | +----->|iijjkkllmm| | | + | message 2: \ | |type=ddp | \ | |stag=s | \ + |offset=o+n| \ \ |notify=y | \ \ |id=i | \ `-->|nnooppqqrr| \ ~ ... ~ `---->|vvwwxxyyzz| Although this picture suggests that DDP decoration information is carried in-line with the message payload, components of the DDP decoration may also be in transport-specific fields, or derived from transport-specific control information if the transport permits. 2.1. Transport Operations For the purposes of this architecture, the transport provides: void xpt_send(socket_t s, message_t m); message_t xpt_recv(socket_t s); msize_t xpt_max_msize(socket_t s); socket_t a transport address, including IP addresses, ports and other transport-specific identifiers. message_t a string of octets. msize_t (unsigned integer) Bailey Expires July 2002 [Page 4] Internet-Draft DDP & RDMA Architecture 4 February 2002 a message size. xpt_send(socket_t s, message_t m) send a transport message. xpt_recv(socket_t s) receive a transport message. xpt_max_msize(socket_t s) get the current maximum transport message size. Corresponds, roughly, to the current path Maximum Transfer Unit (PMTU), adjusted by underlying protocol overheads. Real implementations of xpt_send() and xpt_recv() typically return error indications, but that is not relevant to this architecture. 2.2. DDP Operations The DDP layer provides: void ddp_send(socket_t s, message_t m); void ddp_send_ddp(socket_t s, message_t m, ddp_addr_t d, ddp_notify_t n); ddp_recv_t ddp_recv(socket_t s); bdesc_t ddp_register(socket_t s, buffer_t b); void ddp_deregister(bhand_t bh); msizes_t ddp_max_msizes(socket_t s); ddp_addr_t the buffer address portion of a DDP-decoration: typedef struct { stag_t stag; address_t offset; } ddp_addr_t; stag_t (unsigned integer) a steering tag. A stag_t identifies the destination buffer for DDP-decorated messages. stag_ts are generated when the buffer is registered, communicated to the sender by some client protocol convention and inserted in DDP-decorated Bailey Expires July 2002 [Page 5] Internet-Draft DDP & RDMA Architecture 4 February 2002 messages. stag_t values in this DDP architecture are assumed to be completely opaque to the client protocol, and implementation-dependent. However, particular implementations, such as DDP on a multicast transport (see below), may provide the buffer holder some control in selecting stag_ts. ddp_notify_t the notification portion of a DDP-decoration: typedef struct { bool notify; ddp_msg_id_t i; } ddp_notify_t; ddp_msg_id_t (unsigned integer) a DDP-decorated message identifier. msg_id_ts are chosen by the DDP-decorated message receiver (buffer holder), communicated to the sender by some client protocol convention and inserted in DDP-decorated messages. Whether a message reception indication is requested for a DDP-decorated message is a matter of client protocol convention. Unlike stag_ts, the structure of msg_id_ts is opaque to DDP, and therefore, completely in the hands of the client protocol. bdesc_t a description of a registered buffer: typedef struct { bhand_t bh; ddp_addr_t a; } bdesc_t; `a.offset' is the starting offset of the registered buffer, which may have no relationship to the `start' or `end' addresses of that buffer. However, particular implemenations, such as DDP on a multicast transport (see below), may allow some client protocol control over the starting offset. bhand_t an opaque buffer handle used to unregister a buffer. ddp_recv_t Bailey Expires July 2002 [Page 6] Internet-Draft DDP & RDMA Architecture 4 February 2002 an undecorated message, a DDP-decorated message reception indication, or a DDP-decorated message reception error: typedef union { message_t m; ddp_msg_id_t i; ddp_err_t e; } ddp_recv_t; ddp_err_t indicates an error while receiving a DDP-decorated message, typically `offset' out of bounds, or `stag' is not registered to the socket. msizes_t The maximum undecorated and DDP-decorated messages that fit in a single transport message: typedef struct { msize_t max_undec; msize_t max_dec; } msizes_t; ddp_send(socket_t s, message_t m) send an undecorated message. ddp_send_ddp(socket_t s, message_t m, ddp_addr_t d, ddp_notify_t n) send a DDP-decorated message. ddp_recv(socket_t s) get the next received undecorated message, DDP-decorated message reception indication, or DDP-decorated message error. ddp_register(socket_t s, buffer_t b) register a buffer for DDP on a socket. The same buffer may be registered multiple times on the same or different sockets. Different buffers may also refer to portions of the same underlying addressable object (buffer aliasing). ddp_deregister(bhand_t bh) Bailey Expires July 2002 [Page 7] Internet-Draft DDP & RDMA Architecture 4 February 2002 unregister a buffer from a socket. ddp_max_msizes(socket_t s) get the current maximum undecorated and DDP-decorated message sizes that will fit in a single transport message. 2.3. Transport Characterstics In DDP Certain characteristics of the transport on which DDP is mapped determine the nature of the service provided to client protocols. Specifically, transports are: o reliable or unreliable, o ordered or unordered, o single source or multisource, o single destination or multidestination (multicast or anycast). Some transports support several combinations of these characteristics. For example, SCTP is reliable, single source, single destination (point-to-point) and supports both ordered and unordered modes. In general, these transport characteristics equally affect transport and DDP-decorated message delivery. However, there are several issues specific to DDP-decorated messages. A key component of DDP, is how operations on the receiving side: o set()s, o undecorated messages, and o DDP-decorated message reception indications are ordered among themselves, and how they relate to corresponding operations on the sending side. These relationships depend upon the characteristics of the underlying transport in a way which is defined by the DDP protocol. For example, if the transport is unreliable and unordered, the DDP protocol might specify that the client protocol is subject to the consequences of transport messages being lost or duplicated, rather requiring different characteristics be presented to the client protocol. Multidestination data delivery is the other transport Bailey Expires July 2002 [Page 8] Internet-Draft DDP & RDMA Architecture 4 February 2002 characteristic which may require specific consideration in a DDP protocol. As mentioned above, the basic DDP model assumes that buffer address values returned by ddp_register() are opaque to the client protocol, and can be implementation dependent. The most natural way to map DDP to a multidestination transport is to require all receivers produce the same buffer address when registering a multidestination destination buffer. Restriction of the DDP model to accomodate multiple destinations involves engineering tradeoffs comparable to those of providing non-DDP multidestination transport capability. 3. Remote Direct Memory Access (RDMA) Protocol Architecture Remote Direct Memory Access (RDMA) extends the capabilities of DDP with the ability to read from buffers registered to a socket (RDMA Read). This allows a client protocol to perform arbitrary, bidirectional data movement without involving the remote client protocol. When RDMA is implemented in the NI, arbitrary data movement can be performed without involving the remote host CPU at all. In addition, RDMA protocols usually specify a transport-independent undecorated message service (Send) with characteristics which are both very efficient to implement in an NI, and convenient for client protocols. The RDMA architecture is patterned after the traditional model for device programming, where the client requests an operation using Send-like actions (programmed I/O), the server performs the necessary data transfers for the operation (DMA reads and writes), and notifies the client of completion. The programmed I/O+DMA model efficiently supports a high degree of concurrency and flexibility for both the client and server, even when operations have a wide range of intrinsic latencies. RDMA is implemented as a client protocol on top of DDP: Bailey Expires July 2002 [Page 9] Internet-Draft DDP & RDMA Architecture 4 February 2002 Client Protocol | ^ Sends | | Sends RDMA Read Requests | | RDMA Read Completion indications RDMA Writes v | RDMA Write Completion indications RDMA | ^ undecorated messages | | undecorated messages DDP-decorated messages | | DDP-decorated message reception v | indications DDP ^ | transport messages v . . . In addition to in-line data flow, read (get()) and update (set()) operations are performed on buffers registered with RDMA as a result of RDMA Read Requests and RDMA Writes, respectively. An RDMA `buffer' extends a DDP buffer with a get() operation that retrieves the value of the octet at address `a': typedef struct { const address_t start; const address_t end; void set(address_t a, uint8_t v); uint8_t get(address_t a); } buffer_t; 3.1. RDMA Operations The RDMA layer provides: void rdma_send(socket_t s, message_t m); void rdma_write(socket_t s, message_t m, ddp_addr_t d, rdma_notify_t n); void rdma_read(socket_t s, ddp_addr_t s, ddp_addr_t d); rdma_recv_t rdma_recv(socket_t s); bdesc_t rdma_register(socket_t s, buffer_t b, bmode_t mode); void rdma_deregister(bhand_t bh); msizes_t rdma_max_msizes(socket_t s); Although, for clarity, these data transfer interfaces are synchronous, rdma_read() and possibly rdma_send() (in the presence Bailey Expires July 2002 [Page 10] Internet-Draft DDP & RDMA Architecture 4 February 2002 of Send flow control), can require an arbitrary amount of time to complete. To express the full concurrency and interleaving of RDMA data transfer, these interfaces are also defined to be multithreaded. For example, a client protocol may perform an rdma_send(), while an rdma_read() operation is in progress. rdma_notify_t RDMA Write notification information: typedef struct { bool notify; rdma_write_id_t i; } rdma_notify_t; identical in function to ddp_notify_t, except that the type rdma_write_id_t may not be equivalent to ddp_msg_id_t. rdma_write_id_t (unsigned integer) an RDMA Write identifier. rdma_recv_t a Send message, an RDMA Write completion identifier, or an RDMA error: typedef union { message_t m; rdma_write_id_t i; rdma_err_t e; } rdma_recv_t; rdma_err_t an RDMA protocol error indication. RDMA errors include buffer addressing errors corresponding to ddp_err_ts, and buffer protection violations (e.g. RDMA Writing a buffer only registered for reading). bmode_t buffer registration mode (permissions). Any combination of permitting RDMA Read (BMODE_READ) and RDMA Write (BMODE_WRITE) operations. rdma_send(socket_t s, message_t m) Bailey Expires July 2002 [Page 11] Internet-Draft DDP & RDMA Architecture 4 February 2002 Send a message. rdma_write(socket_t s, message_t m, ddp_addr_t d, rdma_notify_t n) RDMA Write to remote buffer address d. rdma_read(socket_t s, ddp_addr_t s, ddp_addr_t d) RDMA Read from remote buffer address s to local buffer address d. rdma_recv(socket_t s); get the next received Send message, RDMA Write completion identifier, or RDMA error. rdma_register(socket_t s, buffer_t b, bmode_t mode) register a buffer for RDMA on a socket (for read access, write access or both). As with DDP, the same buffer may be registered multiple times on the same or different sockets, and different buffers may refer to portions of the same underlying addressable object. rdma_deregister(bhand_t bh) unregister a buffer from a socket. rdma_max_msizes(socket_t s) get the current maximum Send (max_undec) and RDMA Read or Write (max_dec) operations that will fit in a single transport message. The values returned by rdma_max_msizes() are closely related to the values returned by ddp_max_msizes(), but may not be equal. 3.2. Transport Characterstics In RDMA As with DDP, RDMA can be used on transports with a variety of different characteristics that manifest themselves directly in the service provided by RDMA. Like DDP, an RDMA protocol must specify how: o set()s, o get()s, Bailey Expires July 2002 [Page 12] Internet-Draft DDP & RDMA Architecture 4 February 2002 o Send messages, and o RDMA Read completions are ordered among themselves and how they relate to corresponding operations on the remote peer(s). These relationships are likely to be a function of the underlying transport characteristics. There are some additional characteristics of RDMA which may translate poorly to unreliable or multipoint transports due to attendent complexities in managing endpoint state: o Send flow control o RDMA Read These difficulties can be overcome by placing restrictions on the service provided by RDMA. However, many RDMA clients, especially those that separate data transfer and application logic concerns, are likely to depend upon capabilities only provided by RDMA on a point-to-point, reliable transport. 4. Security Considerations Security considerations are not addressed in this document. Any security considerations resulting from the use of DDP or RDMA must be addressed in the relevant standards. 5. IANA Considerations IANA considerations are not addressed in by this document. Any IANA considerations resulting from the use of DDP or DMA must be addressed in the relevant standards. Author's Address Stephen Bailey Sandburst Corporation 600 Federal Street Andover, MA 01810 USA Email: steph@sandburst.com Bailey Expires July 2002 [Page 13] Internet-Draft DDP & RDMA Architecture 4 February 2002 Full Copyright Statement Copyright (C) The Internet Society (2001). All Rights Reserved. This document and translations of it may be copied and furnished to others, and derivative works that comment on or otherwise explain it or assist in its implementation may be prepared, copied, published and distributed, in whole or in part, without restriction of any kind, provided that the above copyright notice and this paragraph are included on all such copies and derivative works. However, this document itself may not be modified in any way, such as by removing the copyright notice or references to the Internet Society or other Internet organizations, except as needed for the purpose of developing Internet standards in which case the procedures for copyrights defined in the Internet Standards process must be followed, or as required to translate it into languages other than English. The limited permissions granted above are perpetual and will not be revoked by the Internet Society or its successors or assigns. This document and the information contained herein is provided on an "AS IS" basis and THE INTERNET SOCIETY AND THE INTERNET ENGINEERING TASK FORCE DISCLAIMS ALL WARRANTIES, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Bailey Expires July 2002 [Page 14]