Allyn Romanow      (Cisco)
Internet-Draft                                Jeff Mogul            (HP)
Expires: August 2003                          Tom Talpey        (NetApp)
                                              Stephen Bailey (Sandburst)

                     RDMA over IP Problem Statement

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-

     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

     The list of current Internet-Drafts can be accessed at

     The list of Internet-Draft Shadow Directories can be accessed at

Copyright Notice

     Copyright (C) The Internet Society (2003).  All Rights Reserved.


     This draft addresses an IP-based solution to the problem of high
     system costs due to network I/O copying in end-hosts at high
     speeds.  The problem is due to the high cost of memory bandwidth,
     and it can be substantially improved using "copy avoidance."  The
     high overhead has prevented TCP/IP from being used as an
     interconnection network.

Romanow, et al             Expires August 2003                  [Page 1]

Internet-Draft       RDMA Over IP Problem Statement        February 2003

Table Of Contents

     1.   Introduction . . . . . . . . . . . . . . . . . . . . . . .   2
     2.   The high cost of data movement operations in network I/O .   3
     2.1. Copy avoidance improves processing overhead  . . . . . . .   5
     3.   Memory bandwidth is the root cause of the problem  . . . .   6
     4.   High copy overhead is problematic for many key Internet
          applications . . . . . . . . . . . . . . . . . . . . . . .   7
     5.   Copy Avoidance Techniques  . . . . . . . . . . . . . . . .   9
     5.1. A Conceptual Framework: DDP and RDMA . . . . . . . . . . .  11
     6.   Security Considerations  . . . . . . . . . . . . . . . . .  11
     7.   Acknowledgements . . . . . . . . . . . . . . . . . . . . .  12
          References . . . . . . . . . . . . . . . . . . . . . . . .  12
          Authors' Addresses . . . . . . . . . . . . . . . . . . . .  17
          Full Copyright Statement . . . . . . . . . . . . . . . . .  17

1.  Introduction

     This draft considers the problem of high host processing overhead
     associated with network I/O that occurs under high speed
     conditions.  This problem is often referred to as the "I/O
     bottleneck" [CT90].  More specifically, the source of high overhead
     that is of interest here is data movement operations - copying.
     This issue is not be confused with TCP offload, which is not
     addressed here.  High speed refers to conditions where the network
     link speed is high relative to the bandwidths of the host CPU and
     memory.  With today's computer systems, one Gbits/s and over is
     considered high speed.

     High costs associated with copying are an issue primarily for large
     scale systems.  Although smaller systems such as rack-mounted PCs
     and small workstations would benefit from a reduction in copying
     overhead, the benefit to smaller machines will be primarily in the
     next few years as they scale in the amount of bandwidth they
     handle.  Today it is large system machines with high bandwidth
     feeds, usually multiprocessors and clusters, that are adversely
     affected by copying overhead.  Examples of such machines include
     all varieties of servers: database servers, storage servers,
     application servers for transaction processing, for e-commerce, and
     web serving, content distribution, video distribution, backups,
     data mining and decision support, and scientific computing.

     Note that such servers almost exclusively service many concurrent
     sessions (transport connections), which, in aggregate, are
     responsible for > 1 Gbits/s of communication.  Nonetheless, the
     cost of copying overhead for a particular load is the same whether
     from few or many sessions.

Romanow, et al             Expires August 2003                  [Page 2]

Internet-Draft       RDMA Over IP Problem Statement        February 2003

     The I/O bottleneck, and the role of data movement operations, have
     been widely studied in research and industry over the last
     approximately 14 years, and we draw freely on these results.
     Historically, the I/O bottleneck has received attention whenever
     new networking technology has substantially increased line rates -
     100 Mbits/s FDDI and Fast Ethernet, 155 Mbits/s ATM, 1 Gbits/s
     Ethernet.  In earlier speed transitions, the availability of memory
     bandwidth allowed the I/O bottleneck issue to be deferred.  Now
     however, this is no longer the case.  While the I/O problem is
     significant at 1 Gbits/s, it is the introduction of 10 Gbits/s
     Ethernet which is motivating an upsurge of activity in industry and
     research [DAFS, IB, VI, CGZ01, Ma02, MAF+02].

     Because of high overhead of end-host processing in current
     implementations, the TCP/IP protocol stack is not used for high
     speed transfer.  Instead, special purpose network fabrics, using a
     technology generally known as remote direct memory access (RDMA),
     have been developed and are widely used.  RDMA is a set of
     mechanisms that allow the network adapter, under control of the
     application, to steer data directly into and out of application
     buffers.  Examples of such interconnection fabrics include Fibre
     Channel [FIBRE] for block storage transfer, Virtual Interface
     Architecture [VI] for database clusters, Infiniband [IB], Compaq
     Servernet [SRVNET], Quadrics [QUAD] for System Area Networks.
     These link level technologies limit application scaling in both
     distance and size, meaning that the number of nodes cannot be
     arbitrarily large.

     This problem statement substantiates the claim that in network I/O
     processing, high overhead results from data movement operations,
     specifically copying; and that copy avoidance significantly
     decreases the processing overhead.  It describes when and why the
     high processing overheads occur, explains why the overhead is
     problematic, and points out which applications are most affected.

     In addition, this document introduces an architectural approach to
     solving the problem, which is developed in detail in [BT02].  It
     also discusses how the proposed technology may introduce security
     concerns and how they should be addressed.

2.  The high cost of data movement operations in network I/O

     A wealth of data from research and industry shows that copying is
     responsible for substantial amounts of processing overhead.  It
     further shows that even in carefully implemented systems,
     eliminating copies significantly reduces the overhead, as
     referenced below.

Romanow, et al             Expires August 2003                  [Page 3]

Internet-Draft       RDMA Over IP Problem Statement        February 2003

     Clark et al. [CJRS89] in 1989 shows that TCP [Po81] overhead
     processing is attributable to both operating system costs such as
     interrupts, context switches, process management, buffer
     management, timer management, and to the costs associated with
     processing individual bytes, specifically computing the checksum
     and moving data in memory.  They found moving data in memory is the
     more important of the costs, and their experiments show that memory
     bandwidth is the greatest source of limitation.  In the data
     presented [CJRS89], 64% of the measured microsecond overhead was
     attributable to data touching operations, and 48% was accounted for
     by copying.  The system measured Berkeley TCP on a Sun-3/60 using
     1460 Byte Ethernet packets.

     In a well-implemented system, copying can occur between the network
     interface and the kernel, and between the kernel and application
     buffers - two copies, each of which are two memory bus crossings -
     for read and write.  Although in certain circumstances it is
     possible to do better, usually two copies are required on receive.

     Subsequent work has consistently shown the same phenomenon as the
     earlier Clark study.  A number of studies report results that data-
     touching operations, checksumming and data movement, dominate the
     processing costs for messages longer than 128 Bytes [BS96, CGY01,
     Ch96, CJRS89, DAPP93, KP96].  For smaller sized messages, per-
     packet overheads dominate [KP96, CGY01].

     The percentage of overhead due to data-touching operations
     increases with packet size, since time spent on per-byte operations
     scales linearly with message size [KP96].  For example, Chu [Ch96]
     reported substantial per-byte latency costs as a percentage of
     total networking software costs for an MTU size packet on
     SPARCstation/20 running memory-to-memory TCP tests over networks
     with 3 different MTU sizes.  The percentage of total software costs
     attributable to per-byte operations were:

        1500 Byte Ethernet 18-25%
        4352 Byte FDDI     35-50%
        9180 Byte ATM      55-65%

     Although many studies report results for data-touching operations
     including checksumming and data movement together, much work has
     focused just on copying [BS96, B99, Ch96, TK95].  For example,
     [KP96] reports results that separate processing times for checksum
     from data movement operations.  For the 1500 Byte Ethernet size,
     20% of total processing overhead time is attributable to copying.
     The study used 2 DECstations 5000/200 connected by an FDDI network.
     (In this study checksum accounts for 30% of the processing time.)

Romanow, et al             Expires August 2003                  [Page 4]

Internet-Draft       RDMA Over IP Problem Statement        February 2003

2.1.  Copy avoidance improves processing overhead

     A number of studies show that eliminating copies substantially
     reduces overhead.  For example, results from copy-avoidance in the
     IO-Lite system [PDZ99], which aimed at improving web server
     performance, show a throughput increase of 43% over an optimized
     web server, and 137% improvement over an Apache server.  The system
     was implemented in a 4.4BSD derived UNIX kernel, and the
     experiments used a server system based on a 333MHz Pentium II PC
     connected to a switched 100 Mbits/s Fast Ethernet.

     There are many other examples where elimination of copying using a
     variety of different approaches showed significant improvement in
     system performance [CFF+94, DP93, EBBV95, KSZ95, TK95, Wa97].  We
     will discuss the results of one of these studies in detail in order
     to clarify the significant degree of improvement produced by copy
     avoidance [Ch02].

     Recent work by Chase et al. [CGY01], measuring CPU utilization,
     shows that avoiding copies reduces CPU time spent on data access
     from 24% to 15% at 370 Mbits/s for a 32 KBytes MTU using an
     AlphaStation XP1000 and a Myrinet adapter [BCF+95].  This is an
     absolute improvement of 9% due to copy avoidance.

     The total CPU utilization was 35%, with data access accounting for
     24%.  Thus the relative importance of reducing copies is 26%.  At
     370 Mbits/s, the system is not very heavily loaded.  The relative
     improvement in achievable bandwidth is 34%.  This is the
     improvement we would see if copy avoidance were added when the
     machine was saturated by network I/O.

     Note that improvement from the optimization becomes more important
     if the overhead it targets is a larger share of the total cost.
     This is what happens if other sources of overhead, such as
     checksumming, are eliminated.  In [CGY01], after removing checksum
     overhead, copy avoidance reduces CPU utilization from 26% to 10%.
     This is a 16% absolute reduction, a 61% relative reduction, and a
     160% relative improvement in achievable bandwidth.

     In fact, today's network interface hardware commonly offloads the
     checksum, which removes the other source of per-byte overhead.
     They also coalesce interrupts to reduce per-packet costs.  Thus,
     today copying costs account for a relatively larger part of CPU
     utilization than previously, and therefore relatively more benefit
     is to be gained in reducing them.  (Of course this argument would
     be specious if the amount of overhead were insignificant, but it
     has been shown to be substantial.)

Romanow, et al             Expires August 2003                  [Page 5]

Internet-Draft       RDMA Over IP Problem Statement        February 2003

3.  Memory bandwidth is the root cause of the problem

     Data movement operations are expensive because memory bandwidth is
     scarce relative to network bandwidth and CPU bandwidth [PAC+97].
     This trend existed in the past and is expected to continue into the
     future [HP97, STREAM], especially in large multiprocessor systems.

     With copies crossing the bus twice per copy, network processing
     overhead is high whenever network bandwidth is large in comparison
     to CPU and memory bandwidths.  Generally with today's end-systems,
     the effects are observable at network speeds over 1 Gbits/s.

     A common question is whether increase in CPU processing power
     alleviates the problem of high processing costs of network I/O.
     The answer is no, it is the memory bandwidth that is the issue.
     Faster CPUs do not help if the CPU spends most of its time waiting
     for memory [CGY01].

     The widening gap between microprocessor performance and memory
     performance has long been a widely recognized and well-understood
     problem [PAC+97].  Hennessy [HP97] shows microprocessor performance
     grew from 1980-1998 at 60% per year, while the access time to DRAM
     improved at 10% per year, giving rise to an increasing "processor-
     memory performance gap".

     Another source of relevant data is the STREAM Benchmark Reference
     Information website which provides information on the STREAM
     benchmark [STREAM].  The benchmark is a simple synthetic benchmark
     program that measures sustainable memory bandwidth (in MBytes/s)
     and the corresponding computation rate for simple vector kernels
     measured in MFLOPS.  The website tracks information on sustainable
     memory bandwidth for hundreds of machines and all major vendors.

     Results show measured system performance statistics.  Processing
     performance from 1985-2001 increased at 50% per year on average,
     and sustainable memory bandwidth from 1975 to 2001 increased at 35%
     per year on average over all the systems measured.  A similar 15%
     per year lead of processing bandwidth over memory bandwidth shows
     up in another statistic, machine balance [Mc95], a measure of the
     relative rate of CPU to memory bandwidth (FLOPS/cycle) / (sustained
     memory ops/cycle) [STREAM].

     Network bandwidth has been increasing about 10-fold roughly every 8
     years, which is a 40% per year growth rate.

     A typical example illustrates that the memory bandwidth compares
     unfavorably with link speed.  The STREAM benchmark shows that a
     modern uniprocessor PC, for example the 1.2 GHz Athlon in 2001,

Romanow, et al             Expires August 2003                  [Page 6]

Internet-Draft       RDMA Over IP Problem Statement        February 2003

     will move the data 3 times in doing a receive operation - 1 for the
     network interface to deposit the data in memory, and 2 for the CPU
     to copy the data.  With 1 GBytes/s of memory bandwidth, meaning one
     read or one write, the machine could handle approximately 2.67
     Gbits/s of network bandwidth, one third the copy bandwidth.  But
     this assumes 100% utilization, which is not possible, and more
     importantly the machine would be totally consumed!  (A rule of
     thumb for databases is that 20% of the machine should be required
     to service I/O, leaving 80% for the database application.  And, the
     less the better.)

     In 2001, 1 Gbits/s links were common.  An application server may
     typically have two 1 Gbits/s connections - one connection backend
     to a storage server and one front-end, say for serving HTTP
     [FGM+99].  Thus the communications could use 2 Gbits/s.  In our
     typical example, the machine could handle 2.7 Gbits/s at its
     theoretical maximum while doing nothing else.  This means that the
     machine basically could not keep up with the communication demands
     in 2001, with the relative growth trends the situation only gets

4.  High copy overhead is problematic for many key Internet applications

     If a significant portion of resources on an application machine is
     consumed in network I/O rather than in application processing, it
     makes it difficult for the application to scale - to handle more
     clients, to offer more services.

     Several years ago the most affected applications were streaming
     multimedia, parallel file systems and supercomputing on clusters
     [BS96].  In addition, today the applications that suffer from
     copying overhead are more central in Internet computing - they
     store, manage, and distribute the information of the Internet and
     the enterprise.  They include database applications doing
     transaction processing, e-commerce, web serving, decision support,
     content distribution, video distribution, and backups.  Clusters
     are typically used for this category of application, since they
     have advantages of availability and scalability.

     Today these applications, which provide and manage Internet and
     corporate information, are typically run in data centers that are
     organized into three logical tiers.  One tier is typically a set of
     web servers connecting to the WAN.  The second tier is a set of
     application servers that run the specific applications usually on
     more powerful machines, and the third tier is backend databases.
     Physically, the first two tiers - web server and application server
     - are usually combined [Pi01].  For example an e-commerce server
     communicates with a database server and with a customer site, or a

Romanow, et al             Expires August 2003                  [Page 7]

Internet-Draft       RDMA Over IP Problem Statement        February 2003

     content distribution server connects to a server farm, or an OLTP
     server connects to a database and a customer site.

     When network I/O uses too much memory bandwidth, performance on
     network paths between tiers can suffer.  (There might also be
     performance issues on SAN paths used either by the database tier or
     the application tier.)  The high overhead from network-related
     memory copies diverts system resources from other application
     processing.  It also can create bottlenecks that limit total system

     There are a large and growing number of these application servers
     distributed throughout the Internet.  In 1999 approximately 3.4
     million server units were shipped, in 2000, 3.9 million units, and
     the estimated annual growth rate for 2000-2004 was 17 percent
     [Ne00, PA01].

     There is high motivation to maximize the processing capacity of
     each CPU, as scaling by adding CPUs one way or another has
     drawbacks.  For example, adding CPUs to a multiprocessor will not
     necessarily help, as a multiprocessor improves performance only
     when the memory bus has additional bandwidth to spare.  Clustering
     can add additional complexity to handling the applications.

     In order to scale a cluster or multiprocessor system, one must
     proportionately scale the interconnect bandwidth.  Interconnect
     bandwidth governs the performance of communication-intensive
     parallel applications; if this (often expressed in terms of
     "bisection bandwidth") is too low, adding additional processors
     cannot improve system throughput.  Interconnect latency can also
     limit the performance of applications that frequently share data
     between processors.

     So, excessive overheads on network paths in a "scalable" system
     both can require the use of more processors than optimal, and can
     reduce the marginal utility of those additional processors.

     Copy avoidance scales a machine upwards by removing at least two-
     thirds the bus bandwidth load from the "very best" 1-copy (on
     receive) implementations, and removes at least 80% of the bandwidth
     overhead from the 2-copy implementations.

     An example showing poor performance with copies and improved
     scaling with copy avoidance is illustrative.  The IO-Lite work
     [PDZ99] shows higher server throughput servicing more clients using
     a zero-copy system.  In an experiment designed to mimic real world
     web conditions by simulating the effect of TCP WAN connections on
     the server, the performance of 3 servers was compared.  One server

Romanow, et al             Expires August 2003                  [Page 8]

Internet-Draft       RDMA Over IP Problem Statement        February 2003

     was Apache, another an optimized server called Flash, and the third
     the Flash server running IO-Lite, called Flash-Lite with zero copy.
     The measurement was of throughput in requests/second as a function
     of the number of slow background clients that could be served.  As
     the table shows, Flash-Lite has better throughput, especially as
     the number of clients increases.

                Apache              Flash         Flash-Lite
                ------              -----         ----------
     #Clients   Thruput reqs/s      Thruput       Thruput

     0          520                 610           890
     16         390                 490           890
     32         360                 490           850
     64         360                 490           890
     128        310                 450           880
     256        310                 440           820

     Traditional Web servers (which mostly send data and can keep most
     of their content in the file cache) are not the worst case for copy
     overhead.  Web proxies (which often receive as much data as they
     send) and complex Web servers based on SANs or multi-tier systems
     will suffer more from copy overheads than in the example above.

5.  Copy Avoidance Techniques

     There have been extensive research investigation and industry
     experience with two main alternative approaches to eliminating data
     movement overhead, often along with improving other Operating
     System processing costs.  In one approach, hardware and/or software
     changes within a single host reduce processing costs.  In another
     approach, memory-to-memory networking [MAF+02], hosts communicate
     via information that allows them to reduce processing costs.

     The single host approaches range from new hardware and software
     architectures [KSZ95, Wa97, DWB+93] to new or modified software
     systems [BP96, Ch96, TK95, DP93, PDZ99].  In the approach based on
     using a networking protocol to exchange information, the network
     adapter, under control of the application, places data directly
     into and out of application buffers, reducing the need for data
     movement.  Commonly this approach is called RDMA, Remote Direct
     Memory Access.

     As discussed below, research and industry experience has shown that
     copy avoidance techniques within the receiver processing path alone
     have proven to be problematic.  The research special purpose host
     adapter systems had good performance and can be seen as precursors

Romanow, et al             Expires August 2003                  [Page 9]

Internet-Draft       RDMA Over IP Problem Statement        February 2003

     for the commercial RDMA-based NICs [KSZ95, DWB+93].  In software,
     many implementations have successfully achieved zero-copy transmit,
     but few have accomplished zero-copy receive.  And those that have
     done so make strict alignment and no-touch requirements on the
     application, greatly reducing the portability and usefulness of the

     In contrast, experience has proven satisfactory with memory-to-
     memory systems that permit RDMA - performance has been good and
     there have not been system or networking difficulties.  RDMA is a
     single solution.  Once implemented, it can be used with any OS and
     machine architecture, and it does not need to be revised when
     either of these changes.

     In early work, one goal of the software approaches was to show that
     TCP could go faster with appropriate OS support [CJR89, CFF+94].
     While this goal was achieved, further investigation and experience
     showed that, though possible to craft software solutions, specific
     system optimizations have been complex, fragile, extremely
     interdependent with other system parameters in complex ways, and
     often of only marginal improvement [CFF+94, CGY01, Ch96, DAPP93,
     KSZ95, PDZ99].  The network I/O system interacts with other aspects
     of the Operating System such as machine architecture and file I/O,
     and disk I/O [Br99, Ch96, DP93].

     For example, the Solaris Zero-Copy TCP work [Ch96], which relies on
     page remapping, shows that the results are highly interdependent
     with other systems, such as the file system, and that the
     particular optimizations are specific for particular architectures,
     meaning for each variation in architecture optimizations must be
     re-crafted [Ch96].

     A number of research projects and industry products have been based
     on the memory-to-memory approach to copy avoidance.  These include
     U-Net [EBBV95], SHRIMP [BLA+94], Hamlyn [BJM+96], Infiniband [IB],
     Winsock Direct [Pi01].  Several memory-to-memory systems have been
     widely used and have generally been found to be robust, to have
     good performance, and to be relatively simple to implement.  These
     include VI [VI], Myrinet [BCF+95], Quadrics [QUAD], Compaq/Tandem
     Servernet [SRVNET].  Networks based on these memory-to-memory
     architectures have been used widely in scientific applications and
     in data centers for block storage, file system access, and
     transaction processing.

     By exporting direct memory access "across the wire", applications
     may direct the network stack to manage all data directly from
     application buffers.  A large and growing class of applications has
     already emerged which takes advantage of such capabilities,

Romanow, et al             Expires August 2003                 [Page 10]

Internet-Draft       RDMA Over IP Problem Statement        February 2003

     including all the major databases, as well as file systems such as
     DAFS [DAFS] and network protocols such as Sockets Direct [SDP].

5.1.  A Conceptual Framework: DDP and RDMA

     An RDMA solution can be usefully viewed as being comprised of two
     distinct components: "direct data placement (DDP)" and "remote
     direct memory access (RDMA) semantics".  They are distinct in
     purpose and also in practice - they may be implemented as separate

     The more fundamental of the two is the direct data placement
     facility.  This is the means by which memory is exposed to the
     remote peer in an appropriate fashion, and the means by which the
     peer may access it, for instance reading and writing.

     The RDMA control functions are semantically layered atop direct
     data placement.  Included are operations that provide "control"
     features, such as connection and termination, and the ordering of
     operations and signaling their completions.  A "send" facility is

     While the functions (and potentially protocols) are distinct,
     historically both aspects taken together have been referred as
     "RDMA".  The facilities of direct data placement are useful in and
     of themselves, and may be employed by other upper layer protocols
     to facilitate data transfer.  Therefore, it is often useful to
     refer to DDP as the data placement functionality and RDMA as the
     control aspect.

     [BT02] develops an architecture for DDP and RDMA, and is a
     companion draft to this problem statement.

6.  Security Considerations

     Solutions to the problem of reducing copying overhead in high
     bandwidth transfers via one or more protocols may introduce new
     security concerns.  Any proposed solution must be analyzed for
     security threats and any such threats addressed.  [BSW02] brings up
     potential security weaknesses due to resource issues that might
     lead to denial-of-service attacks, overwrites and other concurrent
     operations, the ordering of completions as required by the RDMA
     protocol, and the granularity of transfer.  Each of these concerns
     plus any other identified threats need to be examined, described
     and an adequate solution to them found.

     Layered atop Internet transport protocols, the RDMA protocols will
     gain leverage from and must permit integration with Internet

Romanow, et al             Expires August 2003                 [Page 11]

Internet-Draft       RDMA Over IP Problem Statement        February 2003

     security standards, such as IPSec and TLS [IPSEC, TLS].  A thorough
     analysis of the degree to which these protocols solve threats is

     Security for an RDMA design requires more than just securing the
     communication channel.  While it is necessary to be able to
     guarantee channel properties such as privacy, integrity, and
     authentication, these properties cannot defend against all attacks
     from properly authenticated peers, which might be malicious,
     compromised, or buggy.  For example, an RDMA peer should not be
     able to read or write memory regions without prior consent.

     Further, it must not be possible to evade consistency checks at the
     recipient.  For example, the RDMA design should not allow a peer to
     update a region after the completion of an authorized update.

     The RDMA protocols must ensure that regions addressable by RDMA
     peers be under strict application control.  Remote access to local
     memory by a network peer introduces a number of potential security
     concerns.  This becomes particularly important in the Internet
     context, where such access can be exported globally.

     The RDMA protocols carry in part what is essentially user
     information, explicitly including addressing information and
     operation type (read or write), and implicitly including protection
     and attributes.  As such, the protocol requires checking of these
     higher level aspects in addition to the basic formation of
     messages.  The semantics associated with each class of error must
     be clearly defined, and the expected action to be taken on mismatch
     be specified.  In some cases, this will result in a catastrophic
     error on the RDMA association, however in others a local or remote
     error may be signalled.  Certain of these errors may require
     consideration of abstract local semantics, which must be carefully
     specified so as to provide useful behavior while not constraining
     the implementation.

7.  Acknowledgements

     Jeff Chase generously provided many useful insights and
     information.  Thanks to Jim Pinkerton for many helpful discussions.

8.  References

          N. J. Boden, D. Cohen, R. E. Felderman, A. E. Kulawik, C. L.
          Seitz, J. N. Seizovic, and W. Su. "Myrinet - A gigabit-per-
          second local-area network", IEEE Micro, February 1995

Romanow, et al             Expires August 2003                 [Page 12]

Internet-Draft       RDMA Over IP Problem Statement        February 2003

          G. Buzzard, D. Jacobson, M. Mackey, S. Marovich, J. Wilkes,
          "An implementation of the Hamlyn send-managed interface
          architecture", in Proceedings of the Second Symposium on
          Operating Systems Design and Implementation, USENIX Assoc.,
          October 1996

          M. A. Blumrich, K. Li, R. Alpert, C. Dubnicki, E. W. Felten,
          "A virtual memory mapped network interface for the SHRIMP
          multicomputer", in Proceedings of the 21st Annual Symposium on
          Computer Architecture, April 1994, pp. 142-153

          J. C. Brustoloni, "Interoperation of copy avoidance in network
          and file I/O", Proceedings of IEEE Infocom, 1999, pp. 534-542

          J. C. Brustoloni, P. Steenkiste, "Effects of buffering
          semantics on I/O performance", Proceedings OSDI'96, USENIX,
          Seattle, WA October 1996, pp. 277-291

          D. Black, M. Speer, J. Wroclawski, "DDP and RDMA Concerns",

          concerns-00.txt, Work in Progress, December 2002

          S. Bailey, T. Talpey, "The Architecture of Direct Data
          Placement (DDP) And Remote Direct Memory Access (RDMA) On
          Internet Protocols", Work in Progress,

          arch-01.txt, February 2003

          C-H Chang, D. Flower, J. Forecast, H. Gray, B. Hawe, A.
          Nadkarni, K. K. Ramakrishnan, U. Shikarpur, K. Wilde, "High-
          performance TCP/IP and UDP/IP networking in DEC OSF/1 for
          Alpha AXP",  Proceedings of the 3rd IEEE Symposium on High
          Performance Distributed Computing, August 1994, pp. 36-42

          J. S. Chase, A. J. Gallatin, and K. G. Yocum, "End system
          optimizations for high-speed TCP", IEEE Communications
          Magazine, Volume: 39, Issue: 4 , April 2001, pp 68-74.

Romanow, et al             Expires August 2003                 [Page 13]

Internet-Draft       RDMA Over IP Problem Statement        February 2003

          H.K. Chu, "Zero-copy TCP in Solaris", Proc. of the USENIX 1996
          Annual Technical Conference, San Diego, CA, January 1996

          Jeffrey Chase, Personal communication

          D. D. Clark, V. Jacobson, J. Romkey, H. Salwen, "An analysis
          of TCP processing overhead", IEEE Communications Magazine,
          volume: 27, Issue: 6, June 1989, pp 23-29

          D. D. Clark, D. Tennenhouse, "Architectural considerations for
          a new generation of protocols", Proceedings of the ACM SIGCOMM
          Conference, 1990

          Direct Access File System

          P. Druschel, M. B. Abbott, M. A. Pagels, L. L. Peterson,
          "Network subsystem design", IEEE Network, July 1993, pp. 8-17

          P. Druschel, L. L. Peterson, "Fbufs: a high-bandwidth cross-
          domain transfer facility", Proceedings of the 14th ACM
          Symposium of Operating Systems Principles, December 1993

          C. Dalton, G. Watson, D. Banks, C. Calamvokis, A. Edwards, J.
          Lumley, "Afterburner: architectural support for high-
          performance protocols", Technical Report, HP Laboratories
          Bristol, HPL-93-46, July 1993

          T. von Eicken, A. Basu, V. Buch, and W. Vogels, "U-Net: A
          user-level network interface for parallel and distributed
          computing", Proc. of the 15th ACM Symposium on Operating
          Systems Principles, Copper Mountain, Colorado, December 3-6,

          R. Fielding, J. Gettys, J. Mogul, F. Frystyk, L. Masinter, P.
          Leach, T. Berners-Lee, "Hypertext Transfer Protocol -
          HTTP/1.1", RFC 2616, June 1999

Romanow, et al             Expires August 2003                 [Page 14]

Internet-Draft       RDMA Over IP Problem Statement        February 2003

          Fibre Channel Standard

          J. L. Hennessy, D. A. Patterson, Computer Organization and
          Design, 2nd Edition, San Francisco: Morgan Kaufmann
          Publishers, 1997

     [IB] InfiniBand Architecture Specification, Volumes 1 and 2,
          Release 1.0.a.

          J. Kay, J. Pasquale, "Profiling and reducing processing
          overheads in TCP/IP", IEEE/ACM Transactions on Networking, Vol
          4, No. 6, pp.817-828, December 1996

          K. Kleinpaste, P. Steenkiste, B. Zill, "Software support for
          outboard buffering and checksumming", SIGCOMM'95

          K. Magoutis, "Design and Implementation of a Direct Access
          File System (DAFS) Kernel Server for FreeBSD", in Proceedings
          of USENIX BSDCon 2002 Conference, San Francisco, CA, February
          11-14, 2002.

          K. Magoutis, S. Addetia, A. Fedorova, M.  I. Seltzer, J. S.
          Chase, D. Gallatin, R. Kisley, R. Wickremesinghe, E. Gabber,
          "Structure and Performance of the Direct Access File System
          (DAFS)", accepted for publication at the 2002 USENIX Annual
          Technical Conference, Monterey, CA, June 9-14, 2002.

          J. D. McCalpin, "A Survey of memory bandwidth and machine
          balance in current high performance computers", IEEE TCCA
          Newsletter, December 1995


          A. Newman, "IDC report paints conflicted picture of server
          market circa 2004", ServerWatch, July 24, 2000

Romanow, et al             Expires August 2003                 [Page 15]

Internet-Draft       RDMA Over IP Problem Statement        February 2003

          M. Pastore, "Server shipments for 2000 surpass those in 1999",
          ServerWatch, February 7, 2001

          D. Patterson, T. Anderson, N. Cardwell, R. Fromm, K. Keeton,
          C. Kozyrakis, R. Thomas, K. Yelick , "A case for intelligient
          RAM: IRAM", IEEE Micro, April 1997

          V. S. Pai, P. Druschel, W. Zwaenepoel, "IO-Lite: a unified I/O
          buffering and caching system", Proc. of the 3rd Symposium on
          Operating Systems Design and Implementation, New Orleans, LA,
          February 1999

          J. Pinkerton, "Winsock Direct: the value of System Area

          J. Postel, "Transmission Control Protocol - DARPA Internet
          Program Protocol Specification", RFC 793, September 1981

          Quadrics Ltd.,

          Sockets Direct Protocol v1.0

          Compaq Servernet,

          The STREAM Benchmark Reference Information,

          M. N. Thadani, Y. A. Khalidi, "An efficient zero-copy I/O
          framework for UNIX", Technical Report, SMLI TR-95-39, May 1995

     [VI] Virtual Interface Architecture Specification Version 1.0.

          J. R. Walsh, "DART: Fast application-level networking via

Romanow, et al             Expires August 2003                 [Page 16]

Internet-Draft       RDMA Over IP Problem Statement        February 2003

          data-copy avoidance", IEEE Network, July/August 1997, pp.

Authors' Addresses

     Stephen Bailey
     Sandburst Corporation
     600 Federal Street
     Andover, MA  01810 USA

     Phone: +1 978 689 1614

     Jeffrey C. Mogul
     Western Research Laboratory
     Hewlett-Packard Company
     1501 Page Mill Road, MS 1251
     Palo Alto, CA  94304 USA

     Phone: +1 650 857 2206 (email preferred)

     Allyn Romanow
     Cisco Systems, Inc.
     170 W. Tasman Drive
     San Jose, CA  95134 USA

     Phone: +1 408 525 8836

     Tom Talpey
     Network Appliance
     375 Totten Pond Road
     Waltham, MA  02451 USA

     Phone: +1 781 768 5329

Full Copyright Statement

     Copyright (C) The Internet Society (2003).  All Rights Reserved.

     This document and translations of it may be copied and furnished to

Romanow, et al             Expires August 2003                 [Page 17]

Internet-Draft       RDMA Over IP Problem Statement        February 2003

     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

Romanow, et al             Expires August 2003                 [Page 18]