High performance TCP in ANSNET

This report concentrates on specific requirements and goals of the research networks supported by ANSNET, but applies to any TCP dominated high speed WAN and in particular those striving to support high speed end-to-end flows. Measurements have been made under conditions intended to better understand performance barriers imposed by network equipment queueing capacities and queue drop strategies.The IBM RS/6000 based routers currently supporting ANSNET performed very well in these tests. Measurements have been made with the current software and performance enhanced software. Single TCP flows are able to achieve 40 Mb/s and competing multiple TCP flows achieve over 41 Mb/s link utilization on 44.7 Mb/s DS3 links with delays comparable to US cross continent ANSNET delays. Congestion collapse is demonstrated with intentionally reduced queueing capacity and using window sizes much larger than optimal.A variation of Floyd and Jacobson's Random Early Detection (RED) algorithm [1] is tested. Performance improved with the use of RED for tests involving multiple flows. With RED and queueing capacity at or above the delay bandwidth product, congestion collapse is avoided, allowing the maximum window size to safely be set arbitrarily high.Queueing capacity greater than or equal to the delay bandwidth product and RED are recommended. RED provides performance improvement in all but the single flow case, but cannot substitute for adequate queueing capacity, particularly if high speed flows are to be supported.

[1]  Walter Willinger,et al.  On the Self-Similar Nature of Ethernet Traffic ( extended version ) , 1995 .

[2]  Peter B. Danzig,et al.  Characteristics of wide-area TCP/IP conversations , 1991, SIGCOMM '91.

[3]  Larry L. Peterson,et al.  TCP Extensions Considered Harmful , 1991, RFC.

[4]  Stev Knowles,et al.  IESG Advice from Experience with Path MTU Discovery , 1993, RFC.

[5]  Van Jacobson.  Some Design Issues for High-Speed Networks , 1993 .

[6]  David A. Borman,et al.  Implementing TCP/IP on a cray computer , 1989, CCRV.

[7]  George C. Polyzos,et al.  Tracking long-term growth of the NSFNET , 1994, CACM.

[8]  Van Jacobson,et al.  TCP Extension for High-Speed Paths , 1990, RFC.

[9]  Sally Floyd,et al.  Connections with multiple congested gateways in packet-switched networks part 1: one-way traffic , 1991, CCRV.

[10]  Guy T. Almes.  Engineering Trans-Oceanic 44 Mb/s TCP/IP Networks , 1994, Comput. Networks ISDN Syst..

[11]  John Nagle,et al.  Congestion control in IP/TCP internetworks , 1984, CCRV.

[12]  Robert T. Braden,et al.  Requirements for Internet Hosts - Communication Layers , 1989, RFC.

[13]  Alex M. McKenzie.  Problem with the TCP big window option , 1989, RFC.

[14]  Van Jacobson,et al.  Congestion avoidance and control , 1988, SIGCOMM '88.

[15]  Scott Shenker,et al.  Observations on the dynamics of a congestion control algorithm: the effects of two-way traffic , 1991, SIGCOMM '91.

[16]  Stephen E. Deering,et al.  Path MTU discovery , 1990, RFC.

[17]  kc claffy,et al.  Long-term traffic aspects of the NSFNET , 1993 .

[18]  QUTdN QeO,et al.  Random early detection gateways for congestion avoidance , 1993, TNET.

[19]  Vern Paxson,et al.  Empirically derived analytic models of wide-area TCP connections , 1994, TNET.

[20]  Van Jacobson,et al.  Random early detection gateways for congestion avoidance , 1993, TNET.

[21]  Van Jacobson,et al.  Traffic phase effects in packet-switched gateways , 1991, CCRV.

[22]  Scott Shenker,et al.  Some observations on the dynamics of a congestion control algorithm , 1990, CCRV.

[23]  Sally Floyd,et al.  Wide area traffic: the failure of Poisson modeling , 1995, TNET.

[24]  Richard Fox,et al.  TCP big window and NAK options , 1989, RFC.

[25]  David L. Mills,et al.  Internet Delay Experiments , 1983, RFC.