Multiplexing, TCP, and UDP: Pointers to the Discussion

RUTS:

• Requirements for Unicast Transport/Sessions BOF (ruts). This is a BOF from the Orlando IETF, December 1998.

Multiplexing above the transport layer:

• Jim Gettys and Henrik Frystyk Nielsen, The WebMUX Protocol, August 1, 1998.
  "WebMUX is a session management protocol separating the underlying transport from the upper level application protocols. It provides a lightweight communication channel to the application layer by multiplexing data streams on top of a reliable stream oriented transport... HTTP/1.1 [6] persistent connections and pipelining [11] will reduce network traffic and the amount of TCP overhead caused by opening and closing TCP connections. However, the serialized behavior of HTTP/1.1 pipelining does not adequately support simultaneous rendering of inlined objects - part of most Web pages today; nor does it provide suitable fairness between protocol flows, or allow for graceful abortion of HTTP transactions without closing the TCP connection (quite common in HTTP operation)."

• S. Spero, Session Control Protocol (SCP), 1996.
  "SCP is a simple protocol which lets a server and client have multiple conversations over a single TCP connection."

Multiplexing and TCP:

• J. Touch, TCP Control Block Interdependence, RFC 2140, April 1997.
  "This memo makes the case for interdependent TCP control blocks, where part of the TCP state is shared among similar concurrent connections, or across similar connection instances. TCP state includes a combination of parameters, such as connection state, current round- trip time estimates, congestion control information, and process information."
  While RFC 2104 concentrates on the sharing of TCP state at connection establishment and close, the possibility of sharing RTT or window-size state during a connection is also discussed.

• Lars Eggert, John Heidemann and Joe Touch, Effects of Ensemble-TCP, CCR, January 2000.
  "TCP control block sharing is a technique for reusing information among connections in series and aggregating it among connections in parallel... This paper explores the design space of a modified TCP stack that utilizes these two ideas."

• Michael Savoric, Holger Karl, Ensemble Flow Congestion Management, Performance Evaluation of an Improved Common Congestion Controller for TCP Connections, August 2002.
  "The assumption that an entire ensemble should be at most as aggressive as a single connection appears very conservative; it is here that EFCM and E-TCP differ."

• Hari Balakrishnan, Venkata N. Padmanabhan, Srinivasan Seshan, Mark Stemm, and Randy H. Katz, TCP Behavior of a Busy Internet Server: Analysis and Improvements, Proc IEEE Infocom '98 San Francisco, CA March 1998.
  "To improve the congestion control and loss recovery performance of parallel TCP connections, we present a new integrated approach to congestion control and loss recovery that works across a set of concurrent connections... Our solutions are more general than application-specific enhancements such as the use of persistent connections in P-HTTP and HTTP/1.1." This paper proposes the sharing of a single congestion window across multiple TCP connections.

• Venkata N. Padmanabhan, A Proposal for Supporting Lots of Unicast Multiplexed Connections (SLUMS) ( postscript, pdf). This document argues that TCP is a good basis for multiplexing.

• Venkata N. Padmanabhan, Addressing the Challenges of Web Data Transport, Ph.D. thesis, September 1998.
  Chapters 6 and 7 and Appendix B describe and develop the proposed `TCP session' layer, for shared congestion control and loss recovery across a set of TCP connections between a pair of hosts.

Multiplexing with new transport protocols:

• Rajarshi Gupta, Mike Chen, Steven McCanne and Jean Walrand, WebTP: A User-Centered Web Transport Protocol, a new project at UC Berkeley.
  One of the four design goals of the WebTP project is for session-oriented congestion management. The protocol includes receiver-driven congestion control and allows a client browser to process received packet data in an arbitrary order.

Multiplexing below the transport layer:

• Hari Balakrishnan, Hariharan S. Rahul, and Srinivasan Seshan, An Integrated Congestion Management Architecture for Internet Hosts (html, pdf), SIGCOMM 99, September 1999.
  The Congestion Manager, under development, would provide shared congestion control state for both TCP and UDP.

Related articles on multiplexing issues:

• David Tennenhouse, Layered Multiplexing Considered Harmful, in Protocols for High-Speed Networks, Rudin and Williamson (Editors), North Holland, Amsterdam, 1989.
  This paper is concerned with the jitter that arises from layered multiplexing in the upper and middle layers of the protocol stack. "It is claimed that the extensive duplication of multiplexing functionality across the middle and upper layers is harmful and should be avoided." Examples given are of the OSI architecture, which at the time provided for multiplexing within six of the seven layers.

• David Feldmeier, Multiplexing Issues in Communication System Design, SIGCOMM 1990.
  Abstract: This paper considers some of the multiplexing issues in communication system design by examining overall system issues. In particular, we distinguish physical multiplexing of resources from logical multiplexing of streams. Both physical-resource multiplexing and logical multiplexing determine the service that can be provided by a communication system. We also discuss two issues affected by logical multiplexing: flow control and the relationship between control and data streams of a connection. We conclude that the granularity of physical resource sharing must be fine enough to meet the jitter and latency constraints of demanding applications. Also, high speed communication systems should restrict their logical multiplexing to layer 3.