Reliable communication in data networks is defined as the ability to reproduce at the receiver exactly what has been transmitted by the source, with very high probability. This thesis establishes guidelines for designing network protocols that ensure reliable end-to-end communication. The first section of the thesis focuses on error detection protocols; the second section examines data retransmission schemes. The ability to detect errors in a network, such as bit errors on the data line and lost data due to buffer overflow, is obviously an important issue. Nevertheless, it appears there has been little prior research done on the fundamental issues involved with designing an error protection scheme. A five step methodology is presented that provides insight into, first, the order in which errors should be considered when designing an error detection scheme, second, which types of error detection mechanisms are most effective, and third, which layer should be responsible for detecting a given type of error. The issues of overall effectiveness, efficiency, and robustness of an error detection scheme are addressed. Once an error has been detected, the system must be able to either correct the error or retransmit the portion of data in error. Only retransmission options are explored in this work. A delay criterion for evaluating the performance of general retransmission schemes is presented. This is used to examine the tradeoffs involved with retransmission schemes that are based on a polling mechanism and schemes that employ a timer mechanism. The interplay between the requirements, resources, and error characteristics of a network and the design of a retransmission scheme is discussed. Much of the analysis is devoted to
[1]
M. Schwartz,et al.
Architectures, features, and implementation of high-speed transport protocols
,
1991,
IEEE Network.
[2]
Paul C. van Oorschot,et al.
An Introduction to Error Correcting Codes with Applications
,
1989
.
[3]
R. Damodaram,et al.
Error detection and correction options for data services in B-ISDN
,
1990,
[Proceedings] GLOBECOM '90: IEEE Global Telecommunications Conference and Exhibition.
[4]
Andrew S. Tanenbaum,et al.
Computer Networks, Second Edition
,
1981
.
[5]
Lixia Zhang,et al.
Why TCP timers don't work well
,
1986,
SIGCOMM '86.
[6]
Werner Bux,et al.
Performance of an improved data link control protocol
,
1988,
ICCC.
[7]
Boaz Patt-Shamir,et al.
Self-stabilization by local checking and correction
,
1991,
[1991] Proceedings 32nd Annual Symposium of Foundations of Computer Science.
[8]
R. Gallager.
Information Theory and Reliable Communication
,
1968
.
[9]
Alan G. Konheim,et al.
A Queueing Analysis of Two ARQ Protocols
,
1980,
IEEE Trans. Commun..
[10]
Arun N. Netravali,et al.
Design and implementation of a high-speed transport protocol
,
1990,
IEEE Trans. Commun..
[11]
Douglas Comer,et al.
Internetworking with TCP/IP (2nd ed.), vol. I
,
1991
.
[12]
Subrahmanyam Dravida.
Error control aspects of high speed networks
,
1992,
[Proceedings] IEEE INFOCOM '92: The Conference on Computer Communications.
[13]
J. G. Fletcher,et al.
An Arithmetic Checksum for Serial Transmissions
,
1982,
IEEE Trans. Commun..