Performance of the timed token scheme in MAP

Yue, On-Ching; Brooks, C.A.

doi:10.1109/26.57513

Cited by 28 publications

(1 citation statement)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Examples of timelimited policies can be found in the ANSI/IEEE 802.5 Standard, in the IEEE 802.4 standard proposed for the token bus [18], and in the (T1-T2 ) bandwidth allocation scheme for wideband packet networks proposed in [14]. Establishing a server timeout upon each visit to a queue in the polling cycle is important in order to guarantee a minimum bandwidth to different types of traffic to the system.…”

Section: Introductionmentioning

confidence: 99%

Gated time-limited polling systems

Silva

Gail

Muntz

1998

Performance and Management of Complex Communication Networks

View full text Add to dashboard Cite

Polling systems have found widespread application in the modeling of computer communication systems and protocols. As a consequence, a great. deal of effort has been devoted to analyzing a variety of polling mechanisms. In particular, systems for which a limit on the length of time the server may spend while visiting a queue have recently been the subject of several·studies. For example, protocols and multiplexing schemes for multimedia. networks may employ a. limit on the amount of time the shared channel is allocated to each type of traffic. However, such timelimited policies are difficult to analyze, and relatively few results about them have appeared in the literature. In this paper the·previous methodology we developed to study the exhaustive service case with preemptive timeouts is extended and applied to the analysis of a system operating under the time-limited gated discipline and having either preemptive timeouts or nonpreemptive timeouts. Measures of interest that caD. be calculated include joint distributions at server arrivals, at server departures, and over all time. Additional measures of interest, such as expected wa.iting times and blocking probabilities for each customer class in the system, can also be computed. 'The work of E. de Souza e Silva and R. R.

show abstract

Section: Introductionmentioning

confidence: 99%