Performance analysis of resource sharing policies in CDMA networks
2007 (English)In: International Journal of Communication Systems, ISSN 1074-5351, E-ISSN 1099-1131, Vol. 20, no 2, 207-233 p.Article in journal (Refereed) Published
We consider an interference-limited code division multiple access (CDMA) network, in which some of the service classes are explicit rate controlled. Such rate-controlled service classes include adaptive ones whose holding time does not depend on the sending rate, while the holding time of elastic services becomes longer when their sending rate is slowed down. We develop a Markov model that allows us to study the impact of employing service differentiation (prioritization) during admission control and rate control on the classwise blocking probabilities and throughputs. The admission control takes into account the interference from adjacent cells as a log-normally distributed random variable and employs service class and state-dependent (soft) blocking in addition to hard blocking. The state-dependent rate control algorithm takes into account the minimum and maximum requested rates of admitted sessions in a fairness policy-dependent manner. Our conclusion is that work conserving bandwidth sharing policies may have a major impact on the moments of the time spent in the system by elastic services, but only a minor one on the classwise blocking probabilities. When the admission control algorithm takes into account the neighbour cell caused interference, the blocking probability of all classes increases somewhat, but the throughput of the accepted sessions increases. Blocking prioritization in the soft blocking model affects the classwise total blocking probabilities but has little impact on the throughput of adaptive and elastic traffic.
Place, publisher, year, edition, pages
John Wiley & Sons, 2007. Vol. 20, no 2, 207-233 p.
Engineering and Technology
IdentifiersURN: urn:nbn:se:kth:diva-203083DOI: 10.1002/dac.817ISI: 000244080300005ScopusID: 2-s2.0-33846861308OAI: oai:DiVA.org:kth-203083DiVA: diva2:1080853
UT: 000244080300005; ScopusID: 33846861308. QC 201703152017-03-122017-03-122017-03-15Bibliographically approved