Abstract
Statistical multiplexing of traffic streams results in reduced network bandwidth requirement. The resulting gain increases with the increase in the number of streams being multiplexed together. However, the exact shape of the gain curve, as more and more streams are multiplexed together, is not known. In this paper, we first present the generalized result that the statistical gain of combining homogeneous traffic streams, of any traffic type, is a linear function of the number of streams being multiplexed. That is, given a fixed Quality of Service (QoS) constraint, like percentile delay, D, the bandwidth requirement of n streams to satisfy the delay constraint D is n x R x c where R is the bandwidth requirement of a single stream that satisfies the constraint D and c e (0,1]. We present the linear bandwidth gain result, using an extensive simulation study for video traces, specifically, streaming video (IPTV traces) and interactive video (CISCO Telepresence traces). The linear bandwidth gain result is then verified using analytical tools from two different domains. First, we validate the linearity using Queueing Theory Analysis, specifically using Interrupted Poisson Process (IPP) and Markov Modulated Poisson Process (MMPP) modeling. Second, we formally prove the linear behavior using the Asymptotic Analysis of Algorithms, specifically, the Big-O analysis.
Highlights
Statistical multiplexing and effective bandwidth requirements have been the focus of many research studies, approximate algorithms and network planning heuristics
We present the linear bandwidth gain result, using an extensive simulation study for video traces, streaming video (IPTV traces) and interactive video (CISCO Telepresence traces)
We focus on the nature of statistical multiplexing itself and try to determine the shape of the bandwidth gain curve as more and more streams, constrained by a single end-to-end QoS measure, are multiplexed together
Summary
Statistical multiplexing and effective bandwidth requirements have been the focus of many research studies, approximate algorithms and network planning heuristics. Their main focus remain on utilizing statistical multiplexing to maximize the number of streams a connection carries while satisfying some QoS measures like delay, jitter and/or packet loss. We focus on the nature of statistical multiplexing itself and try to determine the shape of the bandwidth gain curve as more and more streams, constrained by a single end-to-end QoS measure, are multiplexed together. The results carry significant importance in the communications networks as they simplify network dimensioning and resource planning It brings down the problem of allocating bandwidth, for multiplexed traffic streams of the same traffic type, from heuristics and approximations to exact calculations.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
