The Strategy of Traffic Dispersion
E. Martin I. Gustafsson, Gunnar Karlsson
Abstract
Open-access reader
E. Martin I. Gustafsson, Gunnar Karlsson
Abstract
Open-access reader
Traffic dispersion means spreading the traffic from a source over multiple independent paths, transmitting it in parallel through the network. The strategy reduces the effects of bursts in the traffic, and may hence improve the network performance in terms of reduced queuing delay. So far, there have been several reports on traffic dispersion, but to our best knowledge, there has been no in-depth investigation of how the strategy affects different types of traffic under various conditions. This paper focuses on the basic properties of traffic dispersion by defining the strategy and investigating it from the source’s as well as from the network’s point of view. We investigate the influence of dispersion on Poisson traffic, traffic generated by two-state Markov chains and traffic generated by the chaotic FPDI-map. The results indicate a large potential of traffic dispersion to provide fast, secure and fault-tolerant transmission for highly bursty traffic sources.
OpenAlex reports 2 citations for this work. Citation counts describe recorded attention and do not establish research quality.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
Traffic dispersion means spreading the traffic from a source over multiple independent paths, transmitting it in parallel through the network. The strategy reduces the effects of bursts in the traffic, and may hence improve the network performance in terms of reduced queuing delay. So far, there have been several reports on traffic dispersion, but to our best knowledge, there has been no in-depth investigation of how the strategy affects different types of traffic under various conditions. This paper focuses on the basic properties of traffic dispersion by defining the strategy and investigating it from the source’s as well as from the network’s point of view. We investigate the influence of dispersion on Poisson traffic, traffic generated by two-state Markov chains and traffic generated by the chaotic FPDI-map. The results indicate a large potential of traffic dispersion to provide fast, secure and fault-tolerant transmission for highly bursty traffic sources.
Key concepts: Computer science, Dispersion (optics), Traffic generation model, Computer network, Markov chain, Poisson distribution, Traffic congestion reconstruction with Kerner's three-phase theory, Distributed computing