Semi-Fluid content distribution model for fast data dissemination in heterogeneous overlay networks

Abstract

Overlay networks and peer-to-peer networking have emerged as alternative solutions to solve many problems related to massive information distribution and processing tasks by providing enhanced services in the application layer. For instance, overlay networks can improve data dissemination in P2P file sharing overlays and Content Distribution Networks (CDNs). In such overlay networks, delay is considered as a critical performance metric. A viable content distribution model could considerably minimize the completion time and would lead to efficient utilization of resources such as network bandwidth. The main objective of this paper is to design a new flexible content distribution model for fast data dissemination in overlay networks with heterogeneous nodes. The proposed “Semi-Fluid” content distribution model is a hybrid model that enables fast data dissemination in overlay networks by combining Chunk and Fluid content distribution models. In this new content distribution model, the chunk content is distributed into heterogeneous overlay nodes in a fluid manner. The proposed hybrid content distribution model eliminates both a backpressure problem caused by the Fluid content distribution model, as well as a chunk transition delay caused by the Chunk content distribution model. The performance of the proposed Semi-Fluid content distribution model has been evaluated by mathematical analysis as well as by real implementation over the “PlanetLab” platform, and the obtained results show that the Semi-Fluid content distribution model efficiently reduces the total download time of the clients. Hence, the Semi-Fluid content distribution model can be deployed as a promising solution for fast data dissemination in heterogeneous overlay networks.