A Self-Organized Overlay Network Management Mechanism for Heterogeneous Environments

Abstract

The technologies of Cloud Computing and NGN are now growing a paradigm shift where various services are provided to business users over the network. In conjunction with this movement, many studies are active to realize a ubiquitous computing environment in which a huge number of individual users can share their computing resources on the Internet, such as personal computers (PCs), game consoles, sensors and so on. To realize an effective resource discovery mechanism for such an environment, this paper presents an adaptive overlay network that enables a self-organizing resource management system to efficiently adapt to a heterogeneous environment. The proposed mechanism is composed of two functions. One is to adjust the number of logical links of a resource, which forward search queries so that less-useful query flooding can be reduced. The other is to connect resources so as to decrease the communication latency on the physical network rather than the number of query hops on an overlay network. To further improve the discovery efficiency, this paper integrates these functions into a self-organizing resource management system, SORMS, which has been proposed in our previous work. The simulation results indicate that the proposed mechanism can increase the number of discovered resources by 60% without decreasing the discovery efficiency, and can reduce the total communication traffic by 80% compared with the original SORMS. This performance improvement is obtained by efficient control of logical links in a large scale network.