FloRA - Flock-Based Resource Allocation for Decentralized Distributed Virtual Environments

The growth of Massively Multiuser Virtual Environments (MMVEs), increasingly interactive social networking platforms and in particular their likely convergence render today's centralized hosting approaches impracticable. To handle potentially single-instance virtual environments of such massive scale, decentralized systems are necessary that also involve the resources of clients. The expedient design of techniques for enabling this kind of next-generation Decentralized Distributed Virtual Environments (DDVEs) is a growing field of research. We aim at the provision of an infrastructure enabling such DDVEs in the HyperVerse project, focusing on collaboration and self-organization as means to achieve maximum scalability. In this paper we present FloRA, a flock-based resource allocation scheme that helps alleviate the load imposed by regions with a higher user density as they often occur in DDVEs. Exploiting the heterogeneity of clients, only local information is utilized to tackle exigencies. Evaluations show that both for the discovery of these regions and their alleviation the local views converge well to a global one, with favorable effects on the overlay topology.