Proximity-based Networking: Small world overlays optimized with particle swarm optimization
This addresses the need for efficient and resilient peer-to-peer networks for applications like file-sharing and streaming, though it appears incremental as it builds on existing small-world and Chord concepts.
The paper tackles the problem of information dissemination in large, dynamic distributed systems by proposing a networking scheme that combines geographic location in Chord with proximity-based small-world overlays, resulting in efficient data transfer and routing with minimal information loss through FEC and UDP, while offering resilience and fault-tolerance.
Information dissemination is a fundamental and frequently occurring problem in large, dynamic, distributed systems. In order to solve this, there has been an increased interest in creating efficient overlay networks that can maintain decentralized peer-to-peer networks. Within these overlay networks nodes take the patterns of small world networks, whose connections are based on proximity. These small-world systems can be incredibly useful in the dissemination and lookup of information within an internet network. The data can be efficiently transferred and routing with minimal information loss through forward error correct (FEC) and the User Datagram Protocol (UDP). We propose a networking scheme that incorporates geographic location in chord for the organization of peers within each node's partitioned key space. When we combine this with a proximity-based neighborhood set {based on the small world structure} we can mimic the efficient of solutions designed to solve traditional small-world problems, with the additional benefit of resilience and fault-tolerance. Furthermore, the routing and address book can be updated based on the neighborhood requirements. The flexibility of our proposed schemes enables a variety of swarm models, and agents. This enables our network to as an underlying networking model that can be applied to file-sharing, streaming, and synchronization of networks.