Search Algorithms for Distributed Data Structures in P2P Networks

Abstract

The search problem in distributed systems has been evaluated and various proposals have been made to solve them. The problem is particularly pronounced when considering P2P networks, because the different types of P2P network topologies used to dictate the search mechanisms that are used. Specifically, search mechanisms may be key-based, as in structured P2P networks that use distributed data structures such as Distributed Hash Tables (DHTs), Sets, Links, and Trees, or keyword-based for unstructured P2P networks. Online systems today utilize metadata to support search algorithms. Metadata searches are trivial in unstructured networks but are a challenge to the structured networks because of the distributed data structures (DDS). This work proposes to introduce algorithms that support metadata searches in structured P2P networks that utilize DHT overlays. DDSs are introduced into the system as storage layers that support complex data structures. We show that this can be done and searching can be achieved with acceptable performance.