Intelligent query processing in P2P networks: Semantic issues and routing algorithms

Abstract

P2P networks have become a commonly used way of disseminating content on the Internet. In this context, constructing efficient and distributed P2P routing algorithms for complex environments that include a huge number of distributed nodes with different computing and network capabilities is a major challenge. In the last years, query routing algorithms have evolved by taking into account different features (provenance, nodes' history, topic similarity, etc.). Such features are usually stored in auxiliary data structures (tables, matrices, etc.), which provide an extra knowledge engineering layer on top of the network, resulting in an added semantic value for specifying algorithms for efficient query routing. This article examines the main existing algorithms for query routing in unstructured P2P networks in which semantic aspects play a major role. A general comparative analysis is included, associated with a taxonomy of P2P networks based on their degree of decentralization and the different approaches adopted to exploit the available semantic aspects.

Deep Q-Learning Based Optimal Query Routing Approach for Unstructured P2P Network

“…• Basic Searching Algorithms: The most common and foremost algorithms to search relevant resources in unstructured P2P systems are Breadth-First Search (BFS) or flooding [22,23].…”

“…BFS tries to find the maximum number of results. However, this generates a large number of messages and increases the number of connected peers with heavy network traffic compared to other approaches [23]. Napster and Gnutella are the suitable examples who implemented this routing method [24,25].…”

“…In contrast to BFS, which sends queries to all neighbors, DFS allows each peer to choose a candidate neighbor to whom the query should be sent. If the query forwarder doesn't get a response within a specific amount of time (TTL), the peer chooses another neighbor to deliver the query [23]. This is an iterative process that is performed until the query is answered or until all of the neighbors have been queried.…”

“…This algorithm keeps a routing table on each peer that, based on previous experience, advises the position of items in the network. Furthermore, the Exponentially Decaying Bloom Filter (EDBF) data structure compresses probabilistic routing tables and enables for efficient query propagation [23].…”

A Mini Review of Peer-to-Peer (P2P) for Vehicular Communication

“…Moreover, their distributed architecture makes P2P incredibly invulnerable to cyberattacks. Unlike standard models, P2P does not aspect of failure [18]. There are three types of P2P overlay technologies that are called unstructured, structured and hybrid.…”