Active Architecture for Pervasive Contextual Services

TL;DR

This paper proposes an active, peer-to-peer architecture to support scalable, manageable, and evolvable pervasive contextual services over wide-area networks.

Contribution

It introduces a novel peer-to-peer infrastructure design for distributed matching services in pervasive computing environments.

Findings

Demonstrates scalability of the peer-to-peer matching service

Shows manageability and incremental evolution capabilities

Validates approach through prototype implementation

Abstract

Pervasive services may be defined as services that are available "to any client (anytime, anywhere)". Here we focus on the software and network infrastructure required to support pervasive contextual services operating over a wide area. One of the key requirements is a matching service capable of as-similating and filtering information from various sources and determining matches relevant to those services. We consider some of the challenges in engineering a globally distributed matching service that is scalable, manageable, and able to evolve incrementally as usage patterns, data formats, services, network topologies and deployment technologies change. We outline an approach based on the use of a peer-to-peer architecture to distribute user events and data, and to support the deployment and evolution of the infrastructure itself.