MediaBroker: A pervasive computing infrastructure for adaptive transformation and sharing of stream data

MediaBroker is a distributed framework designed to support pervasive computing applications. Key contributions of MediaBroker are efficient and scalable data transport, data stream registration and discovery, an extensible system for data type description, and type-aware data transport that is capable of dynamically transforming data en route from source to sinks. Specifically, the architecture consists of a transport engine and peripheral clients and addresses issues in scalability, data sharing, data transformation, and platform heterogeneity. Details of the MediaBroker architecture, implementation, and a concrete application example are presented in this article. Experimental study shows reasonable performance for selected streaming media-intensive applications. For example, relative to baseline TCP performance, MediaBroker incurs under 11% latency overhead and achieves roughly 80% of the TCP throughput when streaming items larger than 100 kB across our infrastructure. The EventWeb application demonstrates the utility and graceful scaling of MediaBroker for supporting pervasive computing applications. © 2005 Elsevier B.V. All rights reserved. ∗ Corresponding author. E-mail addresses: [email protected] (U. Ramachandran), [email protected] (M. Modahl), [email protected] (I. Bagrak), [email protected] (M. Wolenetz), [email protected] (D. Lillethun), [email protected] (B. Liu), [email protected] (J. Kim), [email protected] (P. Hutto), [email protected] (R. Jain). 1574-1192/$ see front matter © 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.pmcj.2005.03.003 258 U. Ramachandran et al. / Pervasive and Mobile Computing 1 (2005) 257–276