Skip to search boxSkip to navigationSkip to main content

Distributed Reactive XML: an XML-centric coordination middleware

Research Output:
Book / Anthology / Report
Report

Open access

Publication Information

Output type

Research Output:
Book / Anthology / Report
Report

Original language

English

Publication milestones

  • Published - 03/2005

Publication status

Published - 03/2005

Place of publication

Copenhagen

Edition

TR-2005-62

Publisher

IT-Universitetet i København, Denmark

Book series

  • Book series name: IT University Technical Report Series
    Series number: TR-2005-62
    ISSN: 1600-6100

ISBN (Electronic)

87-7949-091-3

Abstract

XML-centric models of computation have been proposed as an answer to the demand for interoperability, heterogeneity and openness in coordination models. We present a prototype implementation of an open XML-centric coordination middleware called Distributed Reactive XML. The middleware has as theoretical foundation a general extendable, distributed process calculus inspired by the theory of Bigraphical Reactive Systems. The calculus is extendable just as XML is extendable, in that its signature and reaction rules are not fixed. It is distributed by allowing both the state of processes as well as the set of eaction rules to be distributed (or partly shared) between different clients. The calculus is implemented by representing process terms as XML documents stored in a value-oriented, peer-to-peer XML Store and reaction rules as XML transformations performed by the clients. The formalism does not require that only process terms are stored---inside process terms one may store application specific data as well. XML Store provides transparent sharing of process terms between all participating peers. Conflicts between concurrent reaction rules are handled by an optimistic concurrency control. The implementation thus provides an open XML-based coordination middleware with a formal foundation that encompasses both the shared data, processes and reaction rules.

Funding Details

This work was funded in part by the Danish Research Agency (grant no.: 2059-03-0031) and the IT University of Copenhagen (the LaCoMoCo project).

Access to documents

Final published version, 227.59 KB