An Architecture for Building Scalable, Web-Based Management Services

We present the architecture of Marvel, a distributed computing environment for building scalable management services using intelligent agents and the world-wide web. Marvel is based on an information model that generates computed views of management information and a distributed computing model that...

Full description

Saved in:
Bibliographic Details
Published inJournal of network and systems management Vol. 7; no. 1; pp. 73 - 104
Main Author Anerousis, Nikolaos
Format Journal Article
LanguageEnglish
Published New York Springer Nature B.V 01.03.1999
Subjects
Online AccessGet full text

Cover

Loading…
More Information
Summary:We present the architecture of Marvel, a distributed computing environment for building scalable management services using intelligent agents and the world-wide web. Marvel is based on an information model that generates computed views of management information and a distributed computing model that makes these views available to a variety of client applications. Computed views consist of monitoring, control and event views of information collected from network elements and subsequently aggregated using a series of spatial and temporal filters. Marvel does not replace existing element management agents but rather builds on top of them a hierarchy of servers that generate computed views and present them to client applications in a number of formats, including Java-enriched web pages. It uses a distributed persistent store to reduce the cost associated with centralized network management systems and mobile agent technology to: (a) support thin clients by uploading the necessary code to access Marvel services; and (b) extend its functionality dynamically by downloading code that incorporates new objects and services. A prototype implementation in Java is presented together with results from its first application on a residential broadband access system using cable modems. [PUBLICATION ABSTRACT]
Bibliography:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 23
ObjectType-Article-2
ObjectType-Feature-1
ISSN:1064-7570
1573-7705
DOI:10.1023/A:1018717900333