Performance and Efficiency in Distributed Enterprise Management

This paper is motivated by the increasing needfor scaleable, distributed management architectures forintegrated network, system and application managementwithin the enterprise network environment. Such integration, extension and wide area deploymentof management functionality impose heavy performancerequirements and produce increased management traffic.Aiming to minimize this traffic and the overall response time, we propose a distributedhierarchical caching scheme that attempts to takeadvantage of the diverse consistency requirements ofmanagement applications. We define coherency conditionsand update policies, identify the appropriateinteraction semantics, and discuss an SNMP-basedimplementation. In order to evaluate the proposed modeland to quantify the expected performance gains weconstruct a simple queuing model that provides analyticalresults on the improvement of response time and thereduction of management traffic. Finally, the analysisof experimental results provides some insight on performance improvement for specific classes ofmanaged objects.