Multi-Dimensional Analysis of ISA Server Logs

Abstract

Network administrators are always interested in analyz-ing the activities and bandwidth usage of network users in order to manage network. Usually, third party tools are being used for such analysis, but third party tools suffer from slow response and lack of customization options. In this paper, we present a novel approach for the effective and very flexible multi-dimensional analysis of network usage, by directly analyzing the ISA Server Proxy Logs. In this approach, we accumulate the ISA Server logs into SQL Server and transform it into the dimensional model using SQL Server Integration Services. Dimensional model is a database modeling technique for the efficient analysis of large datasets. In order to evaluate the effectiveness of our proposed approach, we analyze ISA Server Proxy logs generated at a local university. response of these tools is too slow; furthermore the customization options available in these tools are also limited. In our approach, we connect ISA Server to the SQL Server so that the logs can be accumulated into the database directly. Once the logs are in the structured form, he database model is transformed into the dimensional model using SQL Server Integration Services. Dimensional model is a database design technique used to enhance the query response and for the multi-dimensional analysis of data. The approach is tested on the real time ISA Server proxy logs obtained from a local university and produced very useful multi-dimensional reports and graphs.