Synergy: blending heterogeneous measurement elements for effective network monitoring

Network traffic matrices are important for various network planning and management operations. Previous work for estimation of traffic matrices is based on either link load records obtained from SNMP or flow level sampled data available from Net-flow records. Sampled flow-level data provides us good approximations for traffic matrices, however, static sampling rates are not desired. People have proposed sampling solutions with dynamic sampling rates which adapt according to the changing behavior of the network. SNMP link level reports, on the other hand are computationally less extensive but do not provide optimum solutions for network traffic estimation. This work addresses the problem of measurement in a wider scope and we propose to study merging together data from multiple sources and using it effectively for network monitoring and measurement tasks. We address this problem in both time and spatial domain and intend to develop a framework for dynamically choosing monitoring points...

• Type: Conference
• Year: 2006
• Where: CONEXT
• Authors: Awais Ahmed Awan, Andrew W. Moore
• Keywords: Computer Networks, CONEXT 2006, Network Traffic, Network Traffic Matrices, Traffic Matrices