Title :
Network prediction in a policy-based IP network
Author :
Beigi, Mandis ; Verma, Dinesh
Author_Institution :
IBM Thomas J. Watson Res. Center, Hawthorne, NY, USA
Abstract :
In a quality of service (QoS) enabled network, the Internet service providers (ISPs) need to define the network requirements for the customers by defining a set of policies. Every time a new customer gets added or removed or when a customer´s resource requirements change, the administrator needs to modify the policies installed on the routers/servers in the network. However, these modifications will affect the traffic flowing in the network and in turn might take away some of the existing customers´ resources. Therefore, there is a need to predict how changing the policies will affect the performance of the existing traffic flows in the network. We present a mechanism for predicting whether adding, removing or changing a policy will degrade the performance of the traffic flows belonging to the previous customers. The network administrator can use this information to decide whether such modifications to the policies are desired
Keywords :
Internet; computer network management; network servers; prediction theory; quality of service; resource allocation; telecommunication network routing; telecommunication traffic; Internet service providers; QoS; network administrator; network configuration management; network prediction; network routers; network servers; policy-based IP network; prediction tool; quality of service; resource requirements; Bandwidth; Degradation; Diffserv networks; IP networks; Intelligent networks; Network servers; Quality of service; Telecommunication traffic; Web and internet services; Web server;
Conference_Titel :
Global Telecommunications Conference, 2001. GLOBECOM '01. IEEE
Conference_Location :
San Antonio, TX
Print_ISBN :
0-7803-7206-9
DOI :
10.1109/GLOCOM.2001.966231