Title :
Darkstar: Using exploratory data mining to raise the bar on network reliability and performance
Author :
Kalmanek, Charles R. ; Ge, Ihui ; Lee, Seungjoon ; Lund, Carsten ; Pei, Dan ; Seidel, Joseph ; Van der Merwe, Jacobus ; Ates, J.
Abstract :
Networks have become a critical infrastructure, and performance requirements for network-based applications are becoming increasingly stringent. This trend challenges service providers to raise the bar on the performance and reliability of network services. To achieve this, new network and service management systems are needed that enable providers to continually improve performance, identify issues that are flying under the radar of network operations, and troubleshoot complex issues. This paper presents the Darkstar system, which allows analysts to address these challenges using exploratory data mining and sophisticated correlation tools. We present an overview of key applications that are built on top of the Darkstar system to illustrate the power of the approach.
Keywords :
data mining; telecommunication network management; telecommunication network reliability; telecommunication services; Darkstar system; correlation tools; critical infrastructure; exploratory data mining; network management; network operation radar; network reliability; service management; Data mining; Degradation; Disaster management; IP networks; Jacobian matrices; Multiprotocol label switching; Performance loss; Power system management; Radar; Web and internet services; alarm correlation; exploratory data mining; network management; service management;
Conference_Titel :
Design of Reliable Communication Networks, 2009. DRCN 2009. 7th International Workshop on
Conference_Location :
Washington, DC
Print_ISBN :
978-1-4244-5047-3
Electronic_ISBN :
978-1-4244-5048-0
DOI :
10.1109/DRCN.2009.5340032