Title :
Evaluation for Network Loads Using Data Fusion
Author :
Li Xiaojuan ; Guan Yong ; Yuan Huimei
Author_Institution :
Coll. of Inf. Eng., Capital Normal Univ., Beijing, China
Abstract :
A solution to evaluate network workload by data fusion is put forward, which can be for surveillance the traffic of interconnected communications network in order to keep the network working well by identifying potentially serious problems in the early stages and evaluating network performance. Through fusing the historic network traffic data and network online traffic data, which is based on least squares- SVM and D-S evidence theory, we have been able to identify several network trends of specific chronic problems effectively and provide insights to Internet Service Providers (ISP) for engineering their network traffic, and planning the capacity accordingly. Evaluation experiments indicate that the merging techniques yield improvements in the accuracy and effectiveness.
Keywords :
Internet; computer network performance evaluation; inference mechanisms; least squares approximations; sensor fusion; support vector machines; telecommunication traffic; D-S evidence theory; Internet service providers; data fusion; interconnected communications network; least squares-SVM; network loads; network online traffic data; network performance; traffic surveillance; Capacity planning; Communication networks; Data engineering; Educational institutions; Machine learning; Payloads; Support vector machines; Surveillance; Telecommunication traffic; Web and internet services;
Conference_Titel :
Information Engineering and Computer Science, 2009. ICIECS 2009. International Conference on
Conference_Location :
Wuhan
Print_ISBN :
978-1-4244-4994-1
DOI :
10.1109/ICIECS.2009.5366846
