| Summary: | 碩士 === 國立中央大學 === 資訊管理學系碩士在職專班 === 97 === Internet has been prospering since the end of the 21st century making the Communication between people more close. It brings the rapid growth of the network traffic and also a most knotty problem for internet service providers, that is, how to control their own network quality more efficiently.
Network resources and the circuit bandwidth are limited, whereas users’ demand never be met. Such condition often causes the network bandwidth congestion thereby lowering the service quality. Therefore, internet service providers try to find means to bring more network resources for their customer and reduce the expenditure for businesses itself. Ways to produce resources include buying the new network equipments, building the new control rooms, etc.. However, network suppliers aren’t able to invest all resources on these endlessly.
The primary task reducing the expenditure is to control the network quality more strictly in order to provide better quality of the packet transportation and improve the network performance for business, especially who signed the Service Level Agreements which will guarantee four service items: circuit bandwidth, network availability, packet latency, packet loss ratio.
However, a large-scale service provider has thousands of different network facilities of varied brands including Cisco, Foundry, Juniper, Alcatel, etc. with divergent routing protocol runs on it. In such a complex network environment, once there are any slight changes happened without noticing, it will affect the quality of the network service.
This research through interviews is to investigate the actual working condition of how an enterprise can monitor its network infrastructures with network quality indicators. We hope to understand the rules of how a network supplier monitors their own network, which includes uses of programs, applications, control flows, and benefits after implementing. Finally, we analyze the case and draw a conclusion to establish a more complete evaluation model of the network quality.
|
|---|
