| Summary: | 碩士 === 國立臺灣科技大學 === 電子工程系 === 88 === It is commonly believed that Integrated Service (IntServ) is not scalable to large networks. In this thesis, we study the association of Multiprotocol Label Switching (MPLS) and IntServ to try to overcome the scalability problem. The goal is to provide Quality of Service (QoS) in an efficient and scalable manner in the Internet through enable Label Switching Routers (LSRs) to be able to identify the appropriate reservation state for the aggregated IntServ flows based on the corresponding label. IntServ data flows are aggregated and mapped to one data flow in the MPLS network. IntServ data flows belong to the same Forwarding Equivalence Class (FEC) are aggregated.
The proposed scheme of integration of IntServ and MPLS is efficient and scalable. Efficiency implies that when flow aggregation is employed, one can use less resource to achieve the same required QoS, and less time is spent in transmitting packets. Scalability means that the aggregation of IntServ flows makes IntServ works well even when the volume of the data flow becomes huge. Controlled-load service and guaranteed service are two service types of IntServ. According to the different characteristic of these two services, First Come First Serve (FCFS) and the Rotating Priority Queue+ (RPQ+) packet scheduling are employed for aggregated controlled-load and guaranteed-service flows, respectively.
The Taiwan Academic Network (TANet) and very High Speed Backbone Network Service (vBNS) are used as our simulation environments. Simulation results show that because of flow aggregation, more flows could be served meanwhile the scalability problem of IntServ is also mitigated.
|
|---|
