There have been a variety of researches for the metro networks which connect the access and the backbone networks as the Internet traffic increase. Technologies such as SONET and Ethernet have been the basic means for the metro areas, but it is difficult for these technologies to be adapted to the metro networks due to their unique features. On the other hand, recently Resilient Packet Ring is a representative method to adopt the advantages of them with activities for standardization. Even if RPR can be a solution for provisioning QoS in the metro networks, it is not considered in viewpoints of service users. Therefore, we suggest a scheduling algorithm in order to guarantee quality of service in terms of service user. In addition, we take a look at the proposed algorithm``s performance by comparing itself with WFQ.
In this thesis, we use delay from a source node to a destination node, namely, end-to-end delay, as a criterion to estimate quality of service. To provide real time service, end-to-end delay is an important factor, which affects quality of service, and thus must be limited to a specific value according to the types of service applications. Therefore, we define delay constraint as the limited value of delay to be required in processing a packet in our proposed scheduling algorithm, and calculate the amount of bandwidth to be allocated to each queue by using this value, which is different according to the kinds of applications. In addition, in our thesis, we also explain a dynamic scheduling method, which makes use of queue length of the previous frame to calculate the amount of bandwidth.
We deal with a simulation in two points of views. First of all, we make an observation of the characteristic of delay to ensure that our algorithm complies with delay constraints. And secondly, we measure loss probability in order to feel sure that our algorithm reduces or increases packet loss by dynamically allocating bandwidth. As a result, the simulation sho...