Study on Reducing Implementation Complexity of SRR Scheduler(APCIM2012 Best Papers)

書誌事項

タイトル別名
  • Study on Reducing Implementation Complexity of SRR Scheduler

この論文をさがす

説明

This paper proposes two techniques for reducing implementation complexity of SRR (Smoothed Round Robin) scheduling algorithm. The SRR scheduling algorithm forms a WM (Weight Matrix) by coding the weights of the flows into binary vectors and then schedules packets by scanning the Weight Matrix by using WSS (Weight Spread Sequence). WSS is a specially designed sequence that evenly distributes the output traffic of each flow However area complexity for storing WSS is O(k*2^k), where k is the encoded bit width of weights. We proposed a WSS generation technique with area complexity of O(k^2). It generates WSS by simply in-order depth-first traversing of a graph with k nodes. We also propose an efficient comparator scheme for flow control between queues. The proposed algorithm was modeled and implemented in LINUX using C language. Experimental results show that the proposed technique is effective to reduce the area complexity of WSS generation.

収録刊行物

詳細情報 詳細情報について

問題の指摘

ページトップへ