Network coding techniques are usually applied upon network-layer protocols\nto improve throughput in wireless networks. In scenarios with multiple unicast sessions,\nfairness is also an important factor. Therefore, a network coding-aware packet-scheduling\nalgorithm is required. A packet-scheduling algorithm determines which packet to send\nnext from a nodeââ?¬â?¢s packet backlog. Existing protocols mostly employ a basic round-robin\nscheduling algorithm to give ââ?¬Å?equalââ?¬Â opportunities to different packet flows. In fact, this\nââ?¬Å?equalââ?¬Â-opportunity scheduling is neither fair, nor efficient. This paper intends to accentuate\nthe importance of a coding-aware scheduling scheme. With a good scheduling scheme,\nwe can gain more control over the per-flow throughput and fairness. Specifically, we\nfirst formulate a static scheduling problem and propose an algorithm to find the optimal\nscheduling scheme. We then extend the technique to a dynamic setting and, later, to practical\nrouting protocols. Results show that the algorithm is comparatively scalable, and it can\nimprove the throughput gain when the network is not severely saturated. The fairness among\nflows is drastically improved as a result of this scheduling scheme.
Loading....