Optimal Scheduling for Network Coding: Delay v.s. Efficiency

Network Coding can greatly increase the network transmission efficiency by combining multiple input packets into one packet algebraically before forwarding. However, the first arriving packets need to wait for the later ones before coding, which may cause additional delay. In this work, we will investigate the optimal queuing and scheduling strategy that can minimize the delay introduced by network coding, as well as the tradeoff between delay and efficiency of network coding. We will characterize the arriving packet processes into two classes: Uniform and Random. For uniform arriving processes, we will put forward the delay-minimized queuing schemes for both equal rates and unequal rates flows and analyze their performances. For random arriving processes, a half-opportunistic scheduling strategy that can jointly control delay, queue lengths and efficiency will be proposed. Finally, we will study the effect of processing time on the whole queuing system.