Ottawa, CANADA. June 10-15

This paper deals with the analysis of the impact of different scheduling algorithms on flow level mean delay performance of elastic traffic in a single cell wireless downlink data channel paradigm. An enhanced size-based and channelaware discipline is proposed, which is based on the sized-aware Gittins index approach applied to a time-varying channel context. In order to compare the proposed Opportunistic Gittins rule performance with well-known scheduling algorithms, several simulations have been performed for stochastically arriving flows with Pareto sizes under fading conditions, for different network loads and channel quality indicator reporting rates. As concluded, an approach that combines both flow size-awareness and channel awareness is the best option, which guarantees a trade–off between minimizing uplink overhead due to channel quality reports and improving scheduling performance aimed at reducing overall flow mean delay.