ABSTRACT
We have discussed a detailed framework for hybrid scheduling in asymmetric wireless environments. The scheduling and data transmission strategies can be broadly classified into push and pull scheduling schemes. However, both of these push and pull scheduling strategies suffer from some specific disadvantages. Hence, a hybrid scheduling that explores the advantages of both push and pull scheduling is more attractive. We first develop a basic hybrid scheduling scheme that combines the push and pull scheduling schemes independent of the build-up point, i.e., without restricting the pull queue size to be 1. Our hybrid scheduling system, uses push scheduling to broadcast the popular data items and takes the help of pull scheduling to transmit the less popular ones. The system computes the packet fair scheduling (PFS) for push system and accumulates the clients’ request in the pull queue. The pull system works on most request first (MRF) scheduling. The system alternatively performs one push and one pull method. The cutoff point, that segregates between the push and pull scheduling is chosen in such a manner that the overall average access time experienced by clients is minimized. This hybrid scheduling strategy is enhanced to incorporate heterogeneous data items (i.e., items of variable lengths). Although the basics of the push schedule remain unchanged, pull scheduling now must also consider the item lengths. The underlying reason is that items of variable lengths have different service times. This leads us to use the stretch optimal scheduling principle to choose an item from the pull queue. Performance analysis and simulation results point out the efficiency of this heterogeneous hybrid scheduling scheme.
