Dynamic index coding gain over a complete bi-directional side information graph

Abstract

Dynamic Index coding as a generalization of the traditional index coding problem, deals with transmission of some dynamic data flows in the presence of side information. For a dynamic index coding system, the set of average flow rates for which the system can be stable is named Code-Constraint Capacity region. In this paper, we identify the code-constraint capacity region for a dynamic index coding problem with complete bi-directional side information graph. Furthermore, we introduce the concept of Dynamic Index Coding Gain to quantify the improvement resulted from the dynamic index coding with respect to uncoded transmissions. We also provide some corollaries about system performance metrics such as delay and average number of queued packets. In addition, we show how time-sharing method may reduce transmission delay.