Title: "Instantly Decodable Network Coding: Where We Stand and Where We Head"
Abstract
Instantly decodable network coding (IDNC) emerges as a promising technology to improve the bandwidth efficiency of packet transmission and recovery in wireless multicast networks. Despite its guaranteed superior performance over conventional schemes, achieving the optimal IDNC performance is NP-hard for many parameters, such as the completion time (i.e. throughput) and decoding delay. The first part of this talk will introduce the IDNC layered graph representation, and show how it can be adapted to formulate, analyze and achieve near-optimal performance for these parameters in centralized networks. The last part of the talk will shade light on some of the very interesting open areas of research in IDNC. One important area is the migration towards distributed IDNC, which will open venues for massive contributions in many types of wireless networks, such as sensor, vehicular and cognitive radio networks. Exploring the buffering of non-decodable packets in IDNC is another growth area that can result in strong information theoretic achievements, such as reaching rate optimality in multicast networks, with simple approaches.