Load balanced adaptive routing with reduced overhead for Network on Chip (NoC) Systems
J. Kalaivani, B. Vinayagasundaram · 2013
The Network-on-Chip (NoC) has been emerging as a very promising paradigm for scalable on-chip communication architectures in many fields. To achieve excellent results in Network on Chip (NoC) systems application the load balanced in the routing has to be achieved. Normally, the routing is based on the probability function using the data of the past and current, which means the route may exist or may not, because of which the size of the routing table also increases. To overcome this drawback in the NoC systems, we propose K-Step Look Ahead in Load Balanced Adaptive Routing, where all the details about the routing the routing table is maintained by the every router are achieved by ACO-based Cascaded Adaptive Routing (ACO-CAR). These routing tables are formed based on the probability of pheromones left behind by the ant packets in the past. Then we implement a k-Step Look Ahead (KSLA) algorithm which finds the exact number of steps for the ant packet to reach the destination using the information of the ACO-CAR routing tables and updates the routing table by eliminating the route information which does not exists.