Adaptive Exploration Based Routing for Spatial Isolation in Mixed Criticality Systems

Abstract

Applications of different criticality are increasingly sharing the same System-on-Chip platform to be cost and resource effective. On such mixed criticality systems, spatial partitioning of resources is a commonly utilized technique to prevent interference between applications. At the communication level, Network-on-Chip (NoC) used in such systems can aid by isolating network traffic within application regions. Topologies that can develop in such partitions can be regular or irregular requiring minimal and non-minimal routing. For the NoC to be flexible and support such varying network parameters, it is desirable that the routing algorithm can support communication for all possible topologies. Here, we investigate a topology agnostic routing algorithm based on Ant Colony Optimization (ACO) metaheuristic. The routing algorithm explores the NoC for feasible paths using special ant packets and discovers paths based on history of already utilized paths and local traffic information. We aim to decrease the exploration time overhead, by proposing an adaptive exploration technique. Compared to the static version, the proposed technique can decrease the exploration time overhead by upto 68% while maintaining comparable latency and throughput.