Dynamic Control for Mixed-Critical Networks-on-Chip

Networks-on-Chip (NoCs) for future real-time systems must provide service guarantees for applications with different levels of criticality. In this work, we propose an efficient mechanism for supporting mixed-criticality which combines the global, work-conserving scheduling for the end to end guarantees with the local arbitration in routers. We introduce a dynamic control layer with a central Resource Manager (RM) synchronizing transmissions with a dedicated protocol. The proposed mechanism allows to improve over existing solutions through reducing hardware overhead compared to non-blocking routers with rate control as well as temporal overhead compared to Time-Division Multiplexing (TDM). By using formal analysis, we show that RMs provide efficient service guarantees to all synchronized applications. We validate experimentally, using benchmarks, these guarantees along with the performance of the mechanism and induced overhead.

I. Introduction

Networks-on-Chip (NoCs) [1] constitute the main communication backbone of modern multiprocessor systems on chip (MPSoCs) architectures. Considering their superior efficiency, they show high potential for use in safety critical real-time domains such as avionics and automotive. In such systems, safety- and time-critical applications are frequently integrated and executed together forming mixed-critical systems. This requires “sufficient isolation” between components with different criticality levels (e.g. required by avionics safety standard DO-178B [2]) as well as resource sharing for increased utilization and efficiency. The fulfillment of these demands requires providing guaranteed service for all criticality levels. In realtime NoCs, there exists two established solutions to tackle this issue: Time-Division Multiplexing (TDM) and non-blocking routers with rate control [3].