On the Stability of Best Effort Flow Control Mechanisms in On-Chip Architectures

In this paper we present a centralized flow control scheme in NoCs in the presence of both elastic and streaming flow traffic paradigms. We model the desired best effort (BE) source rates as the solution to an alpha-fair utility maximization problem which is constrained with link capacities while preserving guaranteed service (GS) traffic requirements at the desired level. We propose an iterative algorithm as the solution to the aforementioned problem which has the benefit of low complexity and fast convergence. We also explore the stability and convergence behavior of the proposed algorithm and prove that it is globally asymptotically stable. Such an algorithm may be implemented by a centralized controller with low computation and communication overhead.