Formal probabilistic analysis of distributed resource management schemes in on-chip systems

New paradigms for managing resources in on-chip many-core systems come with various issues. Among them is the key demand for robust verification of (distributed) resource management (RM) schemes before deployment. Moreover, it is important to have a unified framework where different RM schemes can be formally analyzed and compared for their performance efficiency and robustness. Traditional techniques, like simulation or emulation, are inherently in-exhaustive and thus compromise the completeness and accuracy of the analysis results. In this work, we present a formal approach, based on probabilistic model checking, for evaluating and comparing the performance of different distributed RM schemes. To illustrate the benefits and applicability of our formal verification and comparative analysis approach, we perform a case study on the comparison of two state-of-the-art distributed RM schemes using the PRISM model checker.

I. Introduction

With current trends in hardware design, on-chip many-core systems [1] have emerged as a new paradigm. These multicore architectures execute highly parallel and resource demanding applications. This leads to the problem of runtime resource allocation, i.e., which application should use which and how many cores to get the most efficient utilization of available resources [2] [3]. These resources can no longer be managed by one central entity (i.e., in a centralized resource management paradigm) due to increased design optimization space of large-scale many-core systems. Therefore, many distributed Resource Management (RM) schemes have been proposed, e.g., [2] [4] [5] [6], over the past few years, to ensure an efficient utilization of the available resources and for maximizing the overall throughput of on-chip many-core systems in a scalable fashion.