CPU Scheduling in Data Centers Using Asynchronous Finite-Time Distributed Coordination Mechanisms

We propose an asynchronous iterative scheme that allows a set of interconnected nodes to distributively reach agreement within pre-specified bound in finite number steps. While this could be adopted wide variety applications, we discuss it the context task scheduling for data centers. In context, algorithm is guaranteed approximately converge optimal plan, given available resources, Furthermore, by being asynchronous, proposed able take into account uncertainty can introduced from straggler or communication issues form latency variability while still converging target objective. addition, using extensive empirical evaluation through simulations show method exhibits state-of-the-art performance.