With the growing ubiquity of networked computers, there is an ever-increasing potential for executing computations by utilizing distributed resources. However, coordinating delivery of resources to distributed computations is a challenging problem. The challenge is even more pronounced when these computations are executing in open distributed environments, because resources as well as computations needing them can dynamically join or leave the system at any time.  The approach I take is to separate the functionality concerns of concurrent computations from those of coordinating their resource use, with the purpose of reducing programming complexity and aiding development of correct, efficient, and resource-aware concurrent programs.

In this talk, I will present my research on resource coordination in open distributed systems. I will begin by introducing the Actor model of concurrency, and then I will describe the Distributed Resource Estimation and Allocation Model (DREAM), which I developed for representing progress of concurrent computations using distributed resources. Next I will describe the mechanisms and tools I have built for deciding and enforcing fine-grained resource use schedules for deadline-constrained distributed computations. For example, I have developed a tuner, which explicitly balances the overhead of the control mechanisms against the extent of control exercised. Finally, I will present my recent work on extending and applying this approach to multi-core processors to support energy conservation, by calculating fine-grained processor frequencies on a per-core basis.