Dynamic resource allocation for adaptive real-time applications

Dynamic resource allocation is a critical component in any system expected to deliver predictable performance while subject to unpredictable workloads. Examples include military and commercial systems integrating multiple modules servicing sensors and actuators, or providing Internet-based services and advanced human-computer interfaces.; Driven by the goal of improving long-term system performance, previous research has proposed solutions for dynamic resource allocation using load migration and application-specific adaptation capabilities. This thesis improves on such work in two ways: (1) by considering the perturbation effects of adaptation and (2) by enabling complex adaptations by use of general models of target systems. Specifically, the proposed adaptation methods evaluate the timing failures that may be caused by adaptation enactment and handle complex adaptations and dynamic sets of system resources and services.; Our solutions are based on novel models that capture both (1) application-specific information about acceptable adaptations and associated costs and (2) service-specific information like physical resource requirements and timing constraints. Based on this information, adaptation algorithms can determine all appropriate adaptation choices, evaluate their impacts on both long and short term timing constraints, and perform reallocations and/or adaptations such that the perturbation of the running system remains small.; The proposed solutions are part of a prototype framework, called FARA, which provides the basic mechanisms for building adaptation and resource allocation infrastructures for both centralized and distributed systems. FARA is used with two adaptive applications—a video transmission system and a graphics application that emulates sensor-processing code. Experiments show that the proposed models can describe diverse adaptation capabilities and enable adaptation decisions with reduced impact on pending timing constraints, while also resulting in long-term performance improvements.