Adaptive Fair Resource Allocation for Energy and QoS Trade-Off Management

Fumiko HARADA  Toshimitsu USHIO  Yukikazu NAKAMOTO  

IEICE TRANSACTIONS on Fundamentals of Electronics, Communications and Computer Sciences   Vol.E91-A    No.11    pp.3245-3252
Publication Date: 2008/11/01
Online ISSN: 1745-1337
DOI: 10.1093/ietfec/e91-a.11.3245
Print ISSN: 0916-8508
Type of Manuscript: Special Section PAPER (Special Section on Concurrent/Real-time and Hybrid Systems: Theory and Applications)
quality of service,  energy consumption,  dynamic voltage scaling,  resource allocation,  

Full Text: PDF(422KB)>>
Buy this Article

In real-time embedded systems, there is requirement for adapting both energy consumption and Quality of Services (QoS) of tasks according to their importance. This paper proposes an adaptive power-aware resource allocation method to resolve a trade-off between the energy consumption and QoS levels according to their importance with guaranteeing fairness. The proposed resource allocator consists of two components: the total resource optimizer to search for the optimal total resource and QoS-fairness-based allocator to allocate resource to tasks guaranteeing the fairness. These components adaptively achieve the optimal resource allocation formulated by a nonlinear optimization problem with the time complexity O(n) for the number of tasks n even if tasks' characteristics cannot be identified precisely. The simulation result shows that the rapidness of the convergence of the resource allocation to the optimal one is suitable for real-time systems with large number of tasks.