Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
4962799 | Sustainable Computing: Informatics and Systems | 2016 | 78 Pages |
Abstract
In this article, we propose an energy-aware load monitoring model, called e-Mon, for enabling energy-aware load balancing in Peer-to-Peer (P2P) systems. P2P is a scalable and self-organizing technology for utilizing computational resources of the end-user devices for the benefit of a computing system. In P2P systems, the need for fair balance of load is crucial since the end-users need to be incentivized to participate in the system. The short battery life, caused by additional strain on the computational resources of the end-user devices, is a significant negative incentive factor for mobile end-users of current P2P systems. The e-Mon model, proposed in this article, enables moving load from energy-critical to less energy-critical nodes in P2P systems. This is done by including the energy status of a peer node as one of the factors defining a node's load. The model helps saving the energy of mobile P2P nodes, particularly in cases when the remaining battery capacity is low. The article provides a thorough energy-efficiency evaluation demonstrating that e-Mon can significantly improve the battery life of mobile nodes by improving the quality and fairness of load balance between heterogeneous nodes. With a proper selection of a load balancing model for the application scenario, e-Mon is shown to achieve up to 470% battery life extension compared to the case with traditional load balancing with no battery monitoring.
Related Topics
Physical Sciences and Engineering
Computer Science
Computer Science (General)
Authors
Erkki Harjula, Andrei Gurtov, Timo Koskela, Timo Ojala, Mika Ylianttila,