Smartphones as Alternative Cloud Computing Engines: Benefits and Trade-offs

Abstract

Over the past decade, both the popularity and computational power of smartphones have grown at fantastic rates. These devices now represent a wealth of untapped computational power. There have been efforts to incorporate smartphones into volunteer grid frameworks as engines of computing in a manner similar to desktop computers. However, it is not clear what the energy implications of such an integration are. Phones take less energy to create and use less power while running than desktops, but they take significantly longer to complete a task. In this paper, we present an initial investigation of the tradeoffs of integrating smartphones into the cloud. We created a system to generate computation loads on both smartphones and desktops. In an empirical study, we monitored the energy consumption of both systems under similar computational loads. The results of our study show that the desktop could finish our task 10 to over 30 × faster than the smartphones. However, on specific benchmarks, we observed that the desktop consumed up to 2.26 × more energy than the most efficient phone.