Power-Aware Public Resource Management

Abstract

The power of the crowd, more precisely crowd-sourced resources, is in its ubiquity. Accounting for traditional desktop/laptop computers and recent mobile computing devices including tablets and smart phones far surpasses the number of servers in cloud data centres. Besides, the capacity and capability of these resources owned by the crowd (crowd-sourced resources) has increased dramatically with GPUs in particular. Although the public-resource platform has enabled many public-resource (or volunteer) computing projects including SETI@home and Milkyway@home, the lack of more realistic incentive mechanism. Without financial reimbursement, public-resource computing could not attract more participants as high running cost from power usage of the CPU and Graphics Processing Units(GPU). Inspired by Bitcoin network, a paid public resource computing, which resolved the scalability issue and has a sustainable mechanism to improve participating rate. In this thesis, we present Crowdware, a framework for enabling sustainable GPU-based public-resource computing with a realistic financial incentive mechanism. We design an auction-based resource allocation algorithm and a profit-based resource switch algorithm, explicitly taking into account the electricity cost of participating resources. Our results show that Crowdware greatly promotes profitability and cost efficiency for resource providers and resource consumers, respectively. Specifically, Crowdware has enabled the execution of MD5 password recovery jobs. In our testbed, with only 2.2% of the cost of using Amazon EC2 GPU instances while the participation of crowd-sourced resources is profitable with an average profit rate of 9.2%. Crowdware also shows great scalability with its fat-client and thin-server design. Together, Crowdware significantly improves the sustainability of public-resource computing.