Applications of Data Grouping for Effective Mobility

University Of California-Santa Cruz, Santa Cruz CA

Investigators

Abstract

Applications of Data Grouping for Effective Mobility Darrell D. E. Long The problem of reducing a mobile computer's communication and power requirements will be investigated. Specifically, this research will address both issues through improved storage management. Based on successes with automated data grouping and predictive power conservation, research will be conducted into improved data hoarding and disk power management techniques. Effective grouping of data will make it possible to improve the automation of mobile file hoarding, and decrease the effects of network latency and disconnections on the mobile user. Similarly, through the grouping and retrieval of related on-disk data, it is possible to improve disk power management beyond the theoretical limits of any previously attempted scheme. This will be done by actively modifying the access sequence to minimize power requirements. Through transformation of data access workloads, this novel approach reshapes disk idle times to effectively increase the amount of time a disk can be spun down. Improving the battery life of mobile computing devices is critical to their usefulness, and such power efficiency is clearly an environmentally responsible goal. This research will establish the effectiveness of on-line data grouping techniques to increase the amount of data accessed per operation, while simultaneously reducing the total number of disk operations.