Smart caching for efficient information sharing in distributed information systems

Abstract

Remarkable technical advances in cell phones and smart phones have resulted in a worldwide marketplace permeated by mobile devices. These capabilities, in combination with increasing consumers demand to share information show the need for utilizing the mobile devices more efficiently. But, within a distributed network of mobile devices like TwiddleNet, the two most limiting resources are still battery power and bandwidth. By distributing only small sets of data that represent the actual content, use of these resources can be reduced. The information tags can be sent throughout the network, reducing the amount of traffic to share the information. But, once the content itself is shared, the workload on the mobile servers can quickly exceed the mobile device's ability to perform. This thesis offers an algorithm that will conserve battery power and bandwidth, depending on demand and device capabilities. Once a certain limit of bandwidth usage, or a certain battery level, is reached, the algorithm will select the content that will most efficiently relieve these two resources and temporarily upload it to a proxy server that will serve the content on its behalf. This "smart," temporary, caching will last as long as the bandwidth or battery level limits are exceeded.