Mapgrid: context aware grid middleware for mobile applications
Nalini Venkatasubramanian, Yun Huang · 2008
Ensuring acceptable Quality of Service (QoS) for mobile applications while optimizing system resource utilization (e.g. reducing server side storage/computing costs, network bandwidth costs and device energy consumption) is a difficult problem, as mobile environments are highly dynamic and heterogeneous. In this thesis, we show how to leverage heterogeneous and intermittently available grid resources as proxies to support mobile applications; specifically, we focus on devising efficient resource discovery algorithms and data placement strategies. We develop two main categories of solutions for resource discovery and data placement: (1) on-demand techniques for an individual request and (2) aggregated approaches for large numbers of mobile requests. In the first category, we exploit knowledge of an individual client's mobility pattern, device energy profiles and grid resource availability; we apply techniques from graph theory, neural nets, etc. to select optimal localized computational and storage resources within the grid to cache requested mobile data and to provide quality-aware mobile applications (e.g. streaming multimedia). We also propose an integrated solution that adapts to dynamic changes in device energy consumption and unpredictable grid resource availability without compromising application QoS. In the second category, we motivate the importance of applying aggregated mobile data access information in making request scheduling and data placement decisions. We present a novel construct, Mobile Data Overlay (MDO) that captures aggregated mobile data access patterns and proxy availability. We further develop intelligent MDO reconfiguration techniques that exploits spatio-temporal locality of mobile data accesses to determine the granularity of data replication. The reconfigurable MDO approach effectively balances tradeoffs between replication cost and data access cost in making mobile data placement and mobile request scheduling decisions on grid proxies. Finally, we present the design and implementation of MAPGrid, a prototype framework for using volunteer grid machines to service mobile requests. We also implement a driving application, video streaming to mobile users that seamlessly receives segmented streams from multiple grid proxies. We evaluate how different factors (e.g. the number of service proxies) affect the end user experience. We believe that these factors are relevant to effective deployment of next generation mobile applications, including location-based services and third-party storage services, etc.