Caching Techniques In Mobile Real-time Computing Environments

The traditional distributed computing and distributed database is based on limited network and fixed host. It adopts some tolerant crytic hypothesis, such as fixed network connecting, peer to peer communication cost, fixed host etc. But in mobile real-time computing environment, many computing node can connect with network as they move. Then above-mentioned conditions do not exist any more. The technology of traditional distributed database can not apply to mobile database absolutely. Caching is a classical technology to improve the performance of data access. It has been successfully applied to different fields of computing technology. Caching will be even more important in mobile systems because of resources limitations of mobile devices and wireless networks.Data hoarding technology means the technology of prefetching data which may access by user after disconnection later to cache. The aim of data hoarding is that mobile client can manipulate data independently by local cache. Much schemes proposed for data hoarding focused on higher hit ratio of cache when mobile clients were in disconnected state, while scarcely thinking about the utilization of limited battery energy of mobile client and limited wireless network bandwidth. The data item value-evaluated scheme evaluates the system performance based on the stretch of a request, defined to be the ratio of the response time of a request to its service time. The scheme defines a function to get a value of each data item that may be hoarded. Hoarding items with high value computed by the function can achieve lower stretch than any other data hoarding schemes and can achieve better performance. Based on the data item value-evaluated scheme, dynamically adjusts the number of hoard to get better trade off between performance and power.Consistency of cache means the data items in mobile client must keep consistent with that in database in fixed hosts. Due to the limitation of wireless network bandwidth, a large amount of data transfers should be avoided and effective cache invalidation strategies are required to ensure the consistency between the cached data at the clients and the original data stored at the server. Most of strategies based on invalidation report suffer from long latency. And, limited bandwidth can not be utilized efficiently. Based on BS scheme, the method proposed can significantly reduce the query latency and improve the bandwidth utilization, using middle invalidation report and broadcasting requested data items.