Application of mobile user's location prediction to location server restoration

In mobile computing environment, it is important that location server maintains the exact current location of mobile users to locate the users for fast connection. Also, it is necessary to guarantee the availability of location information when one tries to connect to a mobile user. Thus, an explicit restoration against the location server failure is required for high quality of services. We captures some repetitive and regular patterns of user mobility by neuro-fuzzy inference system, and an aggressive restoration scheme using prediction is proposed to reconstruct a location information after location server failure. Our restoration scheme using prediction can locate the users lost by failure with less paging. And it does not have an extreme storage space and periodical backup process like in the checkpointing scheme.