Mobile Computing and Databases Systems

Mobile and wirless networks introduce many new sub-problems in distributed computing and information systems that are not present in traditional computer networks, distributed databases and computing systems. We study the effects of this new environment in the design and development of mobile wireless networks and database systems.

An example of a mobile information and computing system is the military inventory management that provides total asset visibility through accurate tracking and recording tens of thousands of items that are highly mobile over large distance in rapidly changing environment such as the battlefield. An effective solution to this problem is to make each item in the inventory an active entity with Radio Frequency Identification (RFID) package tags that may communication with many other tagged items and mobile support computers via terrestrial and satellites wireless network links. The inventory control system must interface seamlessly with existing large military management information systems used for command and control operations during joint field/force operations.

We have studied the problem through characterization of wireless networks and distributed mobile system, concept study and development, model design and analysis, simulation of behavior, and evaluations of results. We first examine features and properties of the the physical distributed mobile environment in which the Army is required to track and record tens of thousands of mobile items. Then we survey and analyze new concepts that should be adopted in designing various novel techniques in different areas of the system. These new concepts are studied by building analytical and simulation model. The behavior of the new concepts represented by the models are then simulated using computer simulation software. We evaluated the different alternative concepts and techniques by comparing the results of these simulations under different set of operating conditions or workload. There must also be appropriate repositories and information caches to maintain most of the voluminous data from many sources in the radio computer network. We are also re-examining efficient support for transactions in mobile logistics databases involving mobile units. Client server cache strategies must be modified to account for mobility and changes in network attributes of both clients and servers since communication speed and reliability will be adversely affected. New methods for maintaining consistency and recoverability of data may be required.