Generating the identifiers of XML nodes is a crucial task in XML applications. On the other hand, the structural information of XML data is essential to evaluate the XML queries. Several numbering schemes have been proposed so far to express the structural information using the identifiers of XML nodes. In this paper, we introduce a new numbering scheme called recursive UID (rUID) that has been designed to be robust in structural update and applicable to arbitrarily large XML documents. We investigate the applications of rUID to XML query processing in a system called SKEYRUS, which enables the integrated structure-keyword searches on XML data. Experimental results of the performance of SKEYRUS are also reported.

An improved movement-based registration scheme reduces location update cost on conventional movement-based location update scheme but increases uncertainty of the position of a mobile terminal (MT). The effect of this uncertainty of the MT position on the paging cost of the improved movement-based registration scheme is analyzed. We show that the paging cost of the improved movement-based registration scheme in [2] is underestimated, especially for large values of movement threshold. The result shows that the underestimation cannot be ignored.

We propose a location management, 'Mobility/Traffic Adaptive Location Management,' that dynamically changes the Location Area (LA) in response to the terminal's mobility and traffic characteristics to reduce the location management traffic in the air. In this paper, we investigate the following subjects for realizing the proposed location management. First, we present a method of calculating the optimum LA size. Second, we present methods for dynamic LA set up. Third, we present the architecture of the proposed location management. Finally, we evaluate its performance.

It is very important to accurately and effectively track the movement of each mobile terminal in the personal communications networks. Accordingly, various dynamic location update schemes reflecting the mobile terminal's moving pattern are proposed in the literature. Important schemes include time-based, movement-based, distance-based, and direction-based location update schemes. In this paper, we intend to improve the performance of the movement-based location update scheme which is simple to implement and shows good performance. The proposed location update scheme reflects each mobile terminal's movement characteristics and reduces the number of unnecessary location updates. The performance of the proposed scheme is evaluated and compared with the previously proposed movement-based location update scheme through simulation. The proposed location update scheme requires three memory elements to maintain the movement history in the mobile terminal. The simulation results show that the location management cost is significantly reduced especially for the case with small call-to-mobility ratio by minimizing the unnecessary location updates intrinsic in the movement-based location update scheme.

In mobile broadcasting environments, an information server periodically broadcasts a set of data items to a large mobile client population at every broadcast cycle and mobile clients retrieve the data items they need upon arrival at the broadcast channel. In such environments, the cost of data delivery is independent of the number of clients. Many applications such as auctions and stock quotes perform read-only transactions that require the clients to read consistent and current data for accurate result. Previous concurrency control mechanisms designed for such environments ensure that the clients read consistent data, but they cannot ensure latest-bound currency which allows the clients to read the latest or most up-to-date data. In this paper, we propose an efficient concurrency control mechanism that ensures latest-bound currency as well as update consistency, which is appropriate for the mobile broadcasting environments. To ensure latest-bound currency, the server computes control information using "virtual" broadcast cycles. This control information is also used for checking update consistency. Thus, the proposed mechanism allows all data committed in current broadcast cycle to be broadcast. We have performed simulation experiments to measure transaction aborts in order to evaluate the performance of the proposed mechanism. The result confirms that the proposed mechanism produces no stale reads and also shows that the proposed mechanism generates less transaction aborts than previous mechanisms, which implies that we can get higher data currency without increasing data inconsistency.

In this paper, we propose a call arrival history-based location tracking strategy for a variable call arrival rate over time. The basis of the proposed strategy is a time-based location tracking strategy. A mobile terminal obtains the up-to-date information about changes in the call arrival rate by maintaining its call arrival history, from which it can calculate an appropriate timeout interval for a variable call arrival rate. We present a simple analytical model and numerical results to investigate its performance for both a fixed and a variable call arrival rate which is modeled by a Markov-modulated Poisson process.

In logic programming or functional programming languages, data objects, such as terms and lists, are immutable. In a basic implementation of such language, updating one element of an aggregate (contiguous data structure, such as an array) involves making a new copy of the whole aggregate. However, such copying can be expensive, and can be avoided by using a destructive update. We introduce the concept of a wrapper which enables destructive operation on an immutable object. Based on this concept, we designed the reversible functor as a solution to the aggregate update problem. We implemented the reversible functor in the existing SB-Prolog system and carried out several benchmarks. These benchmark results show its effectiveness. When using a large functor and updating it many times, the performance is improved dramatically by implementing the reversible functor. It incurs some overhead at runtime, but the amount is small and acceptable.