A variety of technologies have been proposed and developed to efficiently manage communication network, which is becoming increasingly complex. However, none of them could satisfy all the requirements of management. Recently, Web Services (WS) has emerged as a promising XML based technology with a popularity of Internet. In this paper, I am going to take a close look at general WS system's structure. When we build WS system using Web Services Description Language (WSDL) standardized by W3C, I propose a new efficient design and implementation scheme keeping a whole WS structure such as Universal Description, Discovery and Integration (UDDI). My proposed scheme, Common Type Enveloping/De-enveloping Mechanism (CTEM/CTDM), makes it efficient to implement and to maintain WS systems. CTEM/CTDM scheme gives us a way not only to develop plug-in service implementations which are easily attached or detached if necessary but also to adopt legacy WS implementations easily by loose coupling between service interfaces and service implementations. The concept of CTEM/CTDM also provides us a way to quickly respond to client's additional development requests with low cost and small effort by mean of minimizing the scope of developing not to interface but to application.

Various types of computer systems are used behind the scenes in many parts of the telecommunications network to ensure its efficient and trouble-free operation. These systems are large, complex, and expensive real-time computer systems that are mission critical, and contains a database engine as a critical component. These systems share some of common database issues with conventional applications, but they also exhibit rather unique characteristics that present challenging database issues. Major DBMS issues for network management include choosing the right data model, handling two different kinds of data in terms of integrity and recovery constraints, supporting temporal queries, satisfying real-time performance and high availability requirements, and several miscellaneous issues. Some of these issues have been investigated in various areas of database researches, but most of them largely remain in the research stage. Advances in these areas that result in actual integrated implementations for data-intensive, real-time and temporal applications are eagerly awaited.

There are several issues to be investigated for implementation of database management systems with temporal support. Examples are the types of databases in terms of temporal support, handling of ever-growing storage size, use of optical disks, representation of temporal versions into physical storage, and efficient access methods for both temporal and non-temporal queries. This paper discusses characteristics of tuple versioning vs. attribute versioning, two distinct representation schemes for temporal versions. It then presents various forms of two level storage structures tailored for databases with temporal support.

A prototype of a temporal database management system was built by extending Ingres. It supports the temporal query language TQuel, a superset of Quel, handling four types of database static, rollback, historical and temporal. A benchmark set of queries was run to study the performance of the prototype on the four types of databases. We analyze the results of the benchmark, and identify major factors that have the greatest impact on the performance of the system. We also discuss several mechanisms to address the performance bottlenecks we encountered.