A research report on the laboratory for database systems research : past, present and future*

The Laboratory for Database Systems Research at the Naval Postgraduate School is devoted to research and experimentation on a wide-range of database topics. The major goal of the Laboratory is to maintain a constantly evolving environment that can be utilized by professors and students to conduct research. In this paper, we report on our past, present and future research efforts at the Laboratory for Database Systems Research. The past and present research that we report on is in four major areas, architectures for high-performance database computers, portable database systems, user interfaces for database systems, and methodologies for database systems. The future research we report on is in three areas, real-time database computers, multi-model database systems and multi-medium database systems. In all of our discussions, our main focus is on providing an overview of why we do our research with the secondary and rather limited focus on the actual details of the research. * The work reported herein ia supported by grams, from the Department of Defense STARS Program and conducted at the Laboratory for Database Systems Research, Naval Postgraduate School, Monterey, CA 93943. 1. AN INTRODUCTION The Laboratory for Database Systems Research at the Naval Postgraduate School is devoted to research and experimentation on a wide-range of database topics. Founded at the Ohio State University and moved to its current location in 1982, the Laboratory personnel consists of four professors, three Ph. D. students and fourteen Master's students. The major goal of the Laboratory is to maintain a constantly evolving environment that may be utilized by professors and students to conduct research. In meeting this goal, we make an extensive use of the Laboratory facilities at the Naval Postgraduate School. In addition, this environment utilizes some of the computing equipment of the Department of Computer Science. The entire computing environment is supported by a staff provided by the Department. In this paper, we report on our past, present and future research efforts at the Laboratory for Database Systems Research. The past and present research that we report on is in four major areas, architectures for high-performance database computers, portable database systems, user interfaces for database systems, and methodologies for database computers. Briefly, let us outline each of these areas. Our research in architectures for high-performance database computers is motivated by our disappointment with traditional approaches to database computer architectures. As we show in Section 2, we have found that there are some serious limitations in the performance of traditional approaches. Therefore, we have sought to develop novel architectures for database systems that stress high-performance as a primary design goal. In Section 2, we present our two novel approaches to architectures for high-performance database systems. The first approach is hardware-based, while the second approach is softwarebased. Both approaches achieve highperformance by the utilization of parallelism to increase the throughput. Our research with portable database systems has arisen from the recognition that database systems of the present and future must be able to run on different hardware and operating system configurations. No longer is it sufficient for a database system to be supported on a single type of hardware using a specific series of operating systems. Advances in networking and hardware technology have created a computing environment that is wide-ranging and variable. Consistent database support in such an environment mandates a database system that is capable of functioning on different hardware and operating system configurations. In Section 3 we report how we have designed and developed a highly-portable database system. Our research on user interfaces for database systems is motivated by the need to provide new functionality, in terms of user access methods, to database systems. In Section 4, we investigate the two distinct approaches in our user interface research, both predicated by a desire to overcome the traditional limitation of the single-data-model-and-single-model-based-datalanguage paradigm used by most database systems. In the first approach, we have developed many new one-dimensional (or textual) user interfaces, based on different data models and data languages, to provide the user with a wide-range of database access methods. In the second approach, we are developing a two-dimensional (or visual) user interface to provide the user with a graphical framework for database accesses.