The architecture of a database computer - a summary

Abstract

The motivation for seeking hardware solutions to database management functions traditionally carried out by software has been apparent to data-base designers for sometime now. Firstly, database management software has grown in complexity and size over the years. This growth is prompted by the increase in user requirements, by the formulation of sophisticated models and by the change in data processing mode from an off-line, batched, single user environment to an on-line, concurrent and multi-user environment. Large and complex software systems tend to be failure-prone. Further-more, practical verification methods for software systems are still not in sight. On the other hand, methods for verifying hardware functionality, design and production have long been available. Advanced technology has also overcome some of the problems of the logic complexity and capacity requirements, making the construction of relatively large and complex computers viable. By incorporating basic database management functions into hardware, not only can we provide more reliable basic functions, but we can also improve the software reliability since the software requirements will be less complex and the system software will be smaller in size.