Hardware developments allow wonderful reliability and essentially limitless capabilities in storage, networks, memory, and processing power. Costs have dropped dramatically. PCs are becoming ubiquitous.

The features and scalability of DBMS software have advanced to the point where most commercial systems can solve virtually all OLTP and DSS requirements.

The Internet and application software packages allow rapid deployment and facilitate a broad range of solutions.

Result: We now have the hardware and software to really solve the world’s data management problems. And it is happening: data volumes are growing tremendously and applications are booming.

Problem: Database vendors have succeeded on the traditional measures of VDLB, parallelism, object-relational extensions, OLTP, etc, but they have failed to make their systems easy to use. Most attempts towards providing a simpler model are aimed at a small subset of the entire problem, usually tuning/administering large installations. This is fixing a symptom not the true problem. We need to provide a much broader, integrated solution to the problem of managing the data that will be produced over the next 10 years.

This paper describes the complexity within current database systems and some of the reasons for the current state. Then follows a discussion on many of the dimensions of a broad solution. Finally there’s a section describing Microsoft’s attempts towards an integrated solution.