The need for data warehousing is growing ever more acute as companies seek to leverage their information assets to better support their business. However, what this means in practice, and in very simple terms, is that you have more users querying more data that it is necessary to deliver in a shorter timeframe (and often in real or near real time).
This poses a number of problems for data warehouse providers and users. The biggest is the fact that while the use of dual core and quad core processors means that Moore’s law (that processing capacity doubles every 18 months) continues to be met, the same is not true of disk speeds, where improvements are more of the order of 5 or 10%. Worse, it is estimated that storage requirements are doubling every year. How then, is this circle to be squared? How can organisations meet their users’ demands for the extended access that they are demanding? This paper sets out to explore the possibilities.
However, this is a white paper with a difference. It is not intended to be merely a discussion of data warehousing but an active guide on how to approach the current issues you may have with your existing data warehouse implementation. In particular, it uses a dialogue-based approach along the lines of “if you have this problem then”. (See the diagram in the next section which traces the decision points in this dialogue.) Note that we are assuming that you already have a data warehouse as green field implementations would require a different set of questions. For the same reason we will not be discussing, at least in any depth, what you might do if you decide to completely replace your existing database.
Leading on from this last point, we will also not be discussing what you might do if you are using a data warehouse based on Informix, RedBrick or some other database that is no longer treated by its vendor as a primary data warehousing platform. While you might add a data mart from, say, Netezza, or you might upgrade your hardware, ultimately this will be no more than a stopgap: if you are having concurrency, scalability or performance issues now, with a database that is unlikely to be significantly improved in the foreseeable future, then eventually you will have to replace that system.
What we will be discussing is those environments where you have a viable potential future with your existing supplier and we will consider what you might do, depending on the issues that you face. Note that while this paper has been sponsored by Dataupia it is not limited to discussions of when that company’s product might be suitable for your environment.