IBM is set to show System S today. Just what is System S? Well, “S” must be for streaming. According to the article, this is a system designed to not only gobble up mass quantities of data, but act on that data as it is doing the gobbling. This certainly could solve a problem that many industries face:
The conventional approach to computer analytics and data mining is to collect data, store it in a database program and then search the database for patterns or ask it questions. It is an effective approach, but also tightly structured and often time-consuming.
Bingo! The typical store and query approach doesn’t always work that well. This is especially true when both the quantity of incoming data and the number of analytical tasks required on each piece are very high.
The initial system is running with 800 processors (yes, 800) and can scale up from there. If 800 processors are required, System S will be out of reach for many corporate customers. The article does mention that IBM is entertaining the idea of selling this as a service (on the internet) in addition their traditional sales channels. If done right (IBM, take a look at Amazon’s services), selling this as a service could open this up to a very wide audience.