Summary

Now you really are ready to go real-time with your cubes! To let you in on one of the arcane secrets of "real-time" anything — there is no such thing as real-time in computing, unless you plan to violate the cosmic speed limit (about 186,000 miles per second). There are varying degrees of latency, however, the shortest of which appear instantaneous to the user. Your job as an architect of Business Intelligence solutions is to design your application in such a way as to satisfy the needs of your users. That often means application of these techniques with particular attention paid to keeping the production system online and available during specified hours.

In this chapter you read about three real-world scenarios with one or more recommended solutions for each. These scenarios addressed solutions to long, average, and no latency requirements. A key take-away here is that you should remain calm in the face of massive amounts of source data; there are ways to contend with it. If you get it right, your users will remain blissfully ignorant of the challenges you faced in keeping up the illusion of real-time data presentation. Indeed, your business decision makers using the system won't know (or care) how the system is implemented, they'll know, "…it just works and has amazingly up-to-date data." That is where you want to be.