I am writing my master thesis about QlikView's In-memory approach. Can someone tell me, how this approach works? Why i don't need OLAP-Cubes? Are there any facts that document, that this approach is better than usually approaches like cubes, where i define a star or snowflake schema?
You don't need cubes because everything is in memory and it can search its internal data structures fast enough that getting the data for a chart is... well, vaguely like running SQL against a database, but in real time, even with tens or hundreds of millions of rows.
So far as I know, though, the underlying technology is proprietary, in a "we don't have to tell you how it works, and can change it at any time" way. Even if they were kind enough to share enough so that someone could write the above post, I suspect you're going to have quite a bit of trouble getting enough information to put together a master's thesis on it. I hope that your thesis is at least on something more general, like "in-memory business intelligence" rather than "here's how QlikView's proprietary technology works".
my thesis is not only in memory analysis, as you said, it is about in-memory BI. i have to explain BI (where it comes from, the bi-framework, olap, star/snowflake schema and at least in memory). but in the main part i should use qlikview to document my approach especialle the in memory part. but your link seems to be good. thank you:)
