Inside OLAP Engine: SE cache sharing
After publishing the excerpt from my upcoming whitepaper on AS caching, I received a lot of feedback, both in blog comments and in email. I want to thank everybody who took time to read and assess that piece. Honestly, I was a little bit afraid, because I know that I have tendency to overcomplicate and confuse people when I explain something, but so far the feedback indicated that the technical level was OK, and everybody felt comfortable with the content. Several people asked if they can get preview access to the whitepaper. I, of course, would've done it, but the biggest and most important part of it - namely how to write cache friendly calculations - is not yet written. Everything else is an interesting knowledge, but that part are concrete steps that people can do in order to improve performance by writing their calculations in such a way that they take most benefit of the caching system. While I have a good idea about what I should put there, it is somewhat more difficult material, and I struggle with writing it up. However, I thought that I could share another excerpt here. It seems that readers of my blog are hardcode AS users, and releasing pieces in the blog would be mutually beneficial - readers may find this material useful and I will get free peer review. So in a sense, publishing this staff in the blog is my public "Beta" of the selected material from the whitepaper. This week I decided to publish portion about SE cache sharing. One more note before I start. Some people expressed hope that the acronyms and special terms that are used in the excerpts to be explained. And indeed, the whitepaper will explain all of them. For the purpose of this (and previous piece), SE means Storage Engine - component that deals with processed raw data before calculations apply. FE means Formula Engine - component which deals with all the calculations.