Information on OLAP Databases?
oh-lap asks: "One relatively new topic in computing that hasn't seen much mainstream press (yet) is OLAP or multidimensional databases. OLAP is based on the principle of organizing your data along a number of dimensions which can be pivoted and drilled into. OLAP sounds like a neat idea and represent an abstraction of several problems which I've seen tackled by custom coding jobs. The biggest problem with OLAP right now, seems to be a dire lack of expert knowledge in the area and no standardized API or query language for OLAP servers. But the lack of decent introductory reading material is particularly bothersome. If there are any experts (or even dabblers) in OLAP out there, what are the sources of information (books, websites, forums, webblogs, etc) that you go to for info about OLAP? The issues I'm interested in are wide ins scope and include everything from querying and data modeling to actual design of an OLAP server. Any pointers"
Check out Ralph Kimball's site. A good place to start is his book, The Data Warehouse Toolkit. I haven't read the 2nd edition, but the first edition is full of good practical examples of dimensional modeling.
I'm going to preface this by saying I'm an expert in this field, so my comments may be biased.
OLAP concepts have been around for quite some time, although as a mainstream product only since the early '90s. Today the two major players in the market are Microsoft and Hyperion Solutions, with Cognos a distant third. (To be fair, there are plenty of other players, these are just the three largest). I personally have worked extensively with both Essbase and Analysis Services, and can honestly say that both have very strong points, but generally speaking I have found Hyperion's Essbase to be superior as a whole to Microsoft's Analysis Services.
The original thread commented on how little mainstream press OLAP has received, which I wholeheartedly agree with. However, it is worth noting that many, if not most, large businesses have OLAP implementations (I can't find a source, but if I remember correctly, 499 of the Fortune 500 companies have an implementation.). Considering the market is about $3.5 billion, the products are definately being used extensively. Granted that is a drop in the bucket of the total relational marketplace, but it is hardly insignificant.
Personally, I strongly believe the lack of knowledge outside of the Finance departments of the world about the concepts is directly tied to the lack of exposure the concepts receive at the collegiate level, which continues to focus on relational databases as the primary storage of enterprise data (I am not implying that OLAP database can replace relational databases, they are truely a complimentary product tailored for specific use). The lack of a standardized API & querying languages is definately a problem for the industry, one which the large players have tried to address, with limited success. On the flip side, the industry has pushed (rightly or wrongly) toward selling the total package, with both the backend database and a suite of front end data entry & reporting tools all in one. This makes for an easy sell to many customers, but doesn't encourage much innovation from the outside (IMHO, the front ends to OLAP products are the weakest link, and really do need the most development).
Ironically, one of the best online resources I've found is IBM's web site. They repackage Hyperion's Essbase as DB2/OLAP, so their Redbooks on the concepts are slanted toward Essbase, but they have the most extensive source of free information on the concepts I've seen from a commercial vendor.
Children in the backseats don't cause accidents. Accidents in the back seats cause children.