When you’re working in business intelligence, you’ll see the term OLAP being thrown about. But what exactly does it mean, and why do you need it? It’s usually a pricy add-on for a BI system… so what can it do for your organization.
So what is OLAP anyways? The definition can get a little confusing and technical. It is short for On-Line Analytical Processing. Great, but what does that mean? You’ll hear about theories about databases and relational vs. multidimensional. What it ultimately means is increasing speed and accessibility to the data. The relational version of this is called ROLAP. A multidimensional technology will typically be referred to as MOLAP. The hybrid scenario is called HOLAP. All three have there place in environments, depending on the reporting requirements.
From a technology standpoint, what this usually comes down to is storing data in memory. This is the “on-line” piece of the name. Generally, running a report or otherwise querying your data warehouse will come at a cost of processing time and resource usage on your data server. If you have a number of users, this can quickly tax your database especially if they are all running reports at the same time.
Using in-memory technology means that the larger queries can be run in advance against the data warehouse, usually overnight or in a less busy time. You’ll then have a “cube” or an in-memory set of data against which users can run their reports. Since the data has already been queried from the database, there is no hit against the warehouse for any report runs. Depending on the technology, users may also have access to on the fly manipulation of new report element groups, derived metrics created on the fly, and other report object manipulation.
Most of the major BI vendors have some sort of OLAP capability. Some like Microstrategy, have it as part of their software. Oracle has it as part of its database functionality. Business Objects’ application is based off of the old Crystal Reports technology. Cognos uses PowerCubes with PowerPlay or, more recently, their TM1 product. QlikTech’s major selling point of its Qlikview technology is the in-memory infrastructure. PowerOLAP has CPM capabilities built in. All of them have advantages and disadvantages.
Software is a good way to get this functionality in a reporting area of other business applications. It can be a form of embedded BI or a way to overlay this functionality on a database or data structure.