There has been a drive by increasingly technically savvy business teams to perform self-service analysis on the business information that they intimately know in terms of element relationships and business rules. At the same time these business teams do not want to involve their IT counterparts in all steps of the analysis. The primary reason for this drive is simply the overhead required in terms of time and resources to involve IT in generating critical information that drive business decisions. If business teams could get their hands on tools that allow them pull information in from the data repositories and allow them to do the analysis they need fast and in an easy, non-technically daunting manner, the value of such tools is easily bought by business sponsors. Of course, in no way does this imply that IT is redundant, IT has to be involved in the enterprise reporting level and to maintain the infrastructure that supports the so-called local departmental analytics, and they will always be brought into picture when a certain local analysis is so critical that it needed to be “productionalized” and deployed to a larger audience.
Unfortunately, thus far such “magic” self-service tools that business teams could master and use locally did not exist, or if they did they were either in spreadsheet formats with restricted analysis capabilities, or embedded inside of bigger complex software suites and therefore were cost prohibitive. The fast response times from such local applications where the business user wanted results rather than waiting for response prompts from the reporting applications were absent as well.
The advent of in-memory analytics has sought to address the desire by business users to do their own analysis to a large extent. Such niche tools have been around for a long time. However, the changing dynamics of the industry (consolidations), the decreased cost of processing power and memory, as well as the technical advances in operating systems has suddenly found these niche tools on the “right side of the wall”. These tools are classified in the in-memory analytics niche of business intelligence tools. Thus far they serve to compliment rather than compete against existing business intelligence suites. Several niche players have emerged in this space, the leaders being Qliktech (with Qlikview), and Tibco (Spotfire). The technologies basically load data from sources into the memory of the machine they are running on, rather than landing or reading data (or parts of the data) to the disk as is done in other business intelligence tools. Inherent efficiencies of a 64-bit OS platform, as well as the current typical RAM memory configurations in multiples of GBs allow complex and large data sets to be simultaneously loaded to the memory, providing near-instantaneous response times to user queries.
Bigger players in the business intelligence space have started offering comparable technologies as part of their bigger suites. Of note, Micorosft announced a release of their in-memory engine (thus far known as Project Gemini) in latest release of their flagship SQL Server platform. Oracle has also announced in-memory capabilities in the latest build of their 11g database. Given the increased value especially perceived by end user groups, it will not be surprising if in-memory features become an integral part of all business intelligence suites in the near future.
