Why BI Appliances?

By David Hatch, Guest Author

As I wrote in last week’s blog entry, the term “appliance” is now being applied to BI. This term is used to define a technology offering that combines some form of software with some form of either hardware or other platform software (see Wikipedia URL: https://en.wikipedia.org/wiki/Computer_appliance for a really detailed and deep description of the term itself). The BI world has become more and more familiar with the term “data warehouse appliance”, and a cadre of vendors have entered that space (i.e. Netezza, DATAllegro, Dataupia and others) with solutions that combine a specialized hardware and software solution approach to solving some of data warehousing’s nastiest issues, the primary of them being “data availability”. Large data warehouse vendors have also addressed this issue through taking different technical approaches to their software solutions. Companies like Teradata and IBM battle in the DW trenches and aptly point-out that an appliance is not the only solution for dealing with large volumes of complex data.

Nonetheless, as data volumes have grown, and the complexities of queries and analysis needs have increased, the traditional approaches to accessing, processing and delivering resulting views into data have started to fall short. Simply put, the performance of legacy data warehouses and relational databases has not kept pace with the growth of data volumes and complexity. BI applications have been a driving force behind this. As companies have grown in their maturity with BI capabilities, their questions have become more complex, and the scope of data available to answer questions has become larger.

Now, the term “BI Appliance” has surfaced and through its name alone, there is a certain level of meaning attached to it. In essence, the concept is the same as the DW appliance approach – a combination of specialized software and hardware designed to take-on a specific application or need, in this case BI. The offerings available on the market today vary greatly. SAP and IBM have combined forces to offer a BI appliance solution for very large, very complex enterprise BI requirements. Meanwhile, BI giant Cognos has acquired Celequest and quickly produced the CognosNow! BI appliance that is positioned for SMB BI environments and is deployed either on-site or via a hosted/On-Demand approach. Same “BI Appliance” category… very different purposes. Ingres and Jaspersoft have taken an open source approach to “BI Aplliances” with an offering that consists of a software/software “appliance” that is designed as a virtual machine… it can be loaded onto several generic hardware platforms. This is a reasonable approach for an open source play as it does not assume that the end user’s IT organization will be happy with a specific piece of hardware, but leaves the selection of machine platform (or the pre-existence of a machine) up to the customer.

It will be interesting to see what other BI appliance solutions emerge. Will they be targeted at the large enterprise as a replacement to traditional DW approaches? Will they be the entry-point for SMBs to start dealing with large volumes of data. Will the exponential growth of data (mostly unstructured information such as web content, archived documents, call center notes/dialog, etc) drive customers to adopt appliances or will IT departments resist the siren song of the “one box fits all” approach?

More on this as I conduct more research into the matter at hand.