Update – illumuniate

Since I last got updated on illuminate (see the first look on i-lluminate) they have been focusing on their business model. While illuminate is a general purpose database they have been focusing on data warehousing as an initial market– with business analytics applications to follow. The data warehouse market has various segments based on size. In large Enterprise Data Warehouse (EDW) installs illuminate is not focused on rip and replace but on adding illuminate to improve support for ad-hoc queries. In smaller EDWs illuminate is trying to replace existing installs and they are also targeting those in the small/mid-size market who have not yet adopted an EDW. Illuminate focuses on being the lowest TCO and on reducing what they call “time to analytics” and “time to answer”. They are essentially trying to collapse the time it takes to get from data to decision.

Having decided to go to an indirect model they have been building out alliances for sales and professional services. illuminate is using IT consulting firms to provide an outsourced direct sales force. These folks have experience in DW/BI sales and have to be certified by illuminate for both sales and support/installs. The end result looks like a direct sales force even though they don’t actually work for illuminate. The professional services organizations of these partners vary and there are incentives for them to work together on implementation, though there is no sales competition, and illuminate is expecting partners to give up a degree of autonomy so that the leads and accounts can be managed more coherently across multiple partners.

From a technology perspective they have made some changes too. Illuminate 4.0 supports 64 bit which is critical for them given their use of global ids (the size of the ID representing an upper limit on the size of database they can support). Similarly they are typically processor bound and so multi-threading is important and has been a focus. They have added some support for web services to make SOA integration easier and have provided Java, .Net (and C++) public APIs. LDAP and Active Directory support and some simplified management tools as well as improved ODBC support make integration and development easier.

These changes allow them to support cloud and managed service offerings in principle and have allowed them to improve user scalability and hardware utilization. As processors keep improving illuminate can run in a shared-everything environment because it is processor speed that is typically critical for them. They claim that competitors are often limited by the I/O channel speed which is not improving anything like as fast as processor speed. They continue to focus on query performance and the numbers they show are very impressive.

Clearly the long term value of illuminate is going to be in making analytics work in new areas or in time frames that allow new kinds of solutions. This will require them, however, to develop (or cause to be developed) new algorithms that take advantage of their unique database structure. They are beginning to bring analytic algorithms into the database engine while some things, like product affinity analysis, that used to be data mining queries are now simple queries thanks to this same structure.