All aboard the data integration expressor

Written By:
Published:
Content Copyright © 2008 Bloor. All Rights Reserved.

A new entrant to the data integration space is expressor software. Although there is no shortage of integration technology out there, in practice large scale integration projects involve a large amount of manual scripts which deal with what transformation to apply to a certain column in a source system before sending it on to a target system. This approach often results in business rules being embedded within this code, and hence not re-used in other projects. Moreover current technologies map physical columns to others, and so if (as inevitably happens) changes occur to these structures, then these scripts need to be updated. Consequently, high maintenance costs are a frequent issue for data integration applications. Another issue for many enterprises is the limited batch window, which means that performance of such integration jobs can be a bottleneck, enabling Ab Initio, in particular, to carve out a high-performance niche in the market against players such as Informatica and IBM (who acquired Ascential).

The expressor technology tackles both these real problems head-on. It uses a higher level semantic abstraction model to derive business rules and store these in a metadata repository, maintaining a layer between business terms such as “Full customer name and address” and its physical representation(s). In this way it is easier to re-use business rules across projects, and the abstraction layer means that physical structure changes in the underlying data sources are quicker to adapt to. It has built a series of connectors to common data sources, such as databases, XML and even Cobol copybooks.

The speed issue is dealt with by designing the expressor engine with MPP processing (which is also the approach used by Ab Initio), enabling customers to throw standard hardware at the problem. The company’s principals have long experience in embedded software, and have written a platform neutral engine that they claim shows very high performance in tests relative to other vendors. These claims will need to be validated by independent studies and real-life customer experience; claiming your product is ten times faster than the leading product on the market is something that will require careful testing, though it is intriguing.

The company has been formed by a couple of industry veterans who have been programming in this environment for years and have in-depth experience of the main vendor technologies in the market (and hence know their limitations). The company has received USD 7.5 million dollar funding primarily from two reputable VCs, and has brought in experienced management. Its product hits general availability in June 2008, though it has been conducting trials at a number of customers in spring 2008.

The problems expressor addresses are very real. Although it will be attacking large, entrenched competitors, the success of Netezza in the data warehouse market has shown that such heights can be scaled. The company will need to back up its bold claims on performance with independent (preferably customer) case studies, but even if its performance claims are partly true, then the big boys may have something to worry about soon.