Migrating data from one database to another – to build data warehouses or new enterprise applications – is still a major undertaking for many companies.
While the market is awash with extraction, transformation and loading (ETL) tools to help automate this process, there are few tools that enable developers to examine the source data beforehand to see how suitable or 'clean' it is.
This lack of understanding can have a severe impact on the speed of implementation and ultimate cost of major IT projects, says Avellino co-founder and CEO Tony Rodriguez.
Avellino's flagship product is called Discovery. It automatically analyses source databases to find redundant or duplicate data; discovers data 'joins', 'keys', and 'data dependencies' in the database; and uncovers data parameters, such as minimum and maximum length.
The data is then transferred to Metabase, an in-house built associative database that is less rigid than a standard relational database. It can, therefore, better present the nuances of the source data – how it is structured and how it inter-relates.
This makes it much easier for developers to analyse it, eradicate inconsistencies and out-of-date information, such as old addresses, and prepare it for migration to the data warehouse or the new system.
That promise has proved an enticing sales pitch. The company has more than 25 referencable customers, including BT and Ford. It has also forged partnerships with systems integrators, including Accenture, Cap Gemini and EDS.
Now, it is expanding tentatively overseas. Tentatively, because despite forecast revenues of GBP5.5 million (€8.7m) in the year to the end of June 2002, it has still not raised one single round of funding. Instead of ceding control to a venture capital firm, Rodriguez says he prefers to grow the company organically for now, although VC investment is not ruled out in the future.
But it would be a surprise if they were not to beat a path to his door. The company has a strong, proprietary technology that solves a genuine business need in a market with few direct competitors.