Madan Sheina
Microsoft zooms into data quality
Microsoft has expanded the data management capabilities of its SQL Server platform by acquiring Israeli data quality start-up Zoomix for an undisclosed amount.This acquisition is intended to strengthen Microsoft's SQL Server platform, with Zoomix's software being included in future releases of its database system. Zoomix's development team will be folded into the SQL Server team at Microsoft's R&D centre in Israel. The move is part of Microsoft's strategy to make SQL Server a complete enterprise platform for data management, addressing a critical layer of data improvement. SQL Server already has data integration (extraction, transformation and loading) capabilities and master data management (MDM) courtesy of its acquisition of Stratature last year.Data quality has close links to both these areas, so it makes sense for Microsoft to extend into it. The next challenge for the company is to figure out how Zoomix will tie into the rest of SQL Servers' data management tools? Will Microsoft be able to act quickly enough to make Zoomix a part of Microsoft's first MDM release, code-named 'Bulldog'? Bulldog is already slated to be part of the next version of SharePoint, with further integration points being eyed with Microsoft's Dynamics and CRM applications and Excel. Microsoft is well-advised to add Zoomix's technology to that list. Zoomix's Accelerator software is a bit different to other data quality tools in that it uses so-called 'self-learning' technology which combines advanced semantic and linguistic analysis with machine learning to automatically parse, match, classify and cleanse complex corporate product and financial data. The system is intelligent because it applies the knowledge gained from its previous cleaning actions to new data entering the system - i.e. the more it cleanses the cleverer it gets about it. The company refers to its approach as a 'firewall for data quality', arguing that its use of artificial intelligence minimises manual user intervention and ups consistency. Self-matching data quality methods like these are a welcome, and arguably overdue, addition to the world of data management.Significantly this is also Microsoft's first application of 'semantic' technology to its SQL Server stack. Zoomix's technology touches on 'the semantic web' (Web 3.0) and could also help Microsoft to better understand and leverage the potential of linguistic search to differentiate its online search and advertising initiatives against rivals like Google and Yahoo. Microsoft certainly needs to do something fast (figuratively and literally) to disrupt Google's monopoly in Internet search. Of course a semantic approach to speeding up Internet searches assumes that the semantic web will work as advertised. Nevertheless if Microsoft applies the knowledge and experience it gleans from Zoomix's semantic technology in the right places then Microsoft and its customers will benefit greatly, particularly as more and more business data starts to proliferate in 'dirty clouds' outside of the control of corporate IT departments.Microsoft is believed to have paid in the range of $25-35 million for Zoomix, which is not a big amount. The move underscores the importance of data quality, but it adds fuel to the debate as to which one of your IT vendors should provide it.Today there are many choices. Database, business intelligence, data integration and ERP vendors all now offer data quality tools integrated as part of their platforms in some way or other. Zoomix's technology, however, seems to be tied closely to the SQL Server platform. But dirty data potentially exists across all of an organisation's IT systems, including other database repositories like DB2, Oracle and Sybase, and not just SQL Server. Therefore tying data quality to a specific source system misses a key point of enterprise data management - for information to be managed effectively it needs to be kept agnostic from its consuming front-end transactional databases and applications. So the big question going forward is whether Microsoft will decide to keep its new data quality, or indeed MDM, tools database agnostic or not.
