With Master Data Services, Microsoft may be transforming how SQL Server administrators tackle Master Data Management...
projects, but that doesn’t mean they know how it’s being done.
We spoke with Mark Kromer, a SQL Server expert at Microsoft, about how Master Data Services aims to streamline Master Data Management, which companies are positioned to benefit, and what we can look forward to in subsequent releases of the software.
1. Master Data Services (MDS) is one of the biggest new additions to SQL Server with the release of 2008 R2. Can you briefly explain the hole MDS fills that wasn’t possible with previous versions?
I agree that Master Data Services is a very important addition to the SQL Server data platform. In my discussions with customers about SQL Server as a platform, it has been interesting to see both how many customers are not familiar with MDS, but also how many customers are beginning to plan and execute on Master Data Management (MDM) projects. Up until Microsoft purchased Stratature a few years back, there was not a focused effort at Microsoft around a solid MDM story. That all changed with the acquisition, and it took some time to get the product teams aligned and the strategy in place. But I think they got this right with MDM residing in the SQL Server data platform, as opposed to in Office or SharePoint, where many people thought it would land.
2. Some have said that smaller organizations are unlikely to see as much use for MDS as large enterprises. Would you agree?
That is a bit of a generalization but probably true for the most part. No matter what the size of your organization, you will likely experience the same problems that MDM is there to resolve. That is, data that is critical to your business success that is inconsistent, duplicated and dirty, i.e., lacking in quality control. But MDM projects can be complicated and require a mindset change in many organizations. This surfaces particularly in practice where a data governance or data steward group is identified as the custodian of master data. Not all companies are configured such that implementing these practices and MDM tools is practical. Outside of that generalization, though, is the fact that MDS is a much lighter-weight tool for MDM as opposed to some of the other MDM providers, and since MDS is included as part of SQL Server 2008 R2, MDS has a much lower cost of ownership.
3. Master Data Services is derived from Microsoft’s acquisition of Stratature back in 2007. Why was the Stratature technology a good fit for SQL Server?
Stratature fit perfectly into the classic Microsoft acquisition legacy, particularly in the BI product line. While IBM and SAP were adding MDM hubs into their applications and suites, Microsoft jumped into the MDM market by acquiring a highly respected pure-play MDM vendor which added instant credibility for Microsoft in the MDM technology world.
4. DBAs can use SSIS to load MDS data. Can you explain how the two work together?
SSIS is a key tool to use to load your company’s data into the data models that you build in the MDS web tool. When you build your models with hierarchies, workflow, attributes and security in MDS, SQL Server tables and views are automatically generated. A good practice is to use SSIS to load the data into those objects. I would also say that many customers will likely still also utilize SSIS and its capabilities around data quality with the data profiler task and fuzzy grouping as a way to add another layer of data quality control to the overall MDM process that they practice in their business. With the current version of MDS that ships with SQL Server 2008 R2, I am finding more and more customers put into practice a data quality and standardization project around MDM that includes both MDS as well as SSIS for not just importing, but also for fuzzy grouping and data profiling, which is a new SSIS capability since SQL Server 2008.
5. One of the benefits of using MDS for master data management is obviously cost, but that usually comes at the expense of certain functionality when compared with more mature, third-party products. What are some areas you expect to be improved upon in the next version of MDS?
Some of the areas that I think you can look forward to in MDS would be around built-in data quality capabilities, MDM hubs and improvements in match and merge and fuzzy matching.