While data virtualization tools are a relative newcomer to the data management marketplace, several companies already boast very mature,
enterprisewide deployments, according to Robert Eve, the vice president of marketing at San Mateo, Calif.-based data virtualization software vendor Composite Software Inc.
In the new book Data Virtualization: Going Beyond Traditional Data Integration to Achieve Business Agility, co-authors Eve and Judith R. Davis, an independent IT industry analyst and researcher, explore the experiences of 10 companies -- including pharmaceutical giant Pfizer and wireless technology stalwart Qualcomm Inc. -- that see data virtualization software as an integral element of their overall IT strategies.
An alternative to traditional data warehousing initiatives, data virtualization tools give users the ability to access and report on information housed in disparate data sources throughout an organization's business units.
SearchDataManagment.com recently got on the phone with Eve to learn more about the challenges and benefits experienced by Composite customers and featured in the new book. Here are some excerpts from that conversation:
What would you say is the overriding message of your new book?
Robert Eve: I think the challenges facing so many businesses today -- things like extreme competition, new product introductions [and] risk management -- are real difficult issues to navigate. Businesses agility is the real success factor that's going to make the difference between companies that survive and thrive and companies that don't. And data virtualization is now a proven approach for providing that sort of business agility. As we've kind of moved from this new,
early-adopter image of this technology [to a more] proven approach, people want to see case studies. So, I think the message is that you can gain business agility through the successful adoption of data virtualization, and here are 10 enterprises that have done it.
Do you think data virtualization tools now have a proven track record?
Eve: I would say that at this point. All of these companies [in the book] have broad enterprise adoption, multiple use cases, and they view it not just as some point project capability but as [an actual part of their] enterprise architecture. They are certainly the most advanced companies. But there are other companies who haven't even gotten started. The difference between the [early-adopter phase] and kind of crossing the chasm into a more mainstream market [is evident] when there is broader adoption for a wide variety of use cases and when there are established best practices and domain knowledge that kind of expands around the core technology. I think we've achieved that with data virtualization.
Can data virtualization software be used to access and analyze unstructured data?
Eve: It can be used for unstructured data if we kind of get to that middle ground of semi-structured data. [If you're talking about] just documents and that sort of thing, we can certainly query that and deliver it through. But to really add meaning and align that and join it with other data to create a 360-degree view of customer, [a little bit] of structure is required in order to make that linkage with those other entities. Often people will use something like XML as that middle ground.
What is most common cultural challenge that companies in the book faced when launching a data virtualization project?
Eve: The primary thing that everyone experiences is just resistance to change. The traditional data consolidation paradigm with the data warehouse has really been very successful in solving a number of problems. But [this paradigm] doesn't extend to all these new types of data sources very easily. There is a need for business agility and that [classic data warehousing] approach takes longer to implement. So [a challenge is getting people to] consider a different approach or consider a hybrid mix. Getting on board with that is probably the big challenge that everyone has to go through. It's just a classic change management issue. People tend to stick with what they know.
What is the common technical hurdle associated with data virtualization software and how do users overcome it?
Eve: The concept on which data virtualization is built is around a really simple idea, which is [gaining] a view of the data, and people are pretty formally used to that view. [But] as you get to larger and larger implementations, you really have to think about what the business views need to look like. You can start pretty easily with simple ideas around the views, but as you get to a larger enterprise deployment, you need to think more about your models and consider other things like master data management, et cetera. As you broaden adoption it can get more complicated.
What can we expect in the future from Composite Software in terms of features and functionality?
Eve: With the evolution of fit-for-purpose data stories such as appliances, Hadoop and key value stores, we're constantly challenged to add sources and make them available so that we can consume data from those sources and provide that to the business through a business view of data. I think that's an ongoing challenge that we'll always face.