Q&A: Moving Beyond Traditional Data Virtualization
Composite Software VP Robert Eve has co-written a new book on data virtualization. In this interview, he describes going beyond traditional data virtualization and discusses some of the lessons from the book's 10 high-profile case studies.
- By Linda L. Briggs
- January 17, 2012
"If you want to meet your business agility challenges, you need to go beyond traditional data integration and leverage the power of data virtualization." That's the message in Composite Software VP Robert Eve 's new book, which he co-authored with Judith R. Davis, called Data Virtualization: Going Beyond Traditional Data Integration to Achieve Business Agility. The heart of the book: case studies of 10 companies, some of them very large, that have successfully adopted data virtualization. In this interview, Eve discusses best practices that emerged from the case studies, and explains why data virtualization enables business agility.
BI This Week: What triggered a book on data virtualization?
Robert Eve: As data virtualization has emerged from a solution for early adopters to a proven data integration approach and technology for mainstream enterprises, the demand for success stories and best practices has exploded. The book fills this information void with actual cases studies from ten data virtualization adopters across a range of industries and functional domains.
Why do you see it as important to go beyond traditional data integration?
In the book's foreword, Colin White did a great job answering that when he said, "Growing demand to access data not managed by the data warehouse, coupled with the need to access current data for more agile decision making, means organizations have to change the way data is accessed and used by business users. The solution is to give users a single interface to the data and data services they need, regardless of where the data is stored or how it is organized. This doesn't mean organizations have to abandon the data warehouse, but rather they need to extend it. Data virtualization is a set of techniques and technologies that provide the key underpinnings that enable organizations to do this."
One of the book's central points is that data virtualization enables business agility. How does virtualizing data lead to agility?
It goes back to Charles Darwin: To survive and be competitive, an organization must be adaptable. Business agility -- the ability to quickly take advantage of new or changing business opportunities -- has become the most important factor at enterprises today. While the importance of business agility is well understood, achieving it is a difficult and ongoing challenge, with information the key. In the quest to become an agile business, business and IT leaders must address all three elements of business agility.
- Business decision agility: Data virtualization delivers the complete high-quality, actionable information required for agile business decision making
- Time-to-solution agility: Data virtualization uses a streamlined approach, an iterative development process, and ease of change to significantly accelerate IT time to solution
- Resource agility: Data virtualization directly enables greater resource agility through superior developer productivity, lower infrastructure costs, and better optimization of data integration solutions
The core of the book is a number of case studies of some very large companies that have successfully adopted data virtualization. One example is a group within Pfizer, the huge multinational pharmaceutical corporation. Can you talk about that case study?
Pfizer's PharmSci group is responsible for bringing drugs to market. The group has a complex portfolio of projects that is constantly changing. The biggest challenge faced by the team was the need for an easy and simple method for obtaining complete, high-quality, actionable information so that over 100 information producers at Pfizer, along with 1,000 of their information consumers, can make strategic and tactical business decisions.
Pfizer implemented data virtualization to integrate all its data sources into a single, virtual information schema that can be accessed by all front-end BI tools and users. The agility and flexibility provided by this virtual data layer helps the PharmSci group bring new drugs to market faster and save money.
Another company profiled is the Chicago-based international financial services company Northern Trust. What was unique about what Northern Trust was trying to do with data virtualization?
Northern Trust delivers investment management, asset and fund administration, fiduciary, and banking solutions to corporations, institutions, and affluent individuals. Providing outsourced investment management operations for other institutions is an important line of business for Northern Trust. As the number of institutional customers in the outsourcing pipeline grew, Northern Trust needed to reengineer its process to get these new customers on board faster. By implementing data virtualization, Northern Trust improved time to market by 50 percent and gained a 200 percent return on their data virtualization investment.
In a recent TDWI Best Practices report, Self-Service Business Intelligence, Claudia Imhoff and Colin White promote the use of data virtualization to provide business views of the data. Did any of your case studies deal with that issue?
Providing "business views" of the data is a challenge every organization faces, and the ten we profiled are no exception. The good news is that with data virtualization, all of them are making significant progress. Let me highlight a few, starting with NYSE Euronext.
NYSE Euronext includes 14 different exchanges and trading operations. This organization takes in as many as four billion transactions daily, which translates into two to five terabytes of new trading and reference data every day. The company is using data virtualization to provide a business view of key business entities: trades, orders, receipts, quotes, cancels, and admins across its 14 exchanges and trading operations.
Here's another example of providing business views: The Global 50 Energy Company has defined and bound nearly 600 business canonicals across its upstream line of business. Even the relatively small, not-for-profit Compassion International has made impressive headway in providing business views of its global charity operations information to internal managers and even external donors and sponsors.
Were there any myths about data virtualization that you tried to debunk in the book?
Here's one: Performance and scalability have often been a perceived barrier to adoption of data virtualization. Instead, many of the companies profiled in the book cited the performance, reliability and scalability of their data virtualization solutions as key to their success.
As one example, Comcast improved service-level agreement (SLA) performance from as long as 10 seconds per transaction to 1.2 seconds on average. In addition, on the reliability front, Comcast has run for several years with zero down time. NYSE Euronext uses data virtualization to provide information from massive data volumes (two to five terabytes a day) across all of its exchanges and markets. At the Global 50 Energy Company, if performance SLAs extend beyond core data virtualization capabilities, the company leverages MPP appliances along with data virtualization caching to further accelerate queries.
How long did it take most of the firms profiled to implement data virtualization? Was there any pattern?
Nearly all the organizations started with specific point projects that they were able to get working in just a few months, from data virtualization installation to successful project delivery. They then expanded the number of use cases, data sources, skilled developers, and more as they broadened adoption.
Three of the best practices identified in the book support this implementation strategy, namely: centralize responsibility for implementing data virtualization, educate the business on the benefits of data virtualization, and take a phased approach to implementing data virtualization.
What kind or ROI have the profiled organizations achieved?
Every company interviewed achieved significant, measurable financial returns. The Global 50 Energy Company achieved 40 percent development cost reductions; Compassion International achieved 30 percent. Comcast is saving $730,000 a year in customer service expenses. Northern Trust achieved a 50 percent reduction in time to market and a 200 percent return on investment. NYSE Euronext saved $4.5 million for one project alone. Qualcomm saved $2.2 million across its first five projects.
What were a few of the best practices identified in the book?
All of these user organizations have been using data virtualization for at least two years, some for as long as four to six years. Based on this experience, they identified five best practices.
- Centralize responsibility for implementing data virtualization
- Educate the business on the benefits of data virtualization
- Pay attention to performance tuning and scalability
- Take a phased approach to implementing data virtualization
- Use an experienced vendor partner for data virtualization technology
Across the ten data virtualization adopters you profiled, what is their overall message to traditional data integration users?
The key message from the ten organizations highlighted in this book, and the hundreds of others that have used data virtualization to achieve business agility success, is simple and clear. If you want to meet your business agility challenges, you need to go beyond traditional data integration and leverage the power of data virtualization. If we can do it, you can, too.