This article is part of an Essential Guide, our editor-selected collection of our best articles, videos and other content on this topic. Explore more in this guide:
2. - Data virtualization in the field: Read more in this section
- Benefits of data virtualization technology: Improved decision making
- Pfizer finds answer to data integration issues in virtualization tools
- Data virtualization software expedites chip development for Qualcomm
Explore other sections in this guide:
One of the largest microchip development companies in the world, Qualcomm Inc. designs semiconductors that serve as the brains behind many of today's smartphones, tablets and other wireless devices. And staying competitive in that world means being able to make decisions fast, according to Srinidhi Thirumala, a senior enterprise architect leader at the mobile technology giant Qualcomm.
Qualcomm has been using data virtualization software for the last few years as part of an effort to give company decision makers the most current information about where various chip development and related projects stand. The process required the company to access information housed in various, disparate systems and feed that data through a middleware layer and into useful dashboards that company executives can turn to at any time. The project has been valuable in terms of helping the company speed up chip development cycles and time to market, Thirumala said.
SearchDataManagement.com recently got on the phone with Thirumala to talk about Qualcomm's experience with data virtualization software. Thirumala explained why Qualcomm turned to data virtualization tools and embraced the notion of the logical data warehouse. He also discussed Qualcomm's reasons for choosing Composite Software Inc.'s data virtualization tools over competing products from Informatica Corp. and Denodo Technologies Inc. Here are some excerpts from that conversation, which have been edited for length and clarity:
What were some of the issues or problems Qualcomm faced before it began using data virtualization software?
Srinidhi Thirumala: The design cycle in this high technology space really comes down to speed, and as smartphone adoption and tablet adoption grow, we've got to put more and more features into our chips. Management and senior leaders and program managers executing large programs need an easier way to manage those different geographically dispersed projects. We wanted a quicker time to market so that customers would receive new products sooner.
So the company needed to speed up the design and development cycle to remain competitive. How did that realization lead to data virtualization?
Thirumala: Information was all over the place. We finally said that if company leaders could get to it all quickly -- like in a dashboard -- they could see how the big projects, such as the chip design projects, were progressing. The warehouse approach would have required a large team and many sources and moving the information into one physical data warehouse. It would have been a very labor-intensive development process, and we just couldn't afford it, and that's when we began to look at data virtualization.
You're using data virtualization to connect various sources and populate an informative dashboard for company decision makers. What information can been seen on that dashboard?
Thirumala: We have the one source, which gives us information on feature sets; another source tells us what teams are responsible for design; and there is also fabrication-related data, taxonomy and product data. All of these together provide a dashboard that leaders can look at and see how we are tracking to schedule, which is extremely important. That is what led us to go down the path of data virtualization.
Qualcomm has adopted the logical data warehouse approach, which is becoming increasingly popular. How do you define the phrase logical data warehouse and what is data virtualization's role in that strategy?
For more on data virtualization tools
Get more details on Qualcomm's decision to embrace data virtualization software
Read about pharmaceutical giant Pfizer's foray into the world of data virtualization
Find out how to tell if data virtualization tools are right for your organization.
Thirumala: You have your traditional data warehouse, and then soon you have data in the cloud and elsewhere. There is no point in bringing that data in from the cloud to operate and then mashing it up. And then you also have big data problems. There is no point in moving all of that Hadoop data. Let it reside where it is. The logical data warehouse is an architectural layer that uses data virtualization to abstract data from these various sources. It is disconnected from the actual physical data storage, if you will. In other words, it's a middleware, or architectural layer, that abstracts data independently from sources, rather than moving it all physically into one place.
What vendors did Qualcomm evaluate when it was gearing up to purchase data virtualization software, and why did it ultimately choose Composite?
Thirumala: We did look at Informatica's data virtualization software and another company called Denodo as well. But Composite was not trying to be a data quality company and their vision suited us in this space, and for the long term with our logical data warehouse strategy. Informatica does our ETL data integration and master data management and they are doing very good things as well. But the issue is that we felt Composite helped us quickly start this project and supported us once we reached a critical mass. Denodo was much smaller, and basically they are pretty good as well. We didn't find as many reference customers, and they have a slightly different architecture then Composite. They focus on some other areas in terms of query performance, etc.
What were some of the challenges that Qualcomm faced in implementing data virtualization, and how did the company deal with them?
Thirumala: There are challenges around optimizing performance and stuff, but don't get dissuaded. You can always throw more infrastructure at it. Many people give up too soon when adopting data virtualization because of problems on the periphery of things -- whether it's security or speed or you're getting good performance but not great performance, but you can resolve those.
What advice do you have for IT organizations that want to ramp up support for data virtualization software within their organization?
Thirumala: Focus on the right business reasons to use data virtualization at your organization. You have to do some proof-of-concepts and some piloting, and you have to influence people. Focus less on the technology and more on how quickly you can get stuff done than in the past. Aim for as real time as possible. That is what I would advise. The other thing is not to keep thinking along the old paradigm of just one physical data warehouse. There are mobile mobility trends, there are cloud trends and bring your own device trends, and people have very little patience these days.
Mark Brunelli is the news director for the Business Applications & Architecture Media Group at TechTarget Inc. Follow him on Twitter: @Brunola88