RESEARCH & RESOURCES

Big Blue and Data Warehouses in the Cloud

IBM's Nancy Kopp-Hensley talks about cloud, data warehousing and what's ahead for its new BLU Acceleration for Cloud. She also explains why being a good sales person -- especially if you're an over-taxed IT administrator -- never hurts.

Nancy Kopp-Hensley, director of strategy and marketing for IBM's database systems, was always ahead of the pack. She holds a patent on technology used in IBM's Smart Analytics System and she was using geospatial data in conjunction with her data warehouse systems back in the 1990s. Kopp-Hensley sat down with BI This Week at the recent TDWI World Conference in Chicago to talk about cloud, data warehousing, and why being a good sales person -- especially if you're an over-taxed IT administrator -- never hurts.

She was particularly anxious to talk about IBM's new in-memory-data-warehouse-in-the-cloud, BLU Acceleration for Cloud, which is still in beta. BLU Acceleration for Cloud is based on IBM's DB2 BLU Acceleration database, which isn't a massively parallel processing (MPP) data warehouse (DW) service. An MPP service based on IBM's Netezza platform is coming, Kopp-Hensley assured us. "[BLU Acceleration is] in beta. Right now, it's based on our in-memory database, which is part of DB2 BLU Acceleration. [The service] includes Cognos BI, some data architect licenses, basically everything you need to provision a data warehouse and spin up [a] BI [practice]," said Kopp-Hensley. "You can have your own warehouse, which you can either manage yourself or you can get as a managed service. It's fulfilled on Amazon and SoftLayer, although the managed services one will be only on [BlueMix]." IBM owns SoftLayer and recently launched BlueMix as its platform-as-a-service offering.

She demurred, however, when pressed for a timetable for delivery of a Netezza MPP cloud service. Several vendors -- including Amazon.com (which introduced its Redshift MPP DW-as-a-service 18 months ago), Teradata Corp. (which announced its MPP DW-as-a-service late last year), and GoodData Inc. (which uses the Vertica MPP database from Hewlett-Packard Co. in its newly announced cloud service) -- now offer MPP-in-the-cloud. Scaling a cloud data warehouse is difficult enough; scaling an MPP warehouse is harder still. "Netezza isn't in the cloud today, but that's definitely a strategic direction for us as well. What we're trying to do is to balance how we take the best of our technologies and bring them forward into the cloud," Kopp-Hensley explained.

"We're looking ahead -- say, to three to five years from now -- where, even though we have all of these different technologies, in the cloud and on-premises today, people are going to look at consolidating these things," she continued. "We want to deliver [cloud] solutions that will complement what they're trying to do [in consolidating systems and workloads], between cloud and on-premises. I think a lot of customers will have hybrid [deployments] for quite some time."

All of which begs the question: what does BLU Acceleration have that other DW-in-the-cloud services don't? For one thing, it isn't an MPP service. This isn't necessarily a deal-breaker, of course. IBM's BLU Acceleration service also has the Cognos BI component, which helps differentiate it from other offerings. (IBM officials aren't ready to talk about enterprise pricing yet: the BLU Acceleration site has pricing only for "Trial" and "Solo" plans. Trial sign-ups -- limited to five hours -- are free; Solo, which limits you to a single user with 50 GB of data, is priced "from $0.67 per-hour, according to IBM. That translates into just under $500 a month.)

By contrast, start-up RedRock BI, which offers the Yellowfin BI platform running on top of Amazon's Redshift data warehouse service, starts at $2,500 a month for 2 TB of data. (This also gets you 120 hours of ETL processing powered by an ETL engine from SyncSort Inc. Additional RedRockBI capacity is available for $600 per month. Sources tell BI This Week that RedRock BI, a spin-off of iOLAP, will probably soon be folded back into its parent company.)

So why would customers choose IBM's cloud service over Redshift, Redrock BI, or other offerings? For the same reason, no doubt, that they'd choose IBM's BigInsight's Hadoop offering. "With IBM, I think there's more of an enterprise-class technology stack. Amazon, they're great at what they do, but data warehousing and BI is not their area of expertise," says Kopp-Hensley, who notes that Amazon acquired key intellectual property for Redshift from the former ParAccel Inc. (now Actian). "Customers want a trusted partner who's going to advise them, who's going to help them negotiate this transition [to the cloud], who's going to support them with enterprise-class service."

Kopp-Hensley made the trasition from technology management to technology marketing quite some time ago. As she sees it, she was nevertheless an effective marketer in her IT position.

"When I was on the user side, my predecessors had requested certain things for years. They'd request them, but they wouldn't make a case for them. I knew that if I wanted to get anything done, I had to sell the idea: I had to link it to some need, so I would get some technology, mock up a demo, and take it to the CIO, who would take me to the CFO, and then I'd usually get it," she explains, citing the example of her early use of geospatial data: essentially, this involved overlaying store franchise locations with crime statistics. This permitted her employer (McDonalds) to demonstrate that it was being proactive and was exercising due diligence to secure its stores. The win-win for Kopp-Hensley was additional funding that enabled her to do even more with geospatial.

"It's all in how you sell it. You have to go back to not just selling, but to putting in the effort, by showing value, to make it convincing," she concludes.

TDWI Membership

Get immediate access to training discounts, video library, BI Teams, Skills, Budget Report, and more

Individual, Student, & Team memberships available.