Posted by James Kobielus on March 12, 2010
It was my pleasure to participate in the latest DM Radio podcast panel yesterday. Eric Kavanaugh and Jim Ericson always do a fine job of organizing these events, and, with their stellar industry panels and fun “morning drive-time crew” on-air patter, they keep it lively. And these guys actually know a thing or two about information management.
The latest DM Radio panel was right in my core coverage area. They called it their “Third Annual Appliance Showdown.” That got to me to thinking: early 2008 (when they held their first) was also when Forrester began our coverage of data warehousing (DW) appliances, starting with publication of my report “Appliance Power: Crunching Data Warehousing Workloads Faster And Cheaper Than Ever.” When I published that report, DW appliances were still not quite in the enterprise mainstream, because they were still regarded by enterprise IT as, in the words of Kavanaugh, an “adjunct” to the enterprise DW (EDW) for fast table scans and query processing, rather than as platform that could scale to support all EDW functions.
So much has changed in the market in those 2 years. As I said on the panel yesterday, more and more commercially available DW appliances have the scalability, flexibility, workload management, and other features necessary to support all EDW functions. In fact, the appliance-based platform is now the core go-to-market approach for most long-established EDW solution vendors—including Oracle, IBM, Teradata, Microsoft, and Sybase. Just as key, Forrester now regards Netezza, Greenplum, Vertica, Kognitio, ParAccel, and other DW appliance vendors as offering solutions that are worthy of consideration for EDW deployment. Some are further along the “EDW-readiness” curve than others, but it’s clear that appliances are evolving rapidly to contend with the established leaders in the EDW space.
Later this year, I will update the Forrester Wave for EDW Platforms, and it’s a safe bet that all of the platforms to be evaluated will be exclusively or predominantly appliance-based. I published the first Forrester EDW Platforms Wave a year ago, and then saw it rapidly go out of date due to important appliance product releases from most of the vendors mentioned above. One thing that’s clear now, but wasn’t a year ago, is that the EDW platforms market is shifting toward in-database analytics, prepackaged domain models, and service-oriented architecture (SOA)-based development platforms. In other words, the appliance-based EDW platform is becoming an analytic application server that incorporates SOA interfaces to deliver analytics into the full range of business applications.
With that as context, I took issue on the panel with Dataupia’s Foster Hinshaw, who once again offered the simplistic analogy of the DW appliance as a “toaster.” The problem with that analogy is that it limits DW appliances to the role of adjunct: a simple black box that is designed and optimized for a single function. I pointed out that enterprises’ needs are much more complex, and that people live on a rich, diverse analytics diet that should not be limited to plain white bread singed on either side.
As I noted on the panel, the defining feature of DW appliances is that they are turnkey, pre-built, pre-optimized, modular configurations of hardware and software components that enable quick time-to-value for some range of functions, be it narrow or broad.
But, as I also noted on the call, my true belief is that appliances don’t matter, in the final analysis. What truly matters is analytics-driven business. Down deep, what organizations crave are analytics platforms that are bigger, faster, and cheaper—and which allow them to realize value sooner and more pervasively. Whether that vision is enabled by appliances (however scoped), cloud/SaaS, or some other approach is a secondary issue. However, I do expect cloud/SaaS-based EDWs to mature and gain adoption in coming years for just this reason.
Where last year was the year that appliances entered the DW mainstream, I expect that next year is when we start seeing widespread enterprise uptake of cloud/SaaS-based EDWs. Clouds will take the “bigger, faster, cheaper, sooner” value proposition to the next level, where analytic databases are concerned.
Oh...I'd be remiss if I didn't mention that I'm working on a Forrester report on virtualizing EDWs into the cloud. And, yes, I’m already bracing myself for the never-ending debates on the definition of clouds.