This has been one of the most pivotal years in the evolution of the enterprise data warehousing (EDW) market. Every EDW vendor of note has firmly repositioned its go-to-market strategy around the appliance approach, with some also taking tentative steps into what is sure to be a key theme in 2009 and beyond: EDW in the “cloud.”
Yes, much of the recent “cloud” buzz has been bleeding-edge IT trade-paper fodder. It’s all interesting, to be sure, and there’s plenty of innovation going on. But much of the discussion seems to be a renaming, repackaging, and mashing up--with subtle twists and tweaks--of such well-established themes as service-oriented architecture (SOA), software as a service (SaaS), virtualization, and Web 2.0. And much of the trendy attention to cloud services obscures that fact that public cloud offerings from Amazon, Google, Microsoft, and others are still primarily a work in progress.
Forrester clients have only recently begun to inquire about cloud topics in earnest. But still, it’s hard to stay cynical about cloud computing for long. This solution delivery approach is coming almost certainly, inexorably, to all IT solution segments, including to EDW and business intelligence (BI). Sure, I&KM pros can safely ignore much of the cloud buzz for now, but, a year ago, they said the same about EDW appliances, and look how quickly appliances have become a dominant deployment approach in this market.
In the EDW-related inquiries I take from Forrester information and knowledge management (I&KM) customers, a great many concern the appliance market. CIOs, CTOs, DBAs, and other professionals are actively considering various vendors’ appliances to replace, or at least to supplement and extend, their traditional “roll-your-own” EDWs. Typically, the I&KM pro wants me to help them select the best EDW appliance for their needs from any of several vendors, both venerable blue-chip and sexy start-up.
What validated appliances for I&KM pros this year was the fact that big-name EDW vendors--including Teradata, Oracle, IBM, Microsoft, and Sybase--have gone this route in earnest. The inflection point for the whole industry was this fall when Teradata--who effectively established the EDW appliance space years ago but had long resisted going to market under that label--embraced the approach and significantly expanded their appliance solution portfolio.
EDW cloud services are still a few years away from a similar inflection point. The leading EDW vendors have taken only the most tentative steps into the still-embryonic cloud services market. But they are all beginning to explore the cloud/SaaS channel with greater interest. They simply have to. Customers’ capital budgets are under severe pressure, and a multi-million dollar EDW solution--be it a premises-based appliance or what have you--is a tough sell. In a soft economy, any on-demand pay-as-you-go offering becomes more attractive across all customer segments. Just as important, the increasing scalability, performance, flexibility, and availability demands on the EDW and BI infrastructure are spurring many users to consider managed, hosted, outsourced offerings with fresh interest.
We’re starting to see the next-generation cloud EDW emerge. One noteworthy development this year was Oracle’s partnership with Amazon. Under that agreement, Oracle customers can license Oracle’s core EDW software stack to run in Amazon Web Services' Elastic Compute Cloud (Amazon EC2) environment. They can also use their existing software licenses on Amazon EC2 with no additional license fees.
Another important development was Microsoft’s announcement of its Windows Azure cloud initiative, of which one key component is the (still in beta) SQL Server Data Services (SSDS) subscription offering. When Microsoft SSDS goes into production in 2009, it will offer some basic DW/BI features in addition to transactional database support. Though SSDS will not initially be at a functional par with they licensed SQL Server offerings, it is clear that Microsoft plans to evolve it rapidly toward becoming a feature-competitive DW cloud offering over the next several years.
Microsoft also made a key EDW-related acquisition this year, appliance pure-play DATAllegro. Forrester expects this acquisition to figure centrally into Microsoft’s evolution of SSDS into a massively scalable cloud DW service. Though DATAllegro did not achieve much market adoption as a DW appliance pure-play, it provides Microsoft with a robust scale-out technology called “shared-nothing massively parallel processing” (MPP). By the way, Microsoft is playing catch-up in this regard, since most of its closest competitors implement shared-nothing MPP in their EDW premises-oriented solutions, and such DW appliance pure plays as Netezza, Greenplum, and Vertica also implement it to varying degrees.
When Microsoft ultimately ships a DATAllegro-powered SQL Server EDW appliance under its “Project Madison” in a year or two, we would not be surprised to see it adopted first in SSDS. The cloud EDW offering would benefit immensely from shared-nothing MPP’s ability to manage petabytes of analytic data and parallelize queries and other transactions seamlessly across a grid of hundreds or thousands of nodes.
Indeed, the industry consensus is largely in favor of shared-nothing MPP across the storage and compute tiers, coupled with flexible information-as a-service (IaaS) and server virtualization, as the principal platform for cloud computing. In the next-generation EDW, shared-nothing MPP allows the infrastructure to become more fluid, flexible, and virtualized, while managing ever more massive data sets and providing the agility to handle more complex mixed workloads of reporting, query, OLAP, data mining, data cleansing, transformations, and other functions.
As 2009 approaches, we’ll see pure-play DW cloud vendors come to the fore, appealing both to the early adopters among I&KM pros as well as to those under severe budgetary, headcount, and data center constraints. The established EDW vendors will not come to the cloud in a big way till 2010 at the earliest, it appears. But they will come, and with all guns blazing.
In 2-3 years’ time, the established vendors will own the EDW cloud space just as they’re starting to own the rapidly maturing appliance segment--in part, by acquiring the most promising cloud startups. The longer the economy stays drab and dreary, the faster the cloud EDW segment will expand, mature, and consolidate.
As that happens, commercial EDW cloud offerings will become as diverse and feature-rich as appliances have become, and I&KM pros will almost certainly ramp up their cloud EDW inquiries.