So, I spent a couple of days in San Francisco last week for the Next Generation Data Center conference, and it should come as no surprise that cloud computing was among the most popular topics, but hardly stole the show (from my perspective, at least).
On top of the various sessions and keynotes that focused on cloud computing — including a very informative look at Merrill Lynch’s “stateless computing” infrastructure (see my article on the event for more details) — a clear indicator of how far cloud computing has come can be seen in the event’s awards. GoGrid, which last year took home the LinuxWorld “Best Clustering or Grid Solution” award, this year not only was a finalist in the “Best Virtualization Solution” category (won by Pivot3 — read more about them in Dennis Barker’s feature story on the company) but actually won the “Best in Show” award. When I caught up with GoGrid CEO John Keagy right before I took off to the airport, he was understandably excited about what this award means for the company, as well its significance in validating the cloud paradigm. Where GoGrid can go from here I don’t know, but winning two awards in its first two years of existence is nothing to scoff at. Something tells me it will have a little more cloud-based competition at next year’s show, however, and a cloud computing category can’t be too far off.
Of course, this year’s “Best Clustering or Grid Solution” winner, DataSynapse, also made big news at the show. On top of announcing VMware compatibility and a partnership with Adaptivity, the company also released the results of its survey of 100 IT professionals, most of whom came from the financial services space. As I watched the presentation of the results at the Tuesday-evening cocktail reception, I couldn’t help but notice that some of the top answers seemed to be at odds with one another. For example, while 25 percent of respondents picked “aligning IT with business objectives” as the biggest issue with datacenter operations, 42 percent picked cost reduction as the primary motivation in choosing a datacenter technology. Assuming that “business objectives” meant real-time dynamism, I began to wonder how serious the respondents actually were about achieving this parallelism — after all, there is no shortage of solutions bringing on-demand capabilities to every imaginable aspect of the datacenter. If this alignment is so important, perhaps it should trump cost savings as the No. 1 motivating factor in purchasing.
However, when I spoke with DataSynapse Chief Marketing Officer Joe Schwartz, he reminded me that, especially with the hard times the banking industry is facing, they really do need to do more with less. Thus, DataSynapse and other grid-based technologies actually are doing quite well in this market, as the increased utilization and service-sharing capabilities can significantly reduce CAPEX while delivering substantial gains in real-time performance. Throw in the policy-based allocation tools of a product like FabricServer, and you have your IT-business objectives alignment and your cost savings. Policy-based provisioning actually was among the wishes of the survey respondents, as were related tools like expanded metrics and real-time reporting.
It probably is worth noting, as well, that high-performance computing ranked above both shared services and cloud computing as a driving IT initiative. As Merrill Lynch’s Jeffrey Birnbaum made clear in his keynote, banks definitely are looking at these capabilities, but my impression is that the high-performance trading applications aren’t yet a part of this transformation (Merrill Lynch also is a DataSynapse customer, after all). In fact, Schwartz told me that only 5-10 percent of DataSynapse’s customers have federated their grid environments — a trend that I’ve been told is picking up, though, as the benefits of a shared, service-oriented grid infrastructure become too glaring to ignore. We’ll have more on the survey results in the weeks to come.
I also got a chance to speak with Xkoto Co-Founder and Chief Strategy Officer Albert Lee, who shared with me the latest news about the Waltham, Mass.-based database virtualization company. Among its struggles have been figuring out how to market its Gridscale solutions, which creates a pool of active-active DB2 instances across machines, increasing data performance and creating continuous availability. Lee says the company switched its terminology from “data virtualization” to “database virtualization” in order to minimize confusion about what it does, and also said that although Gridscale is a data grid, Xkoto doesn’t like to advertise this fact, as customers tend to get scared of the g-word. But that doesn’t mean it hasn’t been a good year for Xkoto, which has doubled its customer count since we covered the company back in January, and which plans to support SQL within the next 30 days — a move that should result in a customer spike of its own.
Despite Xkoto’s success, though, Lee acknowledges that neither virtualization nor database virtualization has reached its full potential yet. “When data warehouses are on hypervisors,” he says, “it’s for real, it’s mainstream.” We should have more on Xkoto on the weeks to come, as well.
Finally, as you’ve no doubt noticed if you have been to the home page, this is the last weekly issue of GRIDtoday, and the last issue period under the GRIDtoday moniker. It’s been fun while it lasted, but we’re looking forward to doing business as On-Demand Enterprise starting on Aug. 18. As for this commentary, expect to see more frequent, condensed entries in my blog, “The Essence of On-Demand.”
Elsewhere in the issue, be sure to check out the following items: “AT&T Launches Global Utility Computing Service“; “IBM Unites Enterprise Tools, Mobile Devices in the Cloud“; “GridGain Scales to 512 Nodes on Amazon EC2“; “VMware Enhances Lab Manager, Enables Multi-Team Sharing“; “3Leaf Enhances Server Farm Management w/ I/O Virtualization“; and “Elastra Closes $12M in Series B Financing.”