Although we’re rapidly approaching the new year (which I believe will be 2008), a recent analyst report has me feeling like we’re back in 2004. Just when you thought grid computing was dead and gone, here it comes again – this time as one of Gartner’s 14 IT delivery models that will “completely transform the IT market in the next five years.”
A few years ago, grid computing was the technology on everyone’s mind, and it had the traditional siloed architecture right in its crosshairs. Over the past couple of years, however, the term itself has fallen out of favor and lost some of its shine, as the technology in some cases failed to live up to its tremendous hype and didn’t deliver the out-of-the-box panacea hoped for by so many. In fact, some vendors made efforts to distance themselves from the word “grid,” and predictions had virtualization stealing grid’s much-ballyhooed thunder in the area of server consolidation. (That’s not to say virtualization has not been a rousing success thus far.) So how, then, has grid computing reemerged as one of the most revolutionary IT models on the market?
I can’t speak on Gartner’s take on the matter, as I haven’t yet had a chance to read the report, but in my opinion this has a lot to do with the continued innovation on the part of grid software vendors. Moving beyond the original claims of harnessing CPU power and maximizing resource utilization, today’s grids are better at handling data and working with real-time transactional applications, and new-and-improved licensing models have lowered the upfront costs of jumping on the bandwagon. In addition, a number of smaller vendors — such as Digipede, GigaSpaces, Appistry, etc. — have really started to gain some traction with their generally low-cost, easy-to-use, highly scalable and highly holistic distributed platforms. The distributed caching capabilities employed by GigaSpaces’ and Appistry’s solutions, as well as by recent Oracle acquisition Tangosol and others, also have had a positive impact on the image of grid computing.
With these innovations have come new marketing angles, and grid vendors have stepped up to the plate here, as well. No longer is grid computing being sold almost exclusively as the solution for running compute-intensive applications in traditional HPC-heavy fields like manufacturing, aerospace, pharmaceuticals and automotive engineering. Instead, grids are being offered up as the ideal platforms for all varieties of data management projects, and improvements in latency, availability and scalability have made grid technologies easier sells to anyone doing e-commerce, electronic trading, logistics or pretty much anyone else looking to optimize their enterprise datacenter. Word-of-mouth marketing probably has been a big factor, too, especially when one looks at how heavily major IT users like financial services firms and Internet giants (e.g., Google, Amazon, eBay, etc.) rely on grid-based architectures. It also hasn’t hurt that more and more ISV applications, in fields ranging from e-mail archiving to sales performance management, are becoming grid-enabled, leading to penetration in markets and lines of business outside the technology’s traditional realms.
Grid computing also has been helped along recently by the increasing popularity of complementary technologies and delivery models, many of which are based on grid technologies, so it shouldn’t come as a surprise to see utility computing, capacity on-demand, software as a service (SaaS), and storage as a service on Gartner’s “revolutionary” list. Capacity on-demand and utility computing, in particular, rely heavily on traditional grid computing’s distributed architecture while infusing their own characteristics to deliver enterprise value. By some definitions, in fact, capacity on-demand, utility computing and grid computing can be very difficult to distinguish — especially with certain vendors labeling their in-house platforms as “utility.” Other models, such as SaaS and, to a degree, Web platforms, not only offer up the very griddy idea of on-demand access to applications, but some solutions in these areas also leverage utility or grid models to power their services. In an article we ran in September, Amazon CTO Werner Vogels cited SaaS as one of the up-and-coming use cases for utility services like Amazon’s EC2.
To make a long story short, Gartner is seeing what we’ve been seeing for quite a while now: On-demand platforms, whether you call them “grid,” “utility,” “capacity on-demand,” “application virtualization” or anything else, are going to have a huge impact on tomorrow’s IT landscape.
Elsewhere in this week’s issue, we have a great piece from GemStone Systems’ (another company doing distributed caching) Mike Stolz about the pros of implementing an enterprise data fabric, and we have a slew of big announcements. Among those are “[email protected] Recognized by Guinness World Records,” “GigaSpaces, GridGain to Provide Robust Grid Solution,” “OGF, Open Geospatial Consortium Ink MoU,” “Egenera Launches Software Line of Business,” “Platform Computing Acquires Scali Manage Business” “Microsoft Unveils SOA Vision,” “Voltaire Intros High-Performance Storage Router for Fabrics” and “Fred van den Bosch Named Evergrid CEO.”
Comments about GRIDtoday are welcomed and encouraged. Write to me, Derrick Harris, at [email protected].