Much is being written about how Grid is a paradigm-shifting, barrier-breaking technology that is going to transform not just the data centre, but the way enterprises develop and deploy applications. But, while the marketer in me appreciates such boundless enthusiasm (and the variety of news coverage the topic generates), I thought it would be useful to look at the topic from another angle: Why are more companies not adopting Grid? Or, why are they not adopting more of it, faster?

Lack of understanding
As distributed computing has evolved, many catch-phrases have been used, especially as marketing machines continue pumping millions of dollars into propagating each unique label. For some companies, it’s software. For others, it’s hardware or services. And sometimes, it’s a vision or a brand that encompasses all three. But while vendors develop a new buzzword in the hopes of creating a market distinction and — they hope — a market advantage, in the end, what we’ve really created is confusion.

Without question, if our buyers – you the users – don’t have a common language to discuss problems and solutions, it slows things down. This confusion perpetuates a lack of understanding about this technology. At DataSynapse two to three years ago, the questions we were fielding during evaluations centred around, "What is Grid?". As the market matured, the questions have shifted to: "What does it do, exactly? What will the impact be? Why do I want it?". And, most frequently these days, "How do I get started?".

To address this new need for customer understanding and action, it’s imperative to steer conversations toward the problems Grid can solve, including proven examples of what this technology can do for your business.

Resistance to change
Another hurdle that can’t be discounted is the natural resistance to change that exists within the enterprise. Evangelists of Grid sometimes encounter the attitude that “good enough” is good enough. Interestingly, though, the little adage about not fixing things that aren’t broken doesn’t apply in this case, because while broken may be the wrong word to describe enterprise technology today, there is pain within the enterprise when application performance, scale and reliability issues arise. But still, it’s difficult to battle inertia and to get folks to embrace new ways of solving old problems. This is because change is difficult and as we all know, it’s human nature to avoid change unless absolutely necessary.

For those who’ve spent years building intricate “plumbing,” the care and feeding that legacy distributed systems often require can translate to job security, even if those home-grown solutions may not be getting the job done as efficiently or effectively as possible. And of course, change often equals risk. Proponents of Grid must be able to articulate the risk/rewards scenario and the expected impact of a successful Grid implementation.

The fact of the matter is, Grid represents both an evolution and a revolution. We all acknowledge that most enterprises have been doing some form of distributed computing for years. So perhaps, implementing Grid is merely an evolution from home-grown to packaged technology, so enterprises can redeploy IT resources – away from “minding the infrastructure” and onto other value-add projects.

And yet, the impact of this technology – up and down the entire stack – means that it is also revolutionary. Why? Because it has the power to potentially change the way enterprises buy and deploy software, hardware and ultimately, the way they manage a service-oriented enterprise.

Cultural impact
Closely related to “resistance to change”, the fear of the unknown prevents many a journey. Because it’s not well understood, cultural impact is one of the more widely reported inhibitors to Grid adoption.

As Grid software breaks down the silos that exist between applications and business units, the simple fact is, people have to learn to share. Grid delivers the power to distribute application service requests across a pool of shared resources that are dynamically expanding and contracting according to business demand – regardless of who owns those systems or where they’re located.

The technology exists, but enterprises are simply not set up that way. If one business unit pays for those resources, there’s a proprietary sense of "Why should I share? Let them go pay for their own". Often referred to as “server-hugging”, this is one of the most common sticking points cited early in Grid software evaluations. Even if the resistance to sharing is overcome, there are still other questions to answer.

Users often ask, "How do I know that - if I share, I’ll still get what I need done - when I need it?". What’s lacking is the sense of trust in the Grid’s ability to guarantee execution of service requests, based on policy, priority and user-defined business rules.

Technology impact
Though many companies have already started adopting Grid, there are still many questions around where the technology fits within the IT landscape. How will it impact current and planned infrastructure? Most significantly, what applications fit on the Grid? Which make sense and which don’t?

Unfortunately, the perception exists today that Grid is only good for high performance computing applications. While it is an obvious and easy place to start for most enterprises, it doesn’t represent the sum total of opportunity for Grid within an enterprise. For example during DataSynapse GridServer implementations, applications are assessed based on multiple criteria (eg, unit of work, I/O requirements, whether the workload is synchronous/asynchronous, stateless/stateful, etc). Applications are then plotted in a quadrant that maps ease of integration against business value.

There are two other issues that fall under the heading of 'technology impact': standards and security.

Standards are evolving, but slowly. Because of the overlap with so many other technologies like Web Services, SOA, and traditional distributed computing, a number of standards bodies are developing standards related to Grid computing including the W3C, OASIS, IETF, DMTF, WS/I, and the currently merging EGA/GGF, and others. While it is not practical for vendors to support all of the standards in the space, a combination of industry adoption and standards maturity will eventually clear away some of the confusion.

Security also gets a lot of airtime, especially in situations where the enterprise is deploying Grid across desktops. In a shared environment like this, IT must be able to reassure users that the only thing being scavenged is processing cycles – not proprietary, business critical information.

Software licensing
Arguably, software licensing is probably the reason most talked about to explain why companies are slow to adopt Grid, right behind the cultural inhibitors to Grid adoption.

In a comprehensive report on software licensing (GARS4, Impact of Software Licensing), The 451 Group asserts: "As [enterprises] evolve into using grids as more mainstream technology, the restrictions of current software licensing will become an even greater obstacle."

It’s a pretty succinct summation of the limitations that current licensing practices (per CPU, per seat, per user) place on Grid adoption. Without question, the new computing models will require new licensing models. Grid is just one of many catalysts spurring this dialogue.

While much has been reported about how ISVs are uninterested or unwilling to address Grid, there is progress. A growing list of ISVs have embraced Grid because it’s a way to boost customer satisfaction (eg., Algorithmics, Calypso, Milliman, Reuters, etc). In some cases, they’re announcing OEM agreements that embed Grid capabilities in their software to offer out-of-the-box integration to their install base – and all the inherent benefits in improved application performance that come with it.

Summary
So, who’s NOT afraid of Grid computing? Actually, there is a prestigious and growing list of global firms – many household brand names – who are willing to speak publicly about the significant and measurable value they are deriving from Grid. Moreover, these are companies which, in many cases, are expanding the size and scope of their existing implementations to move toward enterprise Grids – virtualising multiple applications, across multiple lines of businesses and geographies. The case studies are out there – at events and on Webcasts, like those offered by DataSynapse.