If you are like me, you have probably followed the entire discussion around grid computing with a detached sense of curiosity: a feeling that the idea sounds cool but in practicality seems more science fiction than science.
When we add the fact that every vendor seems to want to bend the notion of grid computing to their own particular strengths and weaknesses, it simply feeds my feeling of confusion and detachment.
The term grid has been conflated to mean everything from a simple cluster of databases to data integration, advanced networking or an ASP. It leaves me wondering if the term grid means anything any more and, if so, does a problem actually exist that would require the need for grid technologies?
My understanding of the original grid concept was a way to achieve coordinated resource sharing either within an organization or more grandly, between many organizations. As for what resources could be shared, it would go way beyond simple file sharing to direct access to computer excess processing cycles, programs or storage, among other things. Some have termed this the virtual IT organization.
Consider the future of IT as a virtual organization where there exist storage providers and compute cycle providers, ASPs along with consulting organizations hired by a single entity to work on a single problem.
It might be analyzing data to determine the optimal location for the next retail outlet of a major drug store or for an insurance company to compute models and set rates for insuring coastal communities.
The truth is most of us cant conceive yet how we would utilize a true grid computing infrastructure. Indeed, I would imagine that even the brains within the Globus Consortium who are working on grid standards cannot predict what innovation—and therefore what uses—there will be when grid computing becomes a reality.
Think of the Internet, which was created by the U.S. Defense Department as a way to link computers among university researchers and government. Do you think that 30 years ago, when the Internet went live, that they envisioned the marriage of hypertext to TCP? Did they think what they were building would change the world of commerce and media? I doubt it.
The grid computing problem seems to me to be quite daunting in scope. It requires an entirely new set of standards or, at best, the significant evolution of standards used today by the Internet, business-to-business exchanges or even distributed computing technologies such as CORBA or Enterprise Java. While we have some early examples of sharing resources in limited ways, none can accommodate the broad range of resources that a virtual IT organization would require.
All the experts talk about the need to develop agile organizations, and certainly grid computing and the whole notion of service oriented architectures, of which I believe grid computing to fall under, will help make true IT agility a reality some day. Maybe not for another five years but it will happen.
Imagine the impact on the entire information technology industry of dynamic relationships among organizations. When one company may "rent" resources of all kinds to other companies and vice versa? The implications to licensing, sales, maintenance—indeed the entire vendor-client relationship—could change forever. While Im still confused enough about the whole grid concept to not see the future implications yet, I sense they will indeed be interesting times.
Here lies the cautionary message. The emergence of true grid computing technology will undoubtedly create expectations, give birth to new companies, and make claims to creating a new world order where old rules no longer apply. Gee, that sounds awfully familiar.
I just hope that this time I catch on with a company about to be wildly overvalued even though no one understands what value we deliver. I just hope that this time I get out before everyone realizes its just Mexican food. The same stuff folded differently.
Charles Garry is an independent industry analyst based in Simsbury, Conn. He is a former vice president with META Groups Technology Research Services. He can be reached at firstname.lastname@example.org.