Where did cloud computing go? During 2007-2011 it was all one heard. Was there a cloud computing? Or was it just online storage, apps, and such; the new mainframe, dressed up in graphics and fonts? Well, cloud computing was real, but it was one deployment model among many, useful for some problems and not for others.
Similar dangers in the realm of words lurk in the metaphor of “converged infrastructure.” Again, there are real technologies, deployed instances and real customers doing “it.” But when a major customer asked a vendor at a recent conference, “What is a convergent infrastructure?” I got uneasy. The short answer is that there are no clear definitions.
I guess you could say that I’m not comfortable with the term or the feeling I have that it was invented by marketers with product to move, a polyester suit in a warehouse full of cotton shirts. That said, there are many good things to come that will be tagged to the rubric of convergent infrastructure. According to a Forrester Research report in December 2011, there are people implementing some of the things that people say is CI. That’s good; a stride past the peak of the hype cycle.
The report includes data supporting the conclusion that the top of the IT food chain is driving a move to CIs. According to Forrester, the CIO is driving 70% of converged infrastructure projects reported in the study. That tells me two things: First, that the rest of the organization won’t willingly implement buzzword technology; it’s career limiting. Secondly, CIOs are betting their bonuses, so there must be something real.
If you’ll permit me to take a stab at a definition of CI, I’d say: CI is taking advanced storage, applications, virtualization, and management tools and integrating them into a solution that is better than any of them implemented by themselves. The sprawl of a dilapidated IT shantytown full of old projects and legacy systems is hammering IT.
So the path to this convergent future begins with envisioning bounded projects that can be implemented and then expanded. Start basic use cases:
- One could be a need for servers by development teams; a development infrastructure. That way, the people who will build it will know it.
- Create a managed environment from iron to human that can be turnkey activated, provisioned and managed for special projects that includes all aspects of a 21st century infrastructure.
Another good approach would be to create pilot convergent installations and run legacy software on them before moving those apps to the new infrastructure. But with legacy systems designed for a different world, there will likely be complications and therefore caution is advised.
As you sunset older infrastructures, your team will have the process nailed and you’ll be able lose legacy platforms. First, prove the model with something bounded. That way, you’ll have credibility to spare when your team tackles the legacy systems.
Virtualization is wonderful. Managed environments are wonderful. Turnkey integration is wonderful. NAS and SAM are fabulous. But orchestrating these elements into a virtualized infrastructure with marketing words somewhat ahead of the technology (again), achieving convergence will still take a decade.
But this concept, convergent infrastructure, unlike private clouds, is one with realism embedded in it. It’s time to get started whilst containing expectations.