The Mainframe as a Cloud
September 7, 2011
When the subject of cloud computing comes up, the subject of mainframes never seems to come up. And yet, just about everybody who works with a mainframe that examines the concepts behind cloud computing can’t help but note that mainframes already provide those capabilities…
It appears IBM, however, is finally starting to get that message across. The company this week announced that Yi Lian Zhong Information Technology (YLZ), a provider of information services in China, plans to develop a cloud computing platform based on IBM System z mainframe systems.
What many IT organizations don’t appreciate, says Bill Reeder, worldwide optimization and System z leader for IBM, is that the attributes associated with cloud computing are already present in the mainframe, including the ability to dynamically scale applications up and down. Given the relative expense of the mainframe, Reeder notes that it was always a requirement from the beginning that IBM leverage technologies such as virtualization to make sure customers get the highest server utilization rates possible. Now, Reeder says the x86 world is trying to address these same issues under the heading of cloud computing. By way of comparison, Reeder says x86 systems trace their lineage back to a workstation box that was never really intended to support multiple simultaneous application workloads, so now providers of these systems are trying to retrofit them to support a new management framework.
With the advent of virtualization, adds Reeder, we’re now seeing traditional x86 server sprawl issues now being compounded by the rise of virtual server sprawl, all of which is conspiring to drive the total cost of distributed computing even higher. Reeder says cloud computing is an attempt to contain those costs using concepts to manage both physical and logical resources that IBM originally developed for the mainframe decades ago.
With the launch of the zEnterprise, Reeder says IBM is now trying to extend the management capabilities of the mainframe out to distributed computing systems as part of an effort to promote a hybrid computing model that allows application workloads to dynamically run on the class of system that is best optimized to handle that function. That challenge facing IT organizations is that in order to take advantage of what IBM calls “Smarter Computing,” the IT organization needs to “right-size” their IT infrastructure with an eye towards creating a more balanced IT portfolio.
As the cost of mainframes continues to fall, IBM is betting that the shift to cloud computing models is going to drive up interest in mainframes, especially when it comes to the consolidation x86 servers running Linux. The real issue, of course, may actually be overcoming the internal IT cultural issues that do as much to prevent the convergence of mainframe and distributed computing as the technologies involved.