dcsimg

Virtualization’s MF Future is in its MF Past

Curiosity about computing's future can be found in its past: The mighty Mainframe (MF) lives again.

Quick quiz. What do Mainframes and virtualization have to do with each other? Give up? In a single word: Everything. It may surprise you to know that virtualization existed on Mainframes for almost 3 decades before it existed on any other platform. As I was researching a new virtualization project involving IBM System z, z/VM, and Linux, I realized that we’ve come full circle in computing — from Mainframes and dumb terminals to the wonderful world of mid-range computers, departmental servers, and heavy desktops back to those dinosaur days of yesteryear of Mainframes and dumb terminals. Could it be true that we are returning to our roots as mere users on behemoth systems where an administrator allocates a virtual resource slice to us instead of a complete physical system with which to work? The answer is yes. The answer is also no.

It’s true that Mainframe computing is making a comeback in the virtualization space. It’s also true that, for VDI (Virtual Desktop Infrastructure), Mainframes make a lot of sense. Mainframes provide security, stability, a choice of either pooled or dedicated resources, and 40-year history of success that is unmatched by any other existing technology. They’re also a wise choice for Cloud Computing — a natural for Mainframe scalability and dynamic resource allocation.

Cloud Computing on a Mainframe? Yes, and though this be madness, yet there is method in it — the Mainframe/Cloud Computing paradigm have much in common. Consider the following list of attributes.

  • Users lease computing time and resources.
  • Users neither know, nor care, where their data is kept.
  • Security is a primary concern.
  • Server and resource provisioning is dynamic
  • User spaces are logically separated from each other.
  • Computing resources are dynamic.
  • 100% Availability
  • Remote access is operating system independent.

Are the items in the list describing Mainframe computing or Cloud Computing? If you said both, congratulations, you are correct. Cloud Computing sounds suspiciously like Mainframe computing doesn’t it? It sounds that way to old Mainframers who claim that grid computing, Cloud Computing, or whatever you want to call it, is not some new-fangled technology, but has been around longer than most of you. They also claim that those old school Mainframes are better at the job than the hottest googolplexed x86 processor-based machines will ever be. So what else makes the System z Mainframe desirable as a Cloud Computing platform? z/VM is the most mature type of virtualization, as I’ve already stated, with over 40 years of proven success. Coupled with its history and experience, z/VM is a Type-1 hypervisor much like Xen, VMware’s ESX Server, and Virtual Iron, which aligns it with other current virtualization technologies.

Where’s the downside to using Mainframes for Cloud Computing — or the no part of the answer to the question I asked in the first paragraph? Mainframes still send your bank account into the red. If you duplicated a Mainframe’s processing power in x86 architecture, you would spend at least twice as much on the Mainframe. For your efforts, you’d receive more processing power from the x86 Supercomputer, however there’s more to Cloud-based computing than pure processing power. Mainframes possess superior I/O, which is an often-overlooked aspect of virtualization — especially when considering database virtualization or providing an acceptable user experience for virtualized Desktops.

Does this all mean that you should start looking into the purchase of a Mainframe for your company? Certainly not. It does mean, though, that Cloud Computing vendors who use them, or foresee using them for their services, will be able to meet any resistance to their services by just stating the facts about Mainframe computing.

Since Mainframes aren’t new, or even perceived as new, they don’t garner the same hype that surrounds emerging technologies, there are no buzzwords or trade show excitement following them, and most people think they’re ancient history. They still exist. They’re just there — as they’ve always been — working.

Comments on "Virtualization’s MF Future is in its MF Past"

pfloyd

So many similar articles focus too much on the cost of the mainframe hardware; everyone knows hardware is not the significant cost center it used to be – people and software are the big differentiators. I just completed an in-depth TSO study that compared the cost of running 160 virtual servers on the mainframe to 160 servers on Intel platforms. It wasn’t even close – the mainframe was millions of dollars cheaper when ALL of the costs are totaled out over three years. When the number of servers climbs above 160 the mainframe’s advantage is even greater because floor space, power and environmental factors play a larger part. Most software is licensed on a per-processor or per-core basis, and fewer larger processors (on mainframes) translate to significant savings. Savings on SAN and network port infrastructure is very significant as well; and time to procure, build, harden, and deploy are big manpower differentiators.
Virtualization on Intel will narrow this gap – and we are pursuing the same TCO study comparing against a virtualized Intel infrastructure, but software-per-core will still give the mainframe the advantage I suspect.

Reply

Aw, this was a honestly nice post. In notion I would like to put in writing like this moreover – taking time and actual effort to make a pretty superior article?- but what can I say?- I procrastinate alot and by no means appear to get something completed. allourweb.

Reply

Terrific work! This is the type of info that should be shared around the net. Shame on Google for not positioning this post higher! Come on over and visit my site . Thanks =)

Reply

In this awesome pattern of things you actually get a B- with regard to effort. Where you lost us ended up being in the facts. You know, it is said, the devil is in the details… And it could not be more true at this point. Having said that, allow me tell you precisely what did deliver the results. The authoring can be incredibly powerful and that is possibly the reason why I am taking the effort in order to opine. I do not really make it a regular habit of doing that. 2nd, despite the fact that I can see the jumps in reasoning you come up with, I am definitely not convinced of just how you seem to unite your ideas that produce the actual conclusion. For now I will yield to your point but wish in the future you actually connect your dots much better.

Reply

Leave a Reply

Your email address will not be published. Required fields are marked *

*

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>