A Taste for Virtualization

I am still in Sydney but getting close to leaving – and what a change this second week has seen. The high-powered leaders of many nations have now left, and the city is beginning to look a lot better. The security fences are coming down and access to the city is returning to normal.

Following a couple of positive comments over my use of a picture last week to show the extremes of Sydney’s security measures, I couldn’t help myself when I saw a city bus go by with this plastered across its side – and I just couldn’t come up with a better word to describe the transformation of Sydney as it was locked-down.

I have caught up with a couple of old friends from my time at Nixdorf back in the early ‘80s. David Miller, Kim Brebach, Rolf Jester, Kevin McCormack – all folks I worked closely with for many years at a time when Nixdorf made a very serious push into the IBM Plug Compatible Mainframe marketplace – the 8890 product family.

But it wasn’t about the past we spent time on – but the future. And center-stage was the future of virtualization and operating systems, in particular, Linux. Kim Brebach has done a pretty solid job of developing a multi-part feature on his efforts to load a Linux distribution onto his desktop – a hilarious tale about the road he went down and told pretty much in a medieval style. For those interested, I will forward a link. But after the missteps and stumbling, Kim is a real Linux-on-the-desktop supporter.

I also had a chance to kick around some ideas with Rolf Jester a Vice President, Distinguished Analyst – ITU Services Market Strategies for Gartner. Rolf is looking way out there beyond what we see today and thinks much of our future IT growth will come from the CME – Communications, Media, Entertainment – market segment as traditional IT collapses down to just a handful of global servers. For sure, the transaction loads on these servers will be enormous, but they won’t be exciting to work with at all. It’s all going to happen over in CME-land where new data types and formats will drive innovation. And again, the conversation drifted towards Linux – and the future potential of Linux in this space.

And so it’s been this week – Linux, and almost in lock-step, Virtualization. The two seem to be getting as much media time as each other and this, I believe, is because the underlying metalwork, the chip sets and boards, are heading to where there will be no limit to the number of CPUs, or processing engines, that end up occupying just a single socket on a processing board. To take advantage of this new packaging – essentially, an endorsement of the more traditional Symmetric Multi-Processing (SMP),or Shared-memory Multi Processor as I also saw it defined, model – then, for vendors where any number of Operating Systems (OS) may need to share a basic board package, some level of abstraction between the OS and the metal has advantages.

Hence, the discussion about virtualization and the talk about hypervisors – the software that provides support for a Virtual Machine (VM) environment where any number of OS’s can be deployed. Today there’s a number of different views as to how this is done, with some vendors now talking about “hosted, operating system level vs native” VMs. OS-level virtualization has become visible on servers (Unix, Windows, etc) and looks a lot like partitioning where a server is split into a number of smaller “machines” and where there is only ever one OS kernel. IBM’s zVM is one example of a native VM running, as is VMware and Xen while Java Virtual Machine (JVM) and the Common Language Runtime (CRL) for C#, .Net, etc. are examples. In conversations many years back, with a financial services company’s architect designing new applications for NonStop, he suggested that perhaps we should consider Pathway as an example of an OS-level virtualization and I suspect he may be right.

In the research I was doing for this piece, I ran across a pretty good description on wikipedia “system virtual machines (sometimes called hardware virtual machines) allow the multiplexing of the underlying physical machine between different virtual machines, each running its own operating system. The software layer providing the virtualization is called a virtual machine monitor or hypervisor. A hypervisor can run on bare hardware (Type 1 or native VM) or on top of an operating system (Type 2 or hosted VM).”

I just read in InformationWeek (August 13, 2007) where Ann Livermore, talking about virtualization, said “Innovation has just begun. This is an area where you’ll see whole sets of services and services companies”!For those of us working with NonStop and familiar with the principles of the NonStop OS Kernel, the NSK, some of this may sound alien. But let’s pause for a minute here and look what’s going on inside HP.

For starters, we know that Martin Fink heads up BCS and he has managed the Open Source / Linux Organization (OSLO) in Ft Collins, CO. We know Hal Massey, from NonStop, now works directly for Martin and he is hard at work with the next round of hardware to be shared across a number of OS’s – and that HP is firmly committed to blades. We have also seen Linux as part of Neoview supporting load capabilities, and we also have heard from a number of sources that the new NonStop comms controllers will be Linux. Not all of these observations are connected and some usage is temporary – but I think the dots and the lines are all there and it’s pretty easy to connect them all.

Apart from meeting with my old colleagues at Nixdorf, I had the opportunity to drop in on Marty Turner – now with Integrated Research here in Sydney but formerly with Tandem and one of the team that brought us SNAX/CDF, and later NonStop NET/MASTER. Marty is old-school development – it’s all about cutting code and solving client’s problems. It doesn’t matter to Marty – pTAL, C, C++ … whatever is needed. Marty’s a classic power user and as I looked around his home office it was littered with books and servers. He runs Linux on his desktop – multiple Linux’s on top of VMware’s product offering. Marty absolutely sees the need today to be very aware of getting the most out of today’s new hardware packages with the latest chip sets.

And so it is – as we push ahead in the world of NonStop I will not be surprised what we see happening under the covers. As I connect the dots – and I am the first to admit that a lot of this is speculation on my part – and as I put it into context based on what I see my peers doing, then I am both excited by the potential here as well as extremely curious about how HP will accommodate within NSK.Do we see signs of what I am talking about here – or am I alone in my observation? Am I connecting the dot’s all wrong? I don’t think so – I really do see fundamental changes occurring in the direction the NSK OS is headed.

Will this be a big step forward for NonStop? Will it even look like the NonStop product we are all so familiar with – or will the NSK OS morph and become some derivate of the traditional NonStop Messaging System (a la ServerWare, we saw all so briefly back in the mid ‘90s) layered on top of Linux on top of a highly tailored HP-specific hypervisor? In future blog postings I will take a closer look at virtualization and the potential benefits as well as possible pitfalls of virtualization.

Will we be running native NonStop in the future at all? Or will we deal with the new and fresh flavor of NonStop?