sorted by: recent | see : popular

Content Tagged with infrastructure + virtualization

The virtualization of systems allows for efficient use of server resources and is clearly a trend that many enterprises are embracing. Systems engineers see virtualization as the next generation of tools that can help scale their servers, while network engineers see the virtualization trend headed in their direction as well. Unfortunately, it seems that server virtualization also helps foster trench warfare between the two.

I found myself witness to one small skirmish in this battle today, when I met with a startup looking for funding. The startup is building enterprise services, and for its next generation plans to make heavy use of XenSource’s XenMotion functionality to manage virtual machines on about 50 physical servers. This functionality, which is similar to that of VMware’s VMotion, promises to seamlessly move a virtual machine from one physical server to another. The startup’s service product could be running in one virtual machine on a server and if the server receives too much load or has a failure, the XenMotion functionality could move the virtual machine to another server without resulting in any downtime. For an enterprise services startup, avoiding downtime is a good idea.

I asked some questions about the network and systems architecture and found that the systems engineers had made the assumption that in the new service, any virtual machine could be allocated to any physical server. The network engineers, unfortunately, had not taken this into account. Based on the physical network topology — a classic three-tier architecture — the network engineers had set up firewall rules and access-control lists to appropriately protect the infrastructure. For example, not every server could be accessed from the Internet and only certain physical servers had permission to mount storage area network resources. If using XenMotion meant every server was expected to house any virtual machine at a moment’s notice, these were clearly issues that needed to be resolved.

The systems engineers’ expectation of being able to move any virtual machine to any physical server in the infrastructure meant a complete redesign of the network topology was required. And that is when the skirmish ensued. The systems engineers insisted that the network topology be set up to allow XenMotion to work seamlessly. The network engineers argued that their network topology was necessary for scalability and security. As far as I was concerned, they were both right, so before continuing my due diligence on their business, I sent them off to settle their skirmish amongst themselves.

But it had got me thinking: Has server virtualization added an abstraction layer that further separates systems engineer and network engineers from the physical reality of their environments? Do we need a new engineer — a virtualization engineer — that understands how virtual machines are allocated on physical servers and networks to act as a liaison between the two factions?

If only it were green. That’s what I found myself thinking after Sun Microsystems said it’s paid an undisclosed amount for Innotek, the makers of open-source desktop virtualization software VirtualBox Virtual Machine. With an earth-friendly spin, Innoteck could have created a trifecta of buzzwords that would have really pumped up its valuation.

However open source (please recall Sun just spent $1 billion buying open source database guru MySQL) and virtualization (think VMWare’s buy of Thinstall in January), are plenty good, especially given that Citrix paid $500 million for open source server virtualization company XenSource. For those of you wondering who might be next, think about Parallels (formerly known as SWSoft), whose software competes against VirtualBox.

Right before the Christmas holidays I got a chance to catch up with Dr. Mendel Rosenblum, VMWare’s chief scientist and one of the company’s four five co-founders. Rosenblum is also an associate professor of computer science at Stanford University, where he leads a group focused on operating systems research. It was at Stanford where Rosenblum and three of his graduate students — Scott Devine, Edouard Bugnion and Dr. Edward Wang — came up with the idea that led to VMWare (VMW). Diane Greene joined them as CEO and the fifth co-founder and the company went public in August, garnering a multibillion-dollar valuation that triggered a virtualization frenzy.

Given that VMWare was in a quiet period prior to the release of its quarterly results, my conversation with Rosenblum was quite general. But he did share with me, among other things, the story of how VMWare got started and his outlook for virtualization in 2008. Here are excerpts from the interview:

How did VMWare get started?

I was a professor at Stanford University and we were building a supercomputer called the Flash Machine. I didn’t want to crunch numbers on this machine, but wanted to use virtualization to see if we could run commodity OSes on [it]. We could, and we wrote a paper about it, and that generated a lot of interest, including from Microsoft, who emailed us and wanted us to come and present to them in Redmond. My grad students who worked with me on the project thought we could commercialize the technology, and in 1998 we launched VMWare.

What was the plan when you launched it?

Clearly, the technology was going to be hard to commercialize, and we decided to focus on doing virtualization on the desktop. We worked on the technology and my wife took care of the business side of things.

It seemed to have been a long time in the making.

It took a lot longer than I thought it would take. We released it first on the Linux platform, because we felt the Linux community would adopt it much faster. That proved to be a good move.

Funny now that you have proved it, there is competition coming out of the woodwork. Oracle and Microsoft, for example.

VMWare clearly is going to have competition. Sure it was nice when we were all alone, but we are very different from these other companies. Oracle and Microsoft, for example, are focusing on single machines for now. That’s a nice thing to do. We used to do that. It is good for server consolidation and it is easier and simpler.

What we are doing is basically coming up with a new way to run the data center. So from that perspective, we will continue to have something better than others.

Let’s talk about the data center for a minute. Do you think the whole architecture of the data center needs rethinking?

We went down a rat hole on how we built the data centers. I am not surprised with all the problems we are having with data centers. In my opinion, the architecture has problems because it was built with inferior solutions. What you had was people placing services on servers in a way that led to lightly loaded machines that were idle most of the time. The whole thing was built for peak performance (and not maximum utilization.) Well, idle machines use as much energy as fully utilized machines. The way out of this is to put more on the machines, and get them to be more efficient and take on the work load that will, to some extent, lower the power consumption.

I wrote about pizza boxes becoming a problem, mostly due to low utilization and higher power consumption. It kind of ties in with your thesis.

You have to see them not as boxes but as resources. People are now beginning to utilize virtualization and federate these pizza-box servers. I think if you start to view them as one unit, you can get more utilization out of them. I think in coming months you are going to see a big push to make all servers (and other hardware) inside a data center look more like a single unit. Ironically, if you look at the future — low-end pizza box servers with multicore CPUs running our software — you will start to see the big machine we were building where we got started.

What is your forecast for 2008 from a virtualization standpoint?

We are in a transition period. I think a lot of people dipped their toes in virtualization and got started with server consolidation. They bought into it the “money-saving” argument. In 2008, I expect people to fully embrace virtualization and extend it to other parts of their businesses, even bringing it in-house and using it for optimizing their desktop infrastructure. More importantly, you will start to see the long-term impact of virtualization in the next 12 months.