At The Whiteboard

Enterprise Class Virtualization

Hello. My name is Dan Chu. I'm Senior Director of Product at VM Ware. Today I'm going to talk to you about enterprise class virtualization. In our last white board, we talked about what is virtualization. Now we're going to talk about what it means to actually not just to virtualize a single server environment or just a couple but actually deploy virtualization across the entire enterprise. Today we have over 20,000 enterprise customers who have adopted virtualization broadly and 90 percent of them are actually running in production their business critical applications and workloads in virtualized environments. So we're going to talk about what enables that and what additional benefits companies and CIOs are getting from deploying virtualization across their data centers.

So specifically, we're going to talk about three major concepts and three major requirements for enterprise class virtualization. One is scalability and performance. What we're talking about is deploying virtualization across hundreds or even thousands of virtualized server environments-not just one machine but a large number.

Now, across each of these environments you have many applications running side by side. And as opposed to that one initial environment that you might have been running for development or non-production workloads, these are your databases, these are your e-commerce applications, your web servers, your domain controllers and you want them to run at high performance and with high scalability.

Now, at an entry level oftentimes people use virtualization solutions that run on another operating system. So virtualization software in that case is simply an application that gets installed and it's very easy to get it up and running. On the other hand, for enterprise class virtualization where performance and scalability and fidelity and uptime are critical, customers overwhelmingly have standardized on bare metal architectures or hyper visors that run natively on the underlying hardware.

The benefits of these are manifold. There isn't a single point of failure or the security vulnerabilities of an underlying OS. And just as importantly, there is highly optimized performance because the virtualization layer really runs directly on the hardware and there isn't the inefficiency of the intermediating operating system.

The second one is availability. If you're a CIO and you've deployed virtualization across your entire data center you want to get the benefits of having even better uptime than you had before in your more chaotic physical environment. You want to have your servers be able to allow you to have a much stronger infrastructure and there are a couple of key ways that customers are achieving that today.

One is being able to V-motion or live migrate their applications. So, in essence, if you have a server that you need to upgrade or you need to do maintenance on, you can actually take the applications and have them instantly move over to a neighboring machine without any down time. That business application, that web server moves over without any interruption to the actual customer's users using it and you can move them right back over when you're done upgrading or replacing that machine.

Also, you want distributed availability, which is actually if you have a server that goes down you want your virtualization layer to have the capability to automatically restart those applications on neighboring machines where there's capacity and be able to automatically fail over those business critical applications.

A final significant benefit that customers are getting is management. And again, in moving from the server sprawl of their past physical environments, when they move to these virtualized environments stretching across hundreds or even thousands of servers, they want to be able to manage, provision, update and monitor these servers even better than they could in the past traditional physical environments. And so you want a holistic management capability across all these servers and across all these virtualized environments.

And even more than that, you want the capability to have distributed resource management. In this case, if you have a specific application-and all applications never run at a constant level-if you have a trading payroll application, for instance, and it runs at a low level utilization 99 percent of the time but every two weeks when that payroll goes out it spikes up to full utilization. And you want that capability for that server to be able to automatically take that spiking workload and move it on over to a separate machine that has a lot more capacity and to be able to automatically do that and preserve the service levels you have to your end users. And again, it's another key feature that people are getting in terms of enterprise class virtualization.

So overall, customers are finding huge value to virtualization, not just at a small scale but a whole pool of differential value they're getting from implementing virtualization across their data centers, across their enterprise.