Are there any best practices around virtual servers that will help solution providers optimize their customers' infrastructure without causing performance issues for the end users?
One of the big benefits of advanced virtualization is that it allows things to be much more dynamic. Separating the virtual from the physical allows movement of workloads much more easily within the data center environment, so as I'm trying to better manage my data center, I want to move a workload from one physical server to another for balancing, to patch my servers, and so on. So the data center becomes much more fluid and dynamic in the management of those workloads, and because those workloads require connectivity to other places, it requires a more dynamic nature of the networks as well. As I move the workloads, I need to move the network the connectivity and security that goes with it.
We are also seeing a significant increase in desktop virtualization interest, whereas early on, there were a lot of technical problems around desktop virtualization, many of these issues have now been addressed and so we're seeing a big increase in the number of users that are now moving forward -- twice as many people doing it this year. And so from the solution provider perspective, for those who have some expertise in that area of growth, there will be a lot of activity.
Desktop virtualization has a lot of networking aspects to it, including performance issues, WAN acceleration, and WAN optimization-related issues that will require virtualization expertise and networking expertise.In terms of solution providers, what are some important things networking solution providers should keep in mind when introducing virtualization to a data center or offering virtualized services?
One specific thing to keep in mind is the shift to virtual switches. When I consolidate from physical to virtual servers, multiple virtual workloads occur on one physical server; so if server A used to talk to server B, it went through a physical switch, and I had to do all management through a physical switch. But if they're virtual servers on the same physical server, now the traffic between them only goes through a virtual switch so the management and security issues must be addresses with the virtual switch. And that's going to cause changes.
Another change in virtual switches comes with the newest release of vSphere, which allows virtual switches that run the same software as a physical Cisco switch. So before, when managing my virtual switches, the management generally switched from the Cisco team to the VMware team. Now, because those switches are running the same software as the Cisco physical switches, management can switch back to the Cisco team and is consistent for both physical and virtual. This enables organizational changes as well as better management from a technical standpoint.
For users looking to hear "gotchas" and tips and techniques from real experience, there's a "Virtualization: Life in the Trenches" session, which has a panel of practitioners -- so not the marketing guys but the implementation guys. For example, they see the kinds of network problems that can occur with dynamic movement, too much of which they call vMotion sickness. There is a serious impact when things move around a lot, like security issues and zoning issues. You're moving things to places you didn't expect to, so the implications really have to be thought through. For solution providers, this session will be a good one for hearing about the issues that others run into.
For folks that are new to any specific areas of virtualization, there are primer sessions available in each track -- Virtualizaiton Concepts, Virtualization Management, and Desktop Virtualization. For those who are further along, there are best practices panel sessions in each track as well -- discussing strategies, best practices, and common mistakes to avoid.Should solution providers create their own set of standards to pass onto their customers' networks to ensure efficient virtual server use and minimize problems? Or is every network so different that there isn't much in the way of standardization?
Even if one environment looks just like another, chances are they won't be the same. They may have things in common and still be different, but there are still a lot of best practices that will go across the board both technically and organizationally. We've been in silos for so many years in IT; and, periodically, some technologies have come along that say we need to fix these silos. Storage networking forced some level of improved communications, for example, between storage and networking people in order to make it work. I think virtualization is another one of those technologies that is forcing a need for better communications, because as we're virtualizing, we're doing everything, and all of the pieces need to talk to one another -- networks, storage, server, desktop.
As everything moves toward private clouds and public clouds, it will require all of the groups to work more closely together. We encourage cross-functional teams to be working together, and I think that's one of the best things solution providers can help to make happen. Get the teams to talk so that for example, the server guys don't just go off and implement server virtualization and cause problems for the network guy.
That's an example of organizational best practices, and there are technical aspects of that as well along the same lines. Best practices and communications are key, and that's a piece that solution providers can learn from in their early implementations and carry forward. Doing so will help smooth later implementations and avoid the potholes that you can run into if you don't follow those kinds of concepts.