Microsoft's server and tools head Bob Muglia recently noted that businesses are looking to use cloud computing in different ways: some intend to use it to revamp their IT infrastructure, while others just want to extend their existing setup. Upcoming updates to Microsoft's System Center portfolio of management tools — which includes Virtual Machine Manager, Operations Manager and Configuration Manager — add in cloud support geared to meeting these needs.
At its TechEd event earlier in June, Microsoft gave attendees a demonstration of Virtual Machine Manager (VMM) v.Next 2011, focusing on how it expects server application virtualisation to uncouple applications from the operating system and so give businesses more flexibility in managing workloads.
Ryan O'Hara, Microsoft's senior director for the System Center management suite, talked to ZDNet UK about the background to these latest Microsoft technologies for managing physical and virtual PCs. He also spoke about what the company expects from Intune, a forthcoming cloud-based PC management option targeted at small and medium-sized businesses.
Q: What is behind the integration of Microsoft's Forefront security suite and Configuration Manager?
A: Anti-malware and PC health is now a part of the overall PC lifecycle. The idea of System Center and Forefront working together is to give a more unified experience across that lifecycle.
Does that mean in the future people won't be able to use Forefront without using Configuration Manager?
They will need a Configuration Manager infrastructure, that's absolutely the case. They could use something else for PC management, but it's our hope that we'll demonstrate enough value in Configuration Manager that they won't want to do that.
That makes a lot of sense for organisations above a certain size, but what about smaller companies?
Customers using Configuration Manager and Forefront today span from the largest of the large through to organisations roughly of 500 PCs. Somewhere around 250 to 500 PCs, the dynamics of the IT organization change. There's a line where there's no longer a desktop administration team; you have more of an IT generalist staff. It's at that point where the care and feeding of infrastructure like that no longer makes sense.
For those customers we have Windows Intune, which is in beta testing today and will be brought to market within the coming year. Windows Intune is a hosted cloud-based service for managing and securing PCs. It allows IT generalists to focus on setting policy and managing PCs, instead of the infrastructure of PC management.
We think that over time, enterprises will also want to take a look at Intune, because cloud-based desktop management is really attractive in terms of changing the cost model. So as the features in Intune mature and reach greater parity with [System Center], it becomes a real option for enterprise customers.
Could Intune also be used with cloud-based virtual desktops?
It certainly has that kind of potential for future scenarios. Management of virtual desktops and virtual application packages like App-V are not in the Intune capabilities that are in beta today. But certainly you could foresee something like that in future.
Virtual Machine Manager 2008 is a key component of System Center. How has it evolved in the forthcoming VMM v.Next 2011?
We've been demonstrating a technology called server application virtualisation. This packages a server-based application, so its state is separated from the operating system. Traditional server applications tend to...
...distribute their state across much of the server OS, and this creates a hard link between the application and a specific machine or OS image. By separating that state, server application virtualisation allows that application to be independent of the OS. That's a critical technology for us as it changes some of the downstream operating paradigm.
How does VMM v.Next2011 make it easier to patch server applications?
Think about it as allowing you to have one standard image for your databases, one standard image for your application tiers, one standard image for your web tiers, [and so on]. Now you can manage those three images through monthly patch cycles.
That's very different from having one image per IIS [Internet Information Services] instance, where you might have 15 flavours of IIS packages tethered to operating systems and each one is different, so each time you have a patch you have to go and test 15 operating system images. There are slight variations of them, and the same for your databases and applications. The library gets greatly simplified as a result of those packages that we've been showing lately.
When you're ready you push out those [standard] updated images, you could use server application virtualisation to forklift the workload to replace the underlying OS, and you could also automate the whole process. Today we can automate the patching of clusters using Opalis, where we take a SQL Server pool and drain one of the nodes of the pool, patch it, bring it back on line and let it rebalance. Then take another one off line, drain it and so on — all through a pre-established workflow.
Image-based patching is going to be a very real and majority scenario. It's far more efficient, and you just don't have the error rate [of traditional patching].
Are the new features and products more focused on public or private clouds?
As you think about going from private datacentres to public clouds, the need to manage infrastructure gets diminished. All that matters is the app. So we'll focus our development on application management. For example, new management pack technologies in Operations Manager give you application insight no matter where that app lives.
Our commitment is to follow that application and make sure we can instruct fabric managers on behalf of the app. Today in customer datacentres the Microsoft fabric manager is VMM. In Azure, that fabric manager is the Azure fabric manager. System Center will have enough insight into the application to talk intelligently and instruct those infrastructure managers to provision the right resources.
On the infrastructure side, VMM 2008 knows about the resources that have been provisioned for particular workloads. If those resources have been provisioned by a third party provider using VMM, you could see those resources in your VMM console.
In future, we'll incorporate views that show provisioned capacity, let's say, from Azure. We'll make sure those experiences will be as continuous as possible as you go to a cloud scenario, whether it's private or public.