At Microsoft TechEd 2013 this June, Microsoft began releasing information about some of the changes and additions to the upcoming Windows Server 2012 R2. Virtualization pros should take note, because the changes coming to Hyper-V and System Center R2 (especially System Center Virtual Machine Manager) are especially important.
While the previews at the show indicate a clear direction, Microsoft could still change or drop any feature by the time it releases Windows Server 2012 R2.
There are several important additions to Hyper-V:
Tech Ed video demo
- Live Migration -- The source server will now compress migration traffic before sending it over the network, decreasing migration time by up to 50%. This includes the ability to migrate over Server Message Block 3.0, which provides an even greater performance boost.
- Improvements to the VHDX format -- IT pros can expect to be able to expand or shrink a VHDX file while the virtual machine (VM) is running. VHDX files will be able to be shared between two VMs using a virtual SCSI controller to add flexibility to clustering solutions by removing the iSCSI or virtual Fiber Channel requirement.
- Live exporting and cloning -- A VM can be cloned or exported while it's running
- Storage improvements -- One of the more impressive improvements is the ability to control the storage IOPS per VM.
- Windows Azure -- Because Windows Azure is using Server 2012 Hyper-V, you will be able to take on-premises VMs and run them in Azure.
These are a few of the more impressive improvements to Hyper-V, and there are more, to be sure. I think the near future holds a much more flexible virtualization environment where you can easily scale out from your private cloud to the public cloud (Azure), and then back when resources are no longer needed. The increase in flexibility -- making it easier for IT departments to adjust quickly -- will help companies trying to maximize return on investment.
For me, one of the biggest improvements isn't built into System Center or Hyper-V, but comes in the operating system through PowerShell. In a nutshell, Desired State Configuration (DSC) lets administrators write a simple declarative script (or perhaps use a simple graphical user interface) that describes what a computer should look like. The administrator doesn't need to write all the plumbing scripts, install roles and features, and make registry changes. Instead, DSC takes care of all of that. You simply describe what you want.
This will make physical server (and possibly client) deployments much faster and simpler for virtualization admins to build. The real advantage is in System Center Virtual Machine Manager (SCVMM). Imagine rapidly deploying or scaling out to the needs of the business by clicking a button, then scaling back just as fast.
To really grasp how important this will be to us server virtualization folks, you need to see it in action. Kenneth Hansen and Jeffrey Snover hosted a demonstration at TechEd that is worth watching. See if you can catch the SCVMM part of the demonstration. This is a deep look into DSC, how it works and what you can expect moving forward. I know this is a PowerShell session, but by putting DSC into the OS, many product teams can expose and use this technology. I can't wait to see how the rest of the System Center teams may implement this, but it would appear the SCVMM team has already started.
For the average IT pro, the glitz and glamour of new features in a soon-to-be released product hold interest but not much practicality. Implementing something just off the shelf is typically a no-no, and it's hard to move forward quickly with OS upgrades and product enhancements. Take a moment to consider why Microsoft is adopting a new rapid-deployment strategy -- releasing new products every 12 to 18 months. Think of the new releases as service packs: You have no problem testing and deploying service packs rapidly. In the past, Microsoft has been restricted in releasing new features inside a service pack, which has slowed us all down from being able to implement better solutions in our rapidly changing business environment. The new rapid releases allow Microsoft to build in incremental changes into a service pack.
It has yet to be seen how well this will work going forward for the IT pro who must avoid the dangers of upgrades to production while taking advantage of new technology. I'm sure there will be some “teething" problems along the way -- but it's something to start planning for, even testing to see how agile you can make the process. Ultimately, the business will dictate your strategy, and upgrading "just because it's cool" doesn't make sense. However, I think you might find that some of the problems you're facing today have solutions that you will be able to move toward rapidly.
This was first published in July 2013