If you would like to read the next part of this article series please go to Windows Server 2016 Alive (Part 2)
We’ve all heard the rumors that circulated through the tech community, and they could be boiled down to “Windows Server is dead; long live Azure.” Certainly it’s no secret that Microsoft’s focus – as explicitly stated by CEO Satya Nadella as far back as 2014 – is now “cloud first, mobile first.”
The demise of the datacenter (on-premise, that is) and the eventual irrelevance of the IT professional have been predicted by a legion of pessimistic pundits for almost a decade now. I remember the fierce debate regarding whether the cloud would put us all out of work that ensued way back in 2007 at a Microsoft-sponsored “town hall” event for IT influencers.
Certainly the IT world has changed over the years since, and will change more. However, organizations are still employing IT personnel and the data center infrastructure market was estimated at the beginning of this year to be at approximately $120 billion and growing at about 3% annually. Windows Server runs on a substantial share of those servers. Of course, Windows Server also runs on hundreds of thousands of virtual machines in the Azure and Amazon Web Services public clouds.
Now (amidst speculation that it will be the last one), we have a brand new shiny version of Windows Server to play with. Server 2016 was officially launched in September at this year’s Microsoft Ignite conference, and it’s an apparent attempt to bridge the gap between on-premise and cloud. In this two-part article, we’ll talk about what’s new and what it means to you.
In a post on the Microsoft Hybrid Cloud blog on September 26, General Manager Erin Chapple called Windows Server 2016 “the cloud-ready OS” and said the technologies that it introduces are designed to ease customers’ transitions to the cloud. Microsoft is marketing Server 2016 as the ultimate in flexibility, appropriate both for on-premises servers and cloud deployments.
Some of the new capabilities that are introduced in this version of Windows Server are obviously geared toward cloud computing, such as Software Defined Networking features that are used to manage tenant workloads and tenant virtual networks.
Microsoft also calls Server 2016 a cloud-ready application platform, reflecting a subtle shift in recent years to what I call the “it’s all about the apps” mentality. This in turns goes hand-in-hand with the popular trend toward a DevOps model of computing that is slowly removing the distinctions between developers and operations personnel (IT pros). Increased support for containers give DevOps personnel flexibility to create apps that can be deployed on-premises, in the public cloud or in a hybrid environment without changes to the code. Azure Service Fabric on Server 2016 lets developers build both stateful and stateless microservices that can be more easily scaled into the cloud.
Meanwhile, back home
Despite all this focus on the cloud, many organizations will be looking at what Server 2016 can do for them in the same capacity as Windows Server versions before it: in their on-premise data centers and server rooms. Microsoft hasn’t abandoned those customers; many improvements have been made to the roles and features that you need in your on-premises environment, and new functionalities, such as Nano Server, can benefit both the security and the operating costs of your local network.
Before we look at what some of those additions and enhancements are, let’s address the elephant in the room: the future of Windows Server. The dire predictions that on-premise IT may disappear completely might eventually come to fruition, but I seriously doubt that it will happen nearly as quickly nor as completely as some fear (or others hope). Note that the following are my opinions based on decades of participation in and observation of this industry, and are not to be construed as any sort of insider information.
Windows Server 2016 might very well be the end of WSAWKI (Windows Server As We’ve Known It), in the same way that Windows 10 is a departure from the long-time traditional model for Microsoft desktop operating systems. That is, we can expect OS software on both client and server side to be more “agile,” to move away from fixed “versions” and instead be continuously updated – not just in terms of security and performance fixes but also in terms of features. There are both advantages and disadvantages to this, which are beyond the scope of this article but which I plan to address in another one. The point is that neither is going away in the near future, but both are likely to evolve and morph in ways that their predecessors didn’t.
All that to say this: the new features and functionalities that we see introduced in Windows Server 2016 represent the state of the OS right now, but it’s likely that it will be updated to add more features in the future. Of course, continuous updating works both ways; what Microsoft gives, Microsoft can take away. We might find features removed or interfaces changed in ways that we might or might not like. That might be a factor in considering whether to upgrade or wait. I’m thankful, for example, that I can continue to run the old version of the desktop OS (Windows 8.1) on my media center PC. Although Microsoft dropped it in Win 10 (which I otherwise love), the Windows Media Center feature still remains in Windows 7 and 8.1 for those who use it as their means of recording and timeshifting TV programs.
Coming back to Server, though, let’s look at some of the reasons you might consider it worth your while to upgrade in spite of the possible risks.
Hyper-V has received some tweaking in Server 2016, with improvements to the Hyper-V Management tool that now let you save and use alternate sets of credentials in Hyper-V Manager to connect to a different Windows Server 2016 machine or a remote computer that’s running Windows 10. You can also use the Server 2016 Hyper-V manager to manage earlier versions of Hyper-V on Server 2012/2012 R2 and Windows 8.1.
Hyper-V Manager now uses the WS-MAN (Web Services for Management) protocol to communicate with remote Hyper-V hosts. It was developed as a public standard for consistent interoperability for remotely exchanging management data across different device types. This makes it easier to set up remote hosts for remote management, and it supports CredSSP authentication – which means you can perform a live migration without having to enable constrained delegation in Active Directory.
Another new feature for Hyper-V is discrete device assignment. This is useful when you need to bypass the Hyper-V virtualization stack and give a VM direct access to PCIe hardware devices. Note that there can be some security issues with this, but there are situations – especially with trusted VMs that are part of the hosting fabric – when it’s appropriate and enhances performance. You can read more about discrete device assignment in this article on the TechNet Virtualization Blog.
Host resource protection, which helps to control how large a share of the system resources a VM uses and monitors for high usage, is another new Hyper-V feature. This can prevent an overactive VM from hogging the memory and CPU cycles and thus causing a performance hit for other virtual machines that share the physical machine, or for the host computer.
Here’s another good one that’s going to be appreciated by many folks who’ve wanted this ability for a long time: now you can add or remove a network adapter, and change the amount of memory that’s assigned to a VM (without having Dynamic Memory enabled), while it’s still running – and this is true for both generation 1 and gen 2 VMs that are running on Windows Server 2016 or Windows 10.
Another feature that’s been long-awaited by many people is here with Server 2016. It’s inevitable, when IT pros start working with virtual machines, that eventually we’ll encounter a situation where we think it would be awfully handy to be able to create a VM within a VM. Well, now you can. Nested virtualization gives you the ability to use a virtual machine as a Hyper-V host with VMs running inside it. How cool is that? Note that both the physical host and the virtual host have to be running Server 2016 or Windows 10. There are a few caveats: the physical machine needs to have a processor with Intel VT-x and EPT (Extended Page Tables), and you’re going to need a minimum of 4 GB of RAM assigned to the virtual host. You can find more about nested virtualization in this MSDN article.
At Ignite in late September, Microsoft introduced the latest version of its Windows Server OS, which brings us a number of new features as well as enhancements to some of the old, familiar ones. In this first of a two-part series, we took a look at how Server 2016 improves Hyper-V virtualization with new management capabilities and the WS-MAN protocol, discrete device assignment, host resource protection, ability to make more changes to VMs while they’re running, and nested virtual machines. In Part Two, we’ll move on to discuss new networking and storage features, as well as what’s new in security.
If you would like to read the next part of this article series please go to Windows Server 2016 Alive (Part 2).