For years, command-line enthusiasts of the Unix/Linux persuasion lambasted Windows Server OS for being GUI-based, which opened up a ton of security holes. With the release of Windows Server 2008, Microsoft introduced a minimalistic solution called Server Core. This reduced-codebase, GUI-less option allowed you to go with a lighter footprint on your servers, resulting in a smaller attack surface -- one that should have pleased command-line lovers.
Server Core has evolved in the years since it was released. The first two versions prohibited you from switching to GUI mode from Server Core. But with Windows Server 2012, the ability to switch back and forth was added. Extra enhancements made Server Core more accessible; for example, the inclusion of .Net Framework pieces enabled applications to use Server Core that couldn’t previously. With Windows Server 2012 R2, Windows Defender was added and enabled by default for greater security.
Despite all this, Server Core never really took off. It may have had strong appeal in some circles, but in the main, Server Core was mostly ignored.
This time things may be different in the form of Nano Server, the next-generation (and even smaller) Server Core forthcoming in Windows Server 2016.
Windows Server 2016 is getting fantastic early reviews on multiple fronts due to impressive efforts to modernize the fleet. Hyper-V improvements, containers, storage/disk feature enhancements, and more suggest Server 2016 could be stellar. With the introduction of Nano Server, the idea of a “just-enough OS” version of Windows may finally have legs.
Jeffrey Snover, Technical Fellow at Microsoft, called Nano Server a “purpose-built operating system designed to run born-in-the-cloud applications and containers,” as part of the Nano Server release announcement.
Nano Server will forgo the whole GUI/non-GUI approach of Server Core in favor of a full remote management approach. Don’t think RDP -- think PowerShell or, more appropriately, Core PowerShell, which uses CoreCLR instead of .Net. Also note that this doesn’t mean no GUI, but rather remote GUI. Snover said initial results are promising, with Nano Server having a “93 percent lower VHD size, 92 percent fewer critical bulletins, 80 percent fewer reboots.”
The concept is clear, and the value is obvious: headless, small attack surface, better performance, and so on. The question is whether it will be used, or if it will remain an interesting plaything for the lab. From where I’m sitting, the unique difference between Server Core and Nano Server, aside from the technical distinctions, is the cloud.
Snover says Nano Server focuses on two scenarios: born-in-the cloud applications, which includes support for multiple programming languages and runtimes, and Microsoft Cloud Platform infrastructure, which includes “support for compute clusters running Hyper-V and storage clusters running Scale-out File Server.”
I believe this focus is where Nano Server may see the light of day where Server Core didn’t. Granted, those two scenarios define a narrow place for Nano Server to play in, but we’ll see how that evolves going forward.
Microsoft kicked up a lot of hoopla around Nano Server at last year’s Build conference, with folks like Snover calling it “the future of Windows Server.” And with Windows Server 2016 around the corner, we’ll be able to make a more realistic assessment of that future. In the meantime, the focus on cloud-based apps gives Nano Server a unique selling point this go-round. The bits aren’t here, so we may have to revisit it again next year to see where Nano Server might lead.