Windows Server Servicing

Welcome to part 2 of Windows Server Servicing. Now you’ve got the background, let’s look at the choices we have and why what we’re doing at the moment is okay.

Semi-Annual Channel servicing sounds great…

Well yes, it has its place. But as I said at the end of part 1, this isn’t a shoo-in and there are some considerations before changing update policy.

Some examples:

  • Are the new features of any benefit to the role your server is there for? If not, you’re potentially introducing an additional management overhead and the need to plan for extended maintenance windows every 6 months, for no real benefit. And the updates won’t necessarily align with the most ideal time (e.g. half-term) to take them. There will likely be some control over when the Feature update applies but it remains to be seen whether there will be sufficient control for your environment.

    Microsoft’s view of the Semi-Annual Channel is that it’s for those who are moving at a ‘cloud cadence’. There’s an almost inherent assumption built in therefore, that the number and role of those servers means that they can be regularly updated on a rolling basis – that there are enough of them to take some out of service to be updated while still maintaining a desired level of service to users. That assumption clearly won’t stand for everyone.

  • Is your network well maintained, with patching and updating under control? If it isn’t, you probably have bigger problems than choosing to adopt a new servicing approach for Windows Server, and if you do implement semi-annual servicing those problems could get worse. Each Feature update released into the Semi-Annual Channel will be serviced for 18 months, just like Windows 10. If patching isn’t under control it’s possible that Windows Server could reach end of servicing, and that could lead to a weakening of security – as new threats emerge but new quality updates are not released - and in turn a myriad of other risks and issues.

  • Is your server hardware likely to be suitable as the server operating system evolves and new features demand more resources? With the traditional model (akin to LTSC), it’s not uncommon to refresh server hardware at the point the move is made to the latest server operating system. The reality of funding also means that schools are having to squeeze the most life out of their assets, and so the choice might be taken to start Windows Server Semi-Annual servicing on old hardware. That could result in something of a catch-22 further downstream, where at some stage in the cycle you have the ‘choice’ of upgrading to the next release and having degraded performance or not (be able to) upgrade and eventually drop out of a serviced version.

  • Do you have volume licensing with Software Assurance for Windows Server? Initially at least, the most suitable licensing that entitles you to Semi-Annual Channel serviced Windows Server is volume licensing with Software Assurance. There is also the option to deploy into IaaS from the Azure Marketplace. The choice to adopt Semi-Annual servicing is not available to everyone.

And then there’s the type of server installation…

This one deserves a section all of its own. The Semi-Annual Channel (Windows Server) is only available to deployments of Nano Server, which has become a lot less relevant to our deployments from v1709 onwards, and Server Core. All of our standard server deployments, with very few exceptions, for both CC4 and non-CC4 (‘vanilla’) are currently Server with Desktop.

So right now, we couldn’t easily adopt Semi-Annual servicing for Windows Server even if we wanted to, without introducing new / replacement servers.

You may be aware that we have on occasion deployed Server Core (e.g. CC4 Matrix). There are reasons why it hasn’t resonated with school network managers (e.g. no graphical user interface to manage it, forcing management through PowerShell), but there are changes afoot in the server management space which means we should revisit that – but that’s a subject for another blog.

Okay, so let’s stay with LTSC then?

Well, that is an option. But remember what I said before about making sure it’s right for you? Here are a few things to think about:

  • What role(s) is your server performing? While it’s practically impossible for us to know what features will come with future releases of Windows Server, it is likely that typical features will align to certain Windows Server roles. Remember Microsoft’s view that Semi-Annual servicing is for those moving at a cloud-cadence? On that basis, it’s likely that Windows Server innovation will be focussed on roles to support that, such as Hyper-V and Remote Desktop Services (RDS). More traditional roles such as file and print may receive some attention, but it’s as likely that leveraging those improvements will require a shift in how we use those roles, as technology continues to develop towards cloud adoption.

  • Do you need the latest release to support the broader strategy of cloud adoption? It’s likely that you will want to deploy some new or updated features that are only available to Windows Server ‘as a Service’ in the timescales you need them in.

  • Can you wait 2-3 years for new server features? In time, you’ll become more and more used to regular change as other technologies also progress rapidly (e.g. Windows 10, Azure, SCCM), so may not want to wait 2-3 years for new features that, with different choices, could be available to you sooner.

Conclusion

There isn’t a one-size fits all solution here. Should you adopt semi-annual servicing (and therefore Windows Server ‘as a Service’) across the board? In the longer term, maybe, but in the short term almost certainly not. Should you stay with LTSC in all cases? That’s unlikely to be the best approach for every solution either.

It’s early stages for regular servicing of Windows Server, so as with Windows 10, which has been on this model for a while already there is a need to monitor it until it beds in.

There is certainly potential for taking advantage of the benefits of regular feature releases in more targeted deployments, such as Hyper-V and / or RDS. You should consider the broader impact, as well as the benefits of doing so as part of any introduction though; highlighting the advantages as well as managing transformation will be key to adoption of Server Core, on which this new approach depends.

It certainly isn’t a decision to take lightly. If you would like some advice or help in deciding which server versions best suit your needs, please contact us at networks@rm.com

Key Takeaways

  • Windows Server and Windows Server 2016 are two different things. The former will evolve more quickly than the latter. It’s not possible to upgrade from one to the other – a clean installation is required.
  • There is no ‘one size fits all’ solution here. You should deploy the right operating system for the right job. A lot of deployments will still suit the slower cadence of Windows Server 2016, and some may suit the faster cadence of Windows Server better.
  • Server Core may become more relevant than before.
  • Semi-Annual Channel servicing is only available to Server Core deployments.
  • Licensing is key to having the option to adopt Semi-Annual Channel servicing.

Check out our other Windows Server Servicing blog in this series at www.rm.com/blog/2017/november/windows-server-servicing

Post Comment

Blog Comments

back to top button Let's talk