Windows 10 servicing branches – CBB or LTSB? A discussion
By James Rankin | 21st February 2017
I penned an article recently over on the AppSense (now Ivanti, of course) blog which went into a discussion of the relative merits of Windows 10’s two main servicing branches, Current Branch for Business (CBB), and Long-Term Servicing Branch (LTSB). After publishing this, I seemed to attract comments from some people on Twitter along the lines of “you’re doing it wrong, no-one should use LTSB”. The first time this happened I asked the commenter if they’d mind taking some questions from me about this to justify their stance, but I never received a reply from them. So when I was challenged again about it, I was a little bit snappy about it and invited the commenters to a discussion around it. In this instance, they were as good as their word, so we had a good discussion on this subject last night.
It’s a shame we can’t get anyone from Microsoft directly to join in on these things, but Aaron and Simon in particular are very clued-in Microsoft guys so they gave us some excellent insight. But I don’t know what it says about Microsoft that they are so difficult to get community engagement from in this way. As a company they’re an order of magnitude bigger than the likes of Citrix and VMware, so it might be unfair to expect them to respond in the same way those two would, but I think the reason they won’t engage in this way is that either a) no-one can really answer the questions we’re asking, much like the problem you have when talking licensing, or b) no-one wants to take responsibility for putting an official line on something that could change pretty quickly. That could be just me being cynical, but Microsoft were the first of the tech companies to move towards a solid community program, and now I feel they’ve let it slide somewhat.
Anyways – what were the key takeaways from our discussion? I’ve embedded a video of the discussion within the article (don’t ask me why I did video rather than audio – I will know for next time), but here is a brief summary of the most pertinent points that were made. On the discussion we had myself, Aaron Parker (@stealthpuppy), Simon Binder (@bindertech), Rene Bigler (@dready73) and James O’Regan (@jamesoregan), and we had a bunch of questions both from ourselves and some supplied by people on Twitter around the whole “CBB versus LTSB” debate.
- Windows 10 users should aim to adopt CBB where possible, even if ultimately it isn’t possible for all devices
- Windows 10 brings a new servicing model and people should aim to break away from the old, rather than continuing to do things the same way
- Testing needs to be done better across the board, but there is a resource undertaking involved
- As Windows 10 becomes more widely adopted, vendors will have to accelerate their own remediation and response – essentially, third-party vendors should step up to the Windows 10 release program
- IT departments may have to evolve and run up a learning curve to adapt to Windows 10
- Cloud solutions bring big changes to IT management anyway, so Windows 10 is not really markedly different
- Application and user virtualization can be a big help in Windows 10 environments, but can also expand the device estates
- Smaller businesses may be more likely to use managed services to help with virtualization and Windows 10 adoption
- Universal Apps platform may replace a lot of legacy desktop apps, ones that aren’t can be pushed into a cloud-hosted platform such as XenApp Essentials
- Microsoft isn’t really keen for you to use VDI
- Windows 10 is not designed for non-persistent environments
- LTSB releases will be at least two years apart to differentiate it heavily from CBB
- Controls of Modern Apps and the user environment need to improve
- Corporate branding controls should be improved and easier to find for Enterprise customers
- Microsoft’s advice on CBB definition (that any machine running Office should be considered CBB) should be ignored, this decision should be taken on a per-customer basis as every environment is different
- Upgrade Analytics is possibly a very good idea for fine-tuning Windows 10 deployments, maybe 10% of your estate should be on Current Branch fast ring, with as much application coverage as possible
- On Enterprise Edition, you could potentially deploy a CBB release and leave it in place indefinitely
- We’ve agreed that 8-14 months is a good ballpark figure for the total CBB upgrade window, possibly with some leeway on either side
- Desktop Windows 10 evolution may slow down, whereas other areas of Windows 10 may accelerate, but software evolution will be unlikely to stop
- There may well be a disparity between XenApp desktops on Server 2016 and Windows 10 VDI or fat clients, which may require additional training
- Windows 10 releases can be configured for embedded-type endpoints
- LTSB has mileage for particular verticals and application types
- Azure adopters have been faced with Current Branch anyway for their datacenters, so CBB is not so much of a jump
- Windows 7 and Windows 8 patching changes are aimed at preparing customers for the Windows 10 model
The recording of the session is here below:-
So what conclusion did we reach? I guess in the main, we made the point that CBB should be your goal, but for endpoints that for whatever reason do not meet this criteria, then you can still apply LTSB and maintain a consistency within your environment. I think that is possibly the major point – you should never start from a perspective of aiming for LTSB, but if you need it, then it is always an option.
I hope the video is useful to those of you out there in the community thinking of Windows 10 adoption, please feel free to post questions and comments either here or on the video itself. And a big thanks to Aaron, Simon, Rene and James for giving up some of their time to contribute to this – it is always much appreciated!