Stuipd idea using Windows 10 LTSC or tinkering with V1607

[German]Today I like to discuss within a blog post the question, how reasonable is it, to use a Windows 10 LTSC versions. And I also take a look at the idea of using Windows 10 version 1607 as a poor mans LTSC version.

The initial idea for this article raided from my German blog post Windows 10, geplante Obsoleszenz? Within this article I pointed out some implications of Windows as a service. Furthermore, some blog readers asked why I 'm not mentions Windows 10 LTSC often. Others asked about a 'home made' slimmed-down Windows 10 version wit blocked feature updates and long term support.

On some days, however, the pieces of a puzzle come to light and forces suddenly an image. I'm discussing with some people who are active as administrators in the Windows 10 environment to benefit from their experiences.

Crazy idea Windows Long Term Servicing Channel (LTSC)

I confess, I have occasionally thought about setting up and working with a Windows 10 Enterprise LTSC here. No more feature updates and a operating system base that gets 10 years of support. But Microsoft has made it always clear that Windows 10 LTSC is intended for 'mission-critical' applications and not for day-to-day operation in office environments. 

Within the last hours I became aware (via Twitter) of the Microsoft blog entry Say No to Long Term Servicing Channel (LTSC). British consultant Greg Nottage explains why the LTSC of Windows 10 is not a solution for corporate use. And there are a number of disadvantages that users face with an LTSC solution compared to the SAC (Semi-Annual Channel) approach.

• No new operating system extensions via feature update
• No new security features via feature update – LTSC does not keep pace with the feature enhancements of Windows Defender ATP.
  
• No support for newer hardware – keyword Windows 10 LTSC 2016 does not support Intel chipsets beyond Kaby Lake
• No Windows Analytics Upgrade Readiness support for LTSC
• No support for the Edge browser
• No support for Cortana
• No support for Windows Store
• No support for surface hardware
• LTSC does not support ConfigMgr Express Updates

Nottage wrote that in-place upgrade from Windows 7 to Windows 10 is not supported for LTSC. Also: Microsoft Office 365 on LTSC will no longer be supported from January 2020. How silly. Furthermore, he claims missing support and restrictions in the LTSC area by independent hardware and software manufacturers. And non-security operating system fixes and extensions may not be back ported to LTSC by Microsoft. Loosely defined LTSC release cycles should also make advance planning more difficult. Some people will see it differently, and some arguments a bit silly!

Windows 10 V1607: Poor man's LTSC?

I still had since long time the idea of installing Windows 10 Anniversary Update (V1607), using DISM or tools to remove Store, Cortana, Telemetry and so on. The background for this idea was the fact, that Microsoft promised to support Windows 10 V1607 until 2023 with security updates (see my blog post Windows 10 support for Clover Trail machines till 2023).

Because Microsoft cannot provide Clover Trail support from Windows 10 V1703 upward, security updates are granted until 2013 for Windows 10 V1607. But I didn't know, whether Microsoft checks for Clover Trail chipset during update install. This comment confirmed, that updates still install on officially unsupported Windows 10 Anniversary Update. The user downloaded the June KB4287903 flash update manually and was able to install it on Windows 10 Pro V1607. Nevertheless, it is a crazy idea to use Windows 10 V1607 as a basis, as you will see below.

Looking behind the curtain

Let's go one step further now – I'm not a consultant for Microsoft, so I'm not suspicious of wanting to sell Windows as a service (WaaS). Within the last weeks I received several comments within my blogs. And I had an exchange of ideas with blog reader Karl about the idea of the Windows 10 V1607 LTSC (also the homemade version). So I can lift the curtain of the stage a little and show the dirty floor behind the scenes.

LTSC, just a dead horse …

German blog reader Christian wrote this comment (just scroll down a bit, up to his text) an aspect I wasn't so familiar with. His message: The LTSC is not intended for office computers. Why?

Quote: I now have software here that requires at least Windows 10 – 1607 if you are using Win 10. With Win 7 and 8.1 the software does not make any restrictions. It is a special software, but the problem is not the software itself but in the background and there it can affect any program..

Christian also provides the explanation, which shows that Microsoft can (and probably will) create a rocky road to users who are installing LTSC builds.

Reason: All Win 10 versions below version 1607 cannot install Net Framework 4.7, which automatically results in an LTSB version below version 1607, e.g. 1511, will be cut off from using current software.

For example, paint.net also requires Net Framework 4.7 in its current versions. And if also the provider of the own standard office software takes this step you are quite nailed with your old LTSB version.

Win 7 and 8.1 do not have this problem, Net Framework 4.7 is still fully supported.

And now .NET Framework 4.8 is already in the pipeline. As soon as Microsoft realizes that people don't switch to the most recent Windows 10 builds, you fastens the thumbscrews.

A practical experience

I discuss behind the scenes with some administrators. One, I will call him Karl, just gave up his administrative job some year ago and is now an IT service guy. When I asked him for his opinion about using LTSC or a customized Windows 10 V1607, he came up with a clear statement.

I personally don't believe in LTSC or in cobbling something like Windows 10 V1607 together. What you should ideally do: Install Windows 10 1803 and then use DISM GUI to uninstall the pre-provisioned apps. Or switch off telemetry with local GPOs.

Karl justifies his rejection with the following practical experience, which I cannot reject nor confirm.

Windows 10 V1507, V1511 and V1607 in particular are anything but well designed versions under the hood, which is why I am also not recommending Windows Server 2016 V1607. 

We have 2 well-known customers who have taken this stony path with LTSC and now it is over. All of us wasted money, me included. All clients are now reinstalled and relicensed with Windows 10 Enterprise.

Then he adds the following explanations (that fits with the british consultant's comments cited above).

Reason 1: With the end of mainstream support, the support claim of the same LTSC build against the OEM manufacturer regarding drivers and security updates expired! But it's absurd.

Reason 1a: The LTSC does not allow code changes and is therefore out.

Karl notes also that the OEMs do not live an LTSC ideology. He referred to a discussion with me on Google+, where I wrote, OEMs often make little or no effort to test the insiders and see what happens'. I know this argumentation from Microsoft (see Moving from project to process: digital transformation with Windows as a service, from John Wilcox, Microsoft). That was also the point where Karl went into the discussion and we couldn't find a common sense in the discussion – because Microsoft creates a moving target with Windows 10 semi annual channel (SA)C, which is fundamentally patched one day before the release. And then Karl gave me further reasons for his argument against LTSC:

Reason 2: The updates provided for LTSC are pure security updates, bugs stay in forever if they are not security critical. This is part of the concept.

Reason 3: An LTSC version cannot be upgraded from LTSC to current non-LTSC versions, not even in-place because of other codebase / WIM images .

Reason 4: An LTSC is more expensive to license than other models in the long run. An LTSC is revised approximately every 3 years (LTSC 2015 – 1507,  LTSC 2016 1607,  LTSC 2019 – probably 1809) 

Reason 5: Microsoft does not support Usual Business Use on LTSC, i.e. Office or Office 365 etc. The applications are therefore not road warriors or laptop users, but rather machine controls, telephone systems, devices in the operating room, etc. Devices with little to no changes and high uptime.

Karl argues that the current way with 6 months of upgrades (or if delayed, with 18 months of upgrade support) is the only way to prevent fragmentation like XP, Windows 8 or Windows 7. He told me: By the way, nobody had gotten upset about it at Apple, it had been there since MacOS X.

Note from my side: With macOS you have a closed hardware environment, and still suffer from bugs, but can fix them faster than Microsoft. And Apple recently set itself the internal goal of slowing down the development of new features for higher quality and fewer bugs. Besides (but I'm not 100% in there), it seems to me that the number of collateral damage due to incompatible hardware and software is quite limited with macOS. Eventually a Mac series didn't made it to a new macOS version, but the software can often be used again. But I may be wrong.

So, this has once again become a much too long blog post. However, I think and hope that the aspects mentioned here will lead to a critical and fruitful discussion or new insights. At least now everyone should realize where the journey with Windows should go. I don't envy administrators in the Windows environment for their job.

PS: Finally, a reference to the ZDNet.com article by Mary Foley Microsoft looks to add security, stability with Windows 10 IoT Core Services, where she already outlines in the header text: Microsoft is adding a new, paid option to its Windows 10 IoT Core operating system that will prevent it from getting regular feature updates, among other features.

Similar articles:
Windows 10 support for Clover Trail machines till 2023
Windows 10: End of Life for several builds