Despite the few “DSC Planning Updates” written by Michael Greene (Program Manager for DSC, and DSC Community Committee member), there are still misunderstandings by the community with regards to the directions Microsoft is taking for DSC and its ecosystem.
Here, we’ll try to explain what we understand from the DSC Community.
Microsoft’s DSC Team, about a year and an half ago, was working on a new version of the LCM component of the DSC Platform.
They were thinking they would open-source it shortly after, as you can judge from the DSC Planning update of the time, or sessions delivered ar events by the teams. They labelled, confusingly, the work a new DSC, or DSC Core, before rectifying some of this communication.
Anyway, around June 2018, the medium-term strategy changed, and their work was to be used and developed further for new Azure Services. It’d be a great mistake to be oblivious of Microsoft’s focus to deliver a great cloud where customers are successful.
This new commitments, and having the new Native LCM in production in Azure, meant the teams would not have the resource to commit to open sourcing the engine, a lesson learnt from the DSC Resource Kit and PowerShell team.
Moreover, the solution they developed was primarily focused on audits,
as required by the enabled services: Change Tracking, Inventory, Guest Configuration,
and Update Management, so not a potential replacement for DSC until it
implements the Set functionality (from DSC’s Get, Test, Set).
So in terms of planning, which now usually spans across a 6 months window, open sourcing the work now would be premature. The engine would not be doing enough for the DSC community, and managing a public project while having to support the Azure solutions would stretch thin the time working on for further improvements.
The plan is clear, the focus for now is to focus on delivering the Set and making
sure the quality is up to the task of managing cloud-scale environments,
as required by the demanding Azure services.
Doing it this way, on Azure first, has some great value.
It:
Set
is implemented.With this said, let’s be clear.
There are no plans for DSCv2, DSC vNext or DSC Core.
Plans are only made for a 6 months window (roughly), and its not in there, as far as I know.
Now that’s said loud and clear, there’s definitely a team working on DSC and investments are made on the technology. Azure Policy Guest Configuration relies on DSC, for its DSL, Resources and principle to audit Virtual Machines in Azure, or as long as they’re managed in Azure (Hybrid)…
The team do wish the DSC community will continue to grow, and that there will be increasing usage and demand for open sourcing the LCM they’re working on. They hope the demand will mostly align with Microsoft’s mission to make companies more successful by transitioning towards a cloud model.
The future is bright, and will see changes happening on different time frames.
Investment has been, and is still being made, and we’re just starting to see the results.
Azure Guest Configuration supports custom content, using the DSC DSL to create
custom audits (now Public Preview).
It’s a clear sign that leveraging the DSC ecosystem is still a reality for new services.
The DSC Team is investing in this DSC Community, supporting its growth, and
help with its evolution.
They’re actively supporting more openness, transparency, and community
involvement in the DSC Resource Kit to find a more sustainable path that helps
the users (released now).
The PowerShell and DSC team are planning
to release a cross platform Invoke-DscResource command that would allow to
leverage PowerShell DSC Resources within scripts, running in current context,
without the burden and limitations of running through the LCM as found in
WMF5.1 (in the next few months).
And in the slightly longer term, the DSC team aims to implemented the Set for
their new LCM, and start using it in some Azure services (~6 months+).