SUSE/openSUSE Developing "Adaptable Linux Platform" For Next-Gen SUSE Linux Enterprise

[Charlie68]

I don't understand ... we don't know anything yet, it's hard to say something from those few words. Containers in SUSE are already widely used. Then there are those who still haven't understood, that in the last 10 years the software has changed, it has changed in terms of releases, now much more frequent, it has changed in terms of security where today there is more attention.
An LTS distribution as a fix release is designed, freezes all packages at some point in development, releasing only security fixes, this policy made sense 10 years ago, not today. Because ?
Because the basic concept was "don't touch it until it works", but today with hundreds of security fixes that must be backports, this paradigm, has gone to that country, the software is touched and this often generates weird bugs that software "original" does not have. Hence the need for all enterprise distributions or not, to find a solution to these problems.
Canonical is moving in the direction of application snaps, which are containers, all enterprise LTS distributions are looking for solutions, because now all these fixes heavily modify the distribution and are becoming unmanageable.
Not to mention the fixes that do not arrive (see Debian's situation with browsers a while ago), whereby a lot of software that cannot be fixed remains vulnerable, which is unacceptable for enterprise distributions.

[RahulSundaram]

immutable filesystems in general I find to be a nightmare to work with unless your workload fits nicely into that container.

https://containertoolbx.org/ is used in Silverblue and it should be usable in other distros as well. Makes working with containers much nicer. It doesn't solve all of the problems here but goes a long way.

[pWe00Iri3e7Z9lHOX2Qx]

Host OS and container/VM based userland? I love VMs, but this just sounds like a hassle.

immutable filesystems in general I find to be a nightmare to work with unless your workload fits nicely into that container.

It's a shame, because I think most of the needed pieces exist, but there are still a lot of rough edges. I can even imagine something extreme borrowing ideas from Qubes, just really slickly done. A very lightweight host which you don't even really interact with directly. Happily plowing through your daily tasks in your "main" container / vm without even really knowing it is a separate thing. You click a link in an email, if it's an external (untrusted) domain, it opens in a browser spawned in some throwaway container / vm (maybe with a highlighted border around the window). The experience is seamless no matter what you are really interacting with under the covers.

The hardware is getting fast enough. The software pieces are starting to fall into place. But there's a crap ton of polish that would need to happen. And since my dream is about the desktop, most companies won't care since there isn't obvious money it.

[pWe00Iri3e7Z9lHOX2Qx]

Meanwhile Canonical will continue with their usual LTS paradigm established over a decade ago and still be far more successful.

Reason being that there is simply no enterprise/community split.

It's easy to assume that Ubuntu's relative popularity on the (free) Linux desktop means that Canonical is very successful in the (paid) enterprise, but that isn't the case. SUSE is much larger than Canonical.

In terms of commercial success, Canonical < SUSE < Red Hat, and the jumps between those are all large. Red Hat is enormously successful compared to even SUSE. Canonical's revenue is a drop in the bucket in comparison.

[bobbie424242]

You can bet this will be something based on Micros OS for immutable base OS + FlatPak for consumer apps with maybe a dose of docker/podman containers for server software.

[RahulSundaram]

Meanwhile Canonical will continue with their usual LTS paradigm established over a decade ago and still be far more successful.

Reason being that there is simply no enterprise/community split.

SUSE marketshare has always been much higher than Canonical and therefore continues to be more successful. Also, LTS vs non LTS is enterprise/community split with different branding, not that, it is a bad thing. Enterprises just have different priorities.

[dekernel]

I really hope they don't completely move over to a rolling release methodology because if so, I might have to head over to the Ubuntu LTS which I REALLY don't want because openSUSE Leap has been great for me. I am not interested in getting the latest kernel and such on my workstation. I am paid to produce software for our customers, not debug components of an OS.

[Linuxxx]

Meanwhile Canonical will continue with their usual LTS paradigm established over a decade ago and still be far more successful.

Reason being that there is simply no enterprise/community split.

[pWe00Iri3e7Z9lHOX2Qx]

It sounds like an immutable desktop like Fedora Silverblue, Steam OS, etc...

They already have SLE Micro, MicroOS, etc.

[drake23]

And 10 years from now, the current paradigm will be remarketed as "ultra flexibility to change the host on the fly" as opposed to immutable distributions...

Currently, OpenSuse 15.3 is my all time favourite, but if 16 has container based Userland, I'll switch to FreeBSD :/