It's time to split Linux distros in two

[nedo]

Desktop workloads and server workloads have different needs. Why address them in the same distribution?

For decades, Microsoft has released completely separate operating systems for desktops and servers. They certainly share plenty of code, but you cannot turn a Windows 7 system into a Windows Server 2008 RC2 system simply installing a few packages and uninstalling others. The desktop and the server are completely different, and they are treated as such across the board.

Naturally, that hasn't stopped more than a few folks from making the questionable decision to place server workloads on Windows XP systems, but by and large, there's no mistaking the two. This is not so in Linux.

You can take a Linux installation of nearly any distribution and turn it into a server, then back into a workstation by installing and uninstalling various packages. The OS core remains the same, and the stability and performance will be roughly the same, assuming you tune they system along the way. Those two workloads are very different, however, and as computing power continues to increase, the workloads are diverging even more.

Maybe it's time Linux is split in two. I suggested this possibility last week when discussing systemd (or that FreeBSD could see higher server adoption), but it's more than systemd coming into play here. It's from the bootloader all the way up. The more we see Linux distributions trying to offer chimera-like operating systems that can be a server or a desktop at a whim, the more we tend to see the dilution of both. You can run stock Debian Jessie on your laptop or on a 64-way server. Does it not make sense to concentrate all efforts on one or the other?

If we're homogenizing our distributions by saddling them all with systemd, then there's very little distinction between them other than the package manager and the various file system layouts. Regardless of the big gamble of pursuing desktop Linux as a line of business, would it not make sense for several Linux distributions to focus solely on the desktop while others focus solely on the server? Sure, Ubuntu and others offer "server" and "desktop" versions, or different options at install time, but in reality the only differences are the packages installed. On many distros today, even the kernel is the same; it's been merged.

With the release of popular gaming framework Steam on Linux, we're starting to see some traction for desktop Linux among folks interested in computer gaming and computers in general. They are at least trying Linux on the desktop more than they might have before, and they are finding some success.

However, they're also demanding better performance for desktop-centric workloads, specifically in the graphics department and in singular application processing workloads with limited disk and network I/O, rather than the high-I/O, highly threaded workloads you find with servers. If Linux on the desktop has any real chance of gaining more than this limited share, those demands will need to be met and exceeded on a consistent basis.

Add to that the need for all kinds of hardware support, peripheral support, power management, and other mostly desktop considerations, and the desktop and server distros drift even further apart. Also, I'd wager that there are 10 or 100 times more Linux server systems running on virtual machines than on desktops. That is a completely different scenario that should be accounted for when tailoring a distribution.

Can Linux do all of these things? Sure. Can we stop trying to make every Linux distribution capable of supporting all of these use cases out of the box? That's a very real possibility. There are already desktop-centric distributions like Mint, as well as more server-centric distributions like Gentoo and Debian to some degree (at least before systemd). They aren't full-on in either direction, but they definitely lean one way or the other. I'd be hard-pressed to consider RHEL 7 a truly server-centric distribution, given the use of systemd and the inclusion of desktop packages, but it's not really a desktop system either. It's middle-of-the-road in many regards.

There is enough pushback to systemd to warrant a fork of a major distribution that excises systemd and the GNOME dependencies, while providing a more traditional and stable server platform that has no hint of desktop support. No time need be wasted managing the hundreds upon hundreds of desktop packages present in the distro tree, no need to include massive numbers of desktop peripheral and graphical drivers (RHEL 6.3 ships with 57 xorg drivers, for instance).

There's also the matter of security. The security concerns for a desktop are vastly different than those for a server -- and server security concerns are vastly different among servers, depending on what each server is doing. However, it's safe to say that protecting against malware delivered by clicking through a malicious Web page is not high on the list of possible threats for a Memcached server.

I can clearly see the desire to improve the desktop Linux experience in terms of peripheral hardware support, graphics performance, sound, boot times, and ease of maintenance and management. Those are desktop concerns in a desktop distribution, and if ripping and replacing the plumbing helps to achieve the goals, then there may be some merit. However, there's no reason those same concerns should result in a rip and replace of the plumbing on server-class systems. It's shortsighted and dangerous.

Dedicated and tuned server distributions are a good idea anyway, systemd or not, but if that's the catalyst for the creation of a major, mainstream server-only Linux distribution that remains based on the Unix philosophies that have served us amazingly well over the past 45 years, then maybe all of this heated debate about systemd is not wasted after all.

Sursa articol aici

[Guest]

Poate nu vad eu bine, insa in tot postul asta domina ideea de "Microsoft is better" si de "we should go enterprise with linux".

Microsoft-ul face degeaba cate doua sisteme (enterprise/end-user) daca sunt tot o apa s-un pamant.

Marketing peste marketing.

[nedo]

Nu spun ca microshit e mai bun. Spun ca abordarea lor e mai logica, un sistem pentru desktop, unul pentru servere.

[PingLord]

Nedo, ai recompilat un kernel de Linux ?

Poti sa ii faci shifting de acolo foarte frumos din Desktop in Server.Ai preemption model la CPU , alocatorii de memorie , etc. Deasemenea ai distro-uri ce sunt folosite pentru partea de server si distro-uri ce sunt folosite pentru partea de desktop.