Clear Linux Outlines How You Can Build Your Own Linux Distro In 10 Minutes

[phoronix]

Phoronix: Clear Linux Outlines How You Can Build Your Own Linux Distro In 10 Minutes

While Intel's Clear Linux is known to the most of you for its speed, it's also a distribution that is very easy to build off of for specific use-cases should you want your own pre-configured Linux OS...

http://www.phoronix.com/scan.php?page=news_item&px=Clear-Linux-Mixer-Guide

[FireBurn]

It's not really another distro though is it? It's just your own personalised flavour of their distro - which is great

[Vistaus]

It's not really another distro though is it? It's just your own personalised flavour of their distro - which is great

In that case Ubuntu is just a personalized flavor of Debian and Manjaro a personalized flavor of Arch.......

[smotad]

Does it compile with whatever secret sauce / parameters they use to obtain faster binaries?

[andyprough]

Oh how funny, Clear is pushing home-brewed RPMs as the solution to their lack of packages. With all the time it would take to build all your own custom RPMs and get your own repo up and running, hard to see how this distro would end up being a time-saver.

[Teggs]

Someone at AMD should take advantage of this immediately. Then launch their own resulting 'distro' on about... oh... April 1.

[ermo]

I'm surprised that noone here seems to be getting why this is actually a pretty big deal:

If you run a datacenter, your bottom line depends on efficiency. Efficiency implies getting as much performance as possible for as little energy input as possible. This in turn implies using the capabilities of your system to its fullest. Hang on to that thought.

Traditionally, distros in the x86_64 space are compiled by an ISV (RedHat/Canonical for instance) and distributed with the intent of supporting the least common denominator hardware in order to maximise compatibility. In a cloud setting, this is not exactly a recipe for maximum efficiency.

Now, with the move to containers for everything in the Cloud, the datacenter DevOps staff can request intel hardware *and* get a container-native distro that is *explicitly* designed by intel's own engineers to perform as well as possible on whatever intel hardware they are buying**. Just stop for a second and think what that means for the bottom line of this hypothetical datacenter. The initial cost of creating your container images are quickly amortized when you have hundreds of servers in your datacenter, each with tens to hundreds of customers paying for your services.

And then consider where it leaves everyone else who is trying to create a business of selling support contracts to datacenters? Yeah, that's a pretty bleak outlook if you're the Canonicals/SuSEs etc. of this world. You literally won't stand a chance in a market that's as highly competitive as the Cloud market is.

This has the potential to literally break the software model that has been pervasive since Microsoft started packaging software and selling it separately from hardware in the mid eighties. Call it a return to the UNIX times of old when a computer came with source code and compilers optimised for the hardware you were buying.

And the Cloud (not the client) is where the business revenue is these days. If you don't believe me, take a look at Microsoft's, RedHat's and Canonical's financials and see for yourself. It's no coincidence that IBM bought RedHat: It puts them in a position similar to Clear Linux for their in-house POWER architecture in the Cloud.

... or I could just be talking out my arse. We'll see I guess? I'm just wondering who AMD will team up with.

**: Be it new or used / off-lease with an older-gen intel architecture

EDIT: Tried to clarify that I specifically meant intel hardware.

[Mercyful Fate]

I'm surprised that noone here seems to be getting why this is actually a pretty big deal:

Well, from my perspective, you forgot one big bullet point:

• An Intel distribution tailored specifically for Intel's CPUs*.

This may be the tipping point where chip makers must begin supplying their own Linux distribution rather than relying on the efforts of others or lose out in market-share. Intel could well be on it's way to become the next IBM: The one stop shop for both hardware and software - Intel's.

I would also presume that in the long-term it's not off of Intel's radar to eventually compete with AWS data centers and others in cloud offerings. Edit - I see you also edited your post so here I also agree with you.



*Not that I have anything against a company that uses Linux since I personally prefer and use LinTel rather than WinTel.

[ermo]

Well, from my perspective, you forgot one big bullet point:

• An Intel distribution tailored specifically for Intel's CPUs*.

This may be the tipping point where chip makers must begin supplying their own Linux distribution rather than relying on the efforts of others or lose out in market-share. Intel could well be on it's way to become the next IBM: The one stop shop for both hardware and software - Intel's.

I would also presume that in the long-term it's not off of Intel's radar to eventually compete with AWS data centers and others in cloud offerings. Edit - I see you also edited your post so here I also agree with you.



*Not that I have anything against a company that uses Linux since I personally prefer and use LinTel rather than WinTel.

I guess I didn't make that part obvious enough in my post? It was definitely intended to be the core point:

(...) the datacenter DevOps staff can request intel hardware *and* get a container-native distro that is *explicitly* designed by intel's own engineers to perform as well as possible on whatever hardware they are buying. (...)

The "whatever hardware" was implied to be intel -- be it new or used / off lease.