[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#727708: non-technical reasons



Obviously, then making technical decision one must realize the field
we must work in near future. So we need to make at least
draft analysis of the current situation.

Strongest aspects of Debian that outperforms every other distro:

1) It's universal operating system. In minimal installation it is stays
really small in both disk and memory requirements.

2) It's really fast and effective.

3) It fully controllable by moderate specialist. It does exactly and
only what you want and then you want.

4) Debian core is perfect at its stability. Non-core packet is quite
stable too.

5) Therefore, it is a best base to build effective, robust and
really secure systems on it.

6) It's clearly the best system for small-sized
systems such as web-servers:
http://w3techs.com/technologies/details/os-linux/all/all

6) I believe Debian and its huge and thoroughly tested packet base is the
best starting point for complicated non-standard setups attuned for
individual customer needs.

7) It is rather conservative in tools of managing of the system. This
helps to teach new things and to move further than study how to do the
same things in new system every couple of years.

8) It's best of all distros in freedom of choices how you want things
to be done. At least for skilled people.

The weaknesses of Debian:

1) Lack of full-time developers of distro itself.
We cannot change big things fast. We cannot change things frequently.

2) We don't have much major upstream developers of core Linux components.
We cannot influence driving path of Linux itself much.

3) Lack of paid "official" commonly-trusted technical support with
close connection with major Debian developers.
Customers don't have a telephone number which they can call and all
things magically "start to work".

What happens in Linux world today:

Red Hat is commercial company. They are making money. To make money
they must have high prices. To have prices much higher than real cost
of service a company must be alone on the market segment.
But last years we see trends as Red Hat is have been slowly but
continuously squeezed out from the leader in Linux support and
distribution.

I think it happens due to 2 tendencies:

1) Aggressive marketing of Ubuntu. Many new Linux-people familiar with
Ubuntu, but not Red Hat family of distros.

2) Steady growing of number of high-class Linux specialists. It's easy now
to find really good professional to build custom system with almost any
degree of complexity. And these specialists mostly prefer Debian as it
more transparent, simple, flexible and small-component-driven than
Red Hat and family.
At the same time it's robust and fast at least as Red Hat if not more.

That powers have Red Hat to change the situation:

1) They have strong reputation as leader of Open Source. Many years they
helped to make Linux to become best operating system ever.

2) They have great influence on many key developers of Linux ecosystem.

3) They have outstanding respectfulness and trust credit from
Open Source community.

What Red Had can do:

1) They can change Linux kernel developing scheme so that Red Hat kernel
will somehow has features that not so fast spread to other distros.
In fact they did so in limited form, but Linux kernel already so feature
rich that bleeding-edge kernel features not any critical to most of
users.

2) They can circumvent the GNU/Linux infrastructure so they can fully
control it and any concurrent must be in the wake of Red Had innovations.
Automatically, every distro become either Red Had clone in many aspects or
become geek-only. - I believe this is that happening now.

They developing and hard pushing a product that changes all the way the
system works. The product absorbs more and more system services from
booting and starting services to even user authorization and logging.
Clearly, it makes the system less robust and less secure:
http://ewontfix.com/14/

This is really bad for everybody except Red Hat. They have all systemd
specialists in the company. If you have RH support - you welcome, all you
problems with systemd will be solved fast. If you have non-supported
configuration - just pay more.

If you have some other distro and even non-standard configuration,
you're on your own. I mean all of Debian users.
We are always welcome to patch systemd and send patches to a company,
that created trouble for us.

Yes, probably 95% will never face the problems that somebody already
not solved. But I'm talking about specialists that making complicated
non-standard systems out of Debian.

Even on small and simple web-servers systemd and it additional services
will cause growing of memory footprint of the system, kicking Debian
out of low-end VPS'es exactly as CentOS now. Yes, providers will offer
VPS'es with more memory, but Debian will lose one of its big advantages
over Red Had family.

At any rate, if systemd continues to absorb system level tasks,
diversity between distributives become only cosmetic. But why need
other distributives with minor differencies if we already have the
leader among them: Red Hat?

In such a situation Debian probably will cease to exist.
Or it will try to make really different distro almost from scratch.
But it will be much harder to make it from scratch instead of
trying to not loose control over ecosystem now.

I believe, choice systemd as default init system will slowly make
Debian one of many and eventually destroy it.

Regards,
J


______________________________________________________
powered by Secure-Mail.biz - anonymous and secure e-mail accounts.


Reply to: