Is Red Hat still relevant? You bet.

Scott,

I'd like to thank you for taking a step back, doing some primary source research, and then presenting a narrative that will live past the weekly news cycle.

As one of the folks mentioned in your article (by way of the Wikipedia entry), I am clearly an interested party. But I am also President of the Open Source Initiative, where I try really, really hard to get people to understand the whole picture of open source. I can tell you that it is frustrating when journalists are more concerned with horse-race headlines and juicy tit-for-tat between executives and less concerned with complex narratives that, despite their complexity, reveal powerful truths about the way our industry is changing. And I can tell you how rare it is to find /any/ journalist willing to do the kind of investigation and analysis that you have just demonstrated.

I am proud of what we have done, which is nothing less than to redefine the software industry and the competitive context of the technology industry in the 21st century (as exemplified by this PC World article). And it is clear that neither Red Hat nor the open source community could have reached this point so rapidly if we were as petty, naive, unskilled, unprofessional, and lacking in any business acumen as the conventional narrative would have anybody believe (at least according to the headlines of the conventional stories covering open source).

So, thank you for stating the obvious. It encourages me to keep up the good fight, both at the OSI and at Red Hat.

As an update to your article RHN/Satellite is now Open Source, check http://spacewalk.redhat.com, announced together with the Java Certification for the Free IcedTea project included by Fedora 9 that provides a FREE Java machine

One thing that Red Hat deserves respect for is staying true to their Free Software ideals. All of the software they have written (with a few exceptions like their Red Hat Network service) has been released under the GPL.

...and Red Hat released Red Hat Network Satellite as GPLv2 just last week:

http://spacewalk.redhat.com/

Lots of folks inside and outside Red Hat are glad to see this.

I believe RHN *was* the final product that was not open source... until last week.... now http://spacewalk.redhat.com will take you to the open source RHN.

Kudos to Red Hat for being truly committed to open and transparent computing.

Red Hat Network / Satellite is now free& open source. See http://spacewalk.redhat.com/

I have been a sys admin for many years and I can determine a person's skill level by how much bias they show for a particular OS, and how some confuse the desktop with the OS. Eventually you reach a plateau where it doesn't matter much. You might prefer a particular OS over another, or prefer a Linux distro over another, but that doesn't mean the other ones are bad. Putting something else down to support your opinion is a cop out.

I have used Red Hat for years and I have used other distros as well. For me it depends on what is running in the data center where I currently work. FreeBSD one place, SUSE another. I got Red Hat now and also use Fedora. I am quite happy with Fedora as a workstation and like GNOME. Grabbing the last remnants of the Netscape directory server from AOL was a good move, and the Fedora Directory project is thriving.

I am sure there are other distros out there that are cool, and if you find something you like, thats great! Its silly to fight over which distro is better. Celebrate x86 UNIX and support each other.

Dowdle, (hmmm no tab-complete) just think of canotical as an up-and-coming Red hat. Stay with me now, their business models are nearly identical. Here are the key differences: conotical is working with manufacturers to sell desktops, red hat works withs with manufacturers to sell servers. As to offerings, Red hat offers Enterprise and fedora (guiana pig) versions of their software and canotical does not. This is NOT a warm fuzzy decision to "give" to the community. This is to compete with rh/fedora. I personally purchased 3 versions of red hat prior to the split, the split angered a lot of people including me. Source packages =/= distribution. As far as Red hat's contributions to the linux kernel (linux as opposed to GNU), they aren't doing that for warm fuzzy feelings to "give" to the community either. They are doing that because they need certain things from the kernel that helps their business model. An IBM server with red hat on it has problems, red hat kernel hackers will send a patch to Linus and crew. GPL is live by the sword, die by the sword. As for canotical, I think everyone knows that the ubuntu african warm fuzzies are a farce. They (like red hat) are in it to make money. We can see evidence when you see them pairing up with Dell. When the kernel or gnome are no longer sufficient to keep Dell happy, canotical will have to start patching them. They will have to show an effort to those who are paying the bills, Micheal Dell. They can only do this if they start making some money. My short answer is: yes Red Hat is very relevant and important. So are other companies trying to make money from gpl'ed software, because we benefit from their (via gpl) business needs. After all, they have (hope to have) money to fix things. Big companies reigned in by their workers (software writers) through gpl is capitalism at its finest.

As for active directory killer, linux is a very long way away from it. There are a couple of key misconceptions that unix/linux admins seem to have. One, single-sign-on is all that active directory does. Single-sign-on is a extremely small part of what active directory does. Network wide security groups that have access to local resources is a very important part. These security groups configure the desktops, software access for individual, hardware access, software installs, software updates based on the the users group. These rights and services are replicateable on a very large scale. This difficult (though not impossible) to do push from a linux server to a linux client network wide. That doesn't even say anything about access to network resources. This is all done through the "single-sign-on". One other misconception is that users need zero rights to do their jobs. God is root, (L)users are dirt. Come on we've all said it, at least thought it. There isn't an easy or safe way to offer more permissions to linux users pushed from a central server. Group policy does this and more for windows networks. Until *nix admins stop marginalizing windows servers, and linux directory servers start offering more services to their clients, windows will stay the top dog in corporate intranets.

If you are looking for an AD killer, there is also an interesting Mandriva Directory Server : http://mds.mandriva.org .

Most people ignore, but Ubuntu is not up for make super/turbo best operating system in the entire universe (**gg** that sounds stupid).

All their business is based on providing services for software distribution, (even some open source techies, just call Ubuntu, only a "distributor" and not anything else.)

That's why I will never ever use Ubuntu, their priority, is not quality but speediness and availability, which is obvious very prone to bugs.

Most people don't have a clue when they talk about massive migrations, you simply cannot migrate an entire school or enterprise using Ubuntu, just for the fact that their priority is not quality but quantity and software availability at any cost. Even on bigger enterprises using a non certified distro is prohibited, Novell or Red hat is preferred.

Of course for showing your coolness at an install fest it is perfectly OK!! :UBUNTU users: REALLY get a CLUE.

bye.

I have been using Fedora for several years now, since RH7. I think it has gotten much better recently. My main annoyance is that I have to keep upgrading every 6 months or year, just when I get things like I want. I use some non-free software for my research (Matlab, Mathematica, Intel compiler), and moving/installing those is a pain.

Now I am now trying CentOS, which I think is perfect for my research/work computer. It is free, it is modern enough for my needs and it doesn't keep dropping support for me. It is also similar enough to Fedora/RH that I am familiar right away.

In the next couple years I believe most of the *easy* things will be finished regarding the OS. Sure new things will keep coming along, but once you are not re-writing the basic stuff anymore, it is easier to keep up with those (read at the enterprise time scale). I think enterprise-style distributions will be where it is at, because one if the reasons some people keep windows, is that WinXP is still supported by Microsoft -- after 6-7 years (although who knows for how much longer). I like only having to re-install when I need to, not because of some vendor time-frame BS that isn't important to me actually. I guess there is Ubuntu LTS, but isn't that sort of "enterprise"?

You're partially right about Canonical and kernel development. Canonical has a small kernel team, and it works full time simply keeping their kernel flavors in sync with all the upstream patches that become available. Their mission is to produce kernel flavors which work on the widest possible sets of hardware, and this leaves no time for any upstream kernel participation.

Three things about Debian

1. Debian does not employ anyone. That greatly reduces the number of full time kernel hacker positions.
2. Looking at the 2.6.20 short changelog, there are about 30 changesets from people who are commiters on the Debian kernel team and a few more from people that are Debian Developers or package maintainers. That is a far cry from Rad Hat's 636, but it is more than nothing. They just don't show up as Debian employees because they are volunteers and not all of them use their debian.org address for all stuff.
3. Debian might have liked to claim that Adrian Bunk, but he resigned from Debian a few years ago.

Red Hat is an important and very good player and I would certainly hope it will remain for a long time coming. I do not think they need people taking cheap shots at Debian's contributions to free software in order to assert their relevance.

I used Red Hat from version 5.2 all the way up to Fedora Core 5, mostly on servers and always my laptop, but I finally couldn't stand it anymore.

On the desktop side, I got tired of having to jump through hoops if I wanted to play mp3 files. I got tired of the over the top branding they apply to KDE (to the point it doesn't look any different from GNOME). I got tired of the thousands of useless Java and mono apps they've bloated the os with over the last few years. I finally gave up with Red Hat and now I only run Ubuntu, where *everything just works* and they actually care about the desktop experience.

We still occasionally build a Fedora server, but only for machines that are going to be firewalls or network gateways. RHEL has always been (in my mind) much like SLES (Suse Enterprise) - the packages are too old.

Check this other LWN article for more stats on which companies are sponsoring kernel development.

Red Hat ? relevant ? sure. long term bet ? dunno. I'd bet on rPath. RH has the money, the market and the share. They simply lack the vision and the technology to lead the next big wave. They were great in applying *traditional* business models and revenue streams to an open source ecosystem, but lets face it - they don 't have a clue about what will be coming next. They are too happy where they happy now. The ironic twist is that the rPath gang is what made RH what it is today.

Excellent article, you managed to give Red Hat the credit it deserves for it's contributions, without taking cheap shots to the rest of the community.

Many year ago, not sure of the date and it's too late to comb back threw my records, I was involved in a startup operation where the (in this case) insurance carrier provided the hardware (IBM XT) and I was left with the software of a DOS disk to load the system. Boy did I have a hard time! Many conversations with the seller of the DOS system finally got the system operational.

I guess what I am saying is that if one compares installing Linux (no matter what version) with putting up a Microsoft O/S, then there is really no comparison. But the fact remains that a user does not install the Microsoft O/S but has it installed by the manufacturer of the system. So we, as a Linux Community should not back away from these types of comparisons, but say that installing Linux is easy!

In regards to the suggestion of creating an "Active Directory killer", there is already a Red Hat sponsored project called FreeIPA which is in that ballpark area. It integrates Fedora Directory Server, and the MIT Kerberos Domain Controller together which a bunch of admin tools. It also include FreeRADIUS, NTP, DNS and web based administration. There project is open to any contributors and is making fast progress and beta snapshots available for Fedora.

I think that every problem listed above has been listed about Red Hat since at least 1997 when I started working and definitely when I left in 2001. However at that time, it was Red Hat is the Microsoft and Gentoo will be the new super OS. I do not think that Red Hat will ever get away from the image that it is a Microsoft, it only takes etc.. because there is a core of Linux fanatics who have to have someone in the field who they will label as such. Does Red Hat make mistakes.. hell yes. Does Red Hat want to make money, hell yes. But so does Canonical, etc. However to some of the most vocal people, they have to have something they base their OS choice on by what they are not using.. and so logic doesn't come into play ;).

For a possible AD killer look at the freeipa.org project started by some people inside of Red Hat. It will take about a years worth of work by dedicated people to get it into a good format..

And one thing I found from working inside of companies is that buying closed source software and trying to make it open is not as good a panacea as people hope it to be. Too many times the code uses other closed source stuff, or it is a rats-nest of code that isn't "open" enough for other people to dive in and add/subtract stuff too. [Actually I can't remember any of the closed source projects RH bought that didn't turn into that because it took a year to get legal review on whether the code didn't have other stuff in it.]

Anyway good article and thanks.

Hello Scott,

Excellent post! I've been using Linux for about 3 months now and I tried Fedora 8 the day it came out, November 8th! :) Although I don't have history with other distributions of Red Hat, Fedora 8 was a pleasant experience! I have all of my music files backed up on a 20GB hard drive running Fedora.

About the Ubuntu buzz; it's much easier for newbies like myself to install applications on Ubuntu. I think that accounts for much of the current popularity.

I'm trying to become more program/application oriented. It seems like many of the OS's have similar programs. Gimp, Rhythmbox, OpenOffice.org, Sound Juicer, Evolution, Gaim aka Pidgin, F-spot, Tomboy Notes, Wanda the fortune telling fish, and the Terminal. I think these applications are the future of Linux, and they'll get better as time goes on!

Thomas Fleming
MailTo:Thomas@MontanaReal.com

I have my own favorite distros (*cough* Gentoo), but right now Ubuntu has a huge momentum behind it. It's easy to get up and running, and has an extremely active community behind it. Yes, Fedora is just as easy as Ubuntu in almost every respect, but it doesn't have the same momentum right now. Just imho. :)