Red Hat, SUSE, and Ubuntu in 2008

SearchEnterpriseLinux.com has a good analysis of what 2008 could have in store for the three leading Linux distributions: Red Hat, Novell's SUSE, and Canonical's Ubuntu. Red Hat needs to regain its Fedora passion, according to the article, while Ubuntu needs to learn to channel its Mac-like following into cash. Novell? Novell needs to learn how to market:

The reason that Novell barely gets credit for its work is that its marketing team never leads with anything remotely innovative. If they played it any safer they'd be asleep!...

In the next 52 weeks Novell needs to do what they do best: innovate. Then they need to do well the thing they do worst: they need to lead with their innovation. They need to create a mass marketing campaign around SUSE Linux and its new innovative features that will leave the other vendors in the dust.

When I was at Novell I used to hear constant griping that the market wasn't voting for its products, given how innovative they were. Novell does do innovative work. It just doesn't have a clue as to how to tell people about it. Until it does, the market will continue to belong to Red Hat and Ubuntu.

[alexei_roudnev]

I can't agree.

I used all 3 systems.

RHEL is really a dumb, messy, and primitive. But it keep a few innovations and market itself well - much better then system is in reality. And it _IS STABLE_. It has the worst naming convention, the worst packaging, the worst set of rpm's, teh worst (no any in realiuty) administrative tools... and it is still #1. (Good point for the business researchers - how so primitive system can lead the race...)

Ubuntu is excellent in many areas - good packaging, simple installation, very good support of the community, fast and reliable. But it don't have (just as Redhat) modern administrative tools and it, worst of all, don't have many server-level features included in the base system (and it is not supported by Oracle and many other vendors, which makes it unusable for the servers).

SuSe... It's amazing. They have all components necessary for the success - good Server, Desktop and Open versions of the same generation Linux (so that I can use OpenSuSe when I dont need support, and easily switch to SLES when I need it), they are (still - amazingly, why, but still they are) supported by Oracle, they HAVE, (only one of all) good administrative tools (yast2) so that they can be used by normal (not geeks) people (RedHat can'te be used at all, and Ubuntu can be used with a lot of grain - many operations still requires vi and command line window). And they have a great choice of applications:
- App Armor is much better vs SELinux
- They included OCFSv2 by default, Oracle support by default, DSK is available, SLES SLED and OpenSuSe are inter-compatible.

And.. they are still behind. For the many reasons:
(1) Their contract with Microsoft distracted many (huge number) of community members. I don't have a numbers but looks as this contract became a huge show-stopper for SuSe and caused many companies to remove SuSe from any consideration.
(2) Extremely bad support and feedback in Novell makes it difficult to fix a bugs and report a problems. Extremely poor beta testing cause them to release new systems and service packs always not-ready for the real use (a very few exceptions exists). The overall quality of SUSe dropped every year - when it was German system, it was rock-solid, when it became Novell, it became more and more buggy (what is worst - most bugs are minor, such as for example 'auto installation in SLES10 / OpenSuSe10 works the wrong way if disk has partitions from the previous installation' or 'yast2 users admin tool can't work if NIS database contains duplications' and so on.
(3) Terrible marketing decisions. Who an idiot ever included ZenWorks into SLES and made it primary update system? As a result, if I can install any application in CentOs by simple 'yum install php4' command (for example), the same command takes 20 minutes in SLES because it reads all caches, opens all repositories and so on...
Where is Jboss - users uses it widely, and they should find any way but keep it afloat in repositories... And so , so on... marketing is worst of all.
(4) Bad support for the community. Yes, OpenSuSe has a good support - but SLES don't . When they lost the server with e-mail lists, they spent 2 month restoring it - primary communication server. They never merged together old and new communication forums (e-mail and web based). They never provided access to the bug database (even for the subscribed users). And they start to loose ground in some areas - SLES9 SP3 was an excellent Oracle platform, SLES10 is (still! surprise for me) Oracle certified, but if you read a forums - it is not used much anymore.. people try it, got a different bugs, and switch to more stable platforms...

So they really need to market their Linux - the product itself is excellent, the combination (SLES, SLED, OpenSUSe) is excellent (and not as Fedora - OpenSUSE is a good parent for the next SLES and it includes almost all server related and desktop related things). And they need a better PR and better Feedback. I am surprised reading Linux Magazine - no any mention of SuSe - #1 was Ubuntu , #2 Mandriva - but 50% of my friends who works with Linux uses - guess what - different versions of Suse...

So, it is mystery for me - how Novell could provide so bad PR for the so good product. Looks as Linux is still a stepchild in the Novell family. And it is - I receive a lot of disks by subscription, and only 10- 15% are Linux related... other have Novell stuff so they are going directly to my waste-bucket.. but I can't subscribe to 'linux only' things...

(One example of the bad marketing. SuSe is the first and the only who integrated together OCFSv2, heartbeat2 cluster, and EVMS volume manager. But instead of making it _install out of the box in one click_ and keeping it tested and used everywhere, they abandoned it - so making difficult to use EVMS (which is really GREAT but is not easy to use without a good gui and/or yast2 front-end) and OCFSv2 (which is not stable enough when it uses o2cb cluster software because it is too primitive) - and so making difficult clustering (some of their scripts in heartBeat are broken, and new yast2 module fro heartbeat lost some optiosn from the old one). Instead of testing it, polishing it to the sunshine and then marketing it as 'We have a tested, out of the box, 100% reliable cluster solution' they abandoned the whole project... And it is just a single example.)