SuSE CTO: Novell gives Linux 'muscle'

Novell Inc. is banking its re-entry into the consciousness of IT shops on Linux. Through its acquisitions of Ximian and SuSE Linux AG last year, Novell is working hard to regain its stature as a major technology player. In this interview, SuSE chief technology officer Juergen Geck touches on Novell's Linux strategy, starting with its recent release of the YaST (Yet another Setup Tool) management tool to open source, his thoughts on back-porting features from the 2.6 kernel and SuSE's product road map for the rest of 2004.

This Content Component encountered an error

Juergen Geck
Novell has certainly demonstrated its commitment to Linux since the acquisition of SuSE was announced late last year. But were there any apprehensions from the people at SuSE at first?
When we started to talk about being acquired by Novell, there certainly were a couple. By now, they have all dissolved. I can only say the best about my new colleagues. It's really a good thing to have all the muscle for all the ideas we've been cooking up for the last 12 years. When you look at the sales force we now have selling SuSE Linux on the street worldwide, the number of support engineers we now have backing them up, and the kind of background our engineers now have access to bring to the table, the discussions we have are through the roof when it comes to operating system design, kernel extensions and middleware. Can you talk about the release of YaST to open source under the GNU General Public License (GPL), and how important it was for Novell to make such a statement so early in the Linux game for them?
We've been working on our end for years to get YaST to where it is now -- meaning GPL. Novell fully embraced that, supported us on that move and made that move. Half a year ago, I wouldn't have expected [Novell CEO] Jack Messman to go up on stage at BrainShare and say 'YaST is GPL,' but he did.
I don't think there are any exceptions to open source making perfect sense when it comes to setting standards and establishing a good solid software foundation to run your business on.
Juergen Geck
CTOSuSE LinuxAG
What does YaST enable?
What it means is to enable a comprehensive systems management ecosystem that really can span the industry. On one hand, you have management consoles from IBM Tivoli and HP OpenView. CA Unicenter that manage everything. The challenge is that you have 50,000 dependent entities providing offerings for the IT market, hardware and software alike; all of these 50,000 companies and their offerings are targets for systems management. If you want to play catch-up with that massive amount of offerings, you're likely going to lose out in being comprehensive.

So the only answer possible is to have an open source perspective where we enable the industry itself to do that enablement for their respective offerings. That's what we want to support by releasing YaST under GPL where we already have a good supported base. We have 4,000-plus packages of software open source-enabled to be installed and configured by YaST. We have some cooperation with major partners in the industry like IBM with DB2. IBM, for example, in two weeks managed to get YaST to install and configure DB2 from the YaST background. That is what we have in mind with the release of YaST to have it complement the management consoles of management targets. Anything else that SuSE and Novell may have in mind for a similar open source release?
Well, the commodity water level is rising. Of course, it's something we're looking into right now, but I cannot make a statement about it at this time. I don't think there are any exceptions to open source making perfect sense when it comes to setting standards and establishing a good solid software foundation to run your business on.

It is certainly something different to consciously give up the revenue opportunity around these established products. But, I don't think that necessarily this is how open source will evolve over time. When you look at open source projects in general, you will find that they have always progressed very rapidly. They have conquered the market and been successful in areas where only a few people would have expected them to be in the first place. If you have an established player in the market already, from a closed source or open source perspective, it's highly unlikely to have an open source replacement in that space.

What we're driving is to have additional open source offerings in areas where closed source is not today. That is in the systems management space, for example, and data integration where it really comes back to having all of the industry aligned and forced to be aligned around open source offerings. If you have something for free, it's hardly arguable that it makes sense to re-invent it in a closed-source fashion. What kinds of questions are SuSE enterprise customers asking about managing Linux?
It's mainly a cost discussion. Where I see that open source generally has still a lot to gain in the awareness of the IT market is the additional opportunities that open source opens.

If you have people today cutting costs by having standard management targets applied to their infrastructure, that's only half the coin. The other half is that you can make your system configuration highly standards based also highly customized at the same time. For example, what YaST allows today is to have a company-wide server and client environment rollout pre-configured into standard components like SuSE Linux desktop or SuSE Linux Server where you have one ecosystem to really tell all your management targets what to do and how to behave without having to reach out for all the vendors that you want to put into these systems on the management targets.

The types of discussions we're having are around how we efficiently get there and how do we customize SuSE Linux to really make full use of the tools we already have included in there for free. For example, we've partnered with a couple of big banks to have their desktop and server facilities rolled out pre-configured without really changing the code base, without having the need to replace software packages in there, but merely defining the custom configurations in a standardized way. During your session at the recent Real World Linux conference in Toronto, it was reported that you said you were against back-porting features from the Linux kernel. Could you clarify?
That is not a true statement.

What we bring to customers as an additional value on top of what the open source community provides is, for example, back-porting, back-fixing and customizing the system at large, so that it can run as a homogeneous entity, maintenance and support over time. In short, the community provides technology. We complement that technology offering by keeping the same API (application programming interface) stable for five years. That is the core part of the value proposition of SuSE Linux Enterprise Server.

The remarks made in Toronto were made in a very specific context being 'How do I treat changes I want to make to the vanilla versions of open source code bases or open source projects I'm dealing with, for example, the kernel? The degree that you deviate from the open source vanilla versions defines the degree of efficiency possible between what you bring to the market and what you get from the community as a foundation for what you're doing.

What I said was, If you deviate too much from the open source community standards and the APIs defined in the open source project, you're going out on a limb because the investment you have to bear with and make over time to maintain these proprietary changes will kill the efficiency that you get from open source community. If we were to deviate from our policy of keeping changes to a minimum, we would make a big mistake and it would not be favorable for us, or our customers. What can we expect from SuSE for the rest of the year in terms of a product road map?
SuSE Linux Enterprise Server is due to come out the middle of this year and it's going to be the first enterprise Linux based on the 2.6 kernel. It's got full support for anything from NPTL (Native Posix Threading Libraries) to YaST to whatever you would expect from an enterprise release.

We just released SuSE Linux Professional 9.1, which you could consider a preview of things to come. We're going to have a variety of releases coming up this year where we are lucky enough to have not only one desktop in the pipeline, but we'll stick with the strategy of choice in that area.

What we're looking forward to having again is the only enterprise Linux code base that is available on the server and the desktop. Nobody else has that and it's about the most neglected, most underestimated and most impressive feature of SuSE Linux. SuSE Linux has only one code base and one source code repository from which we've built the desktop, the server, all the language-specific and hardware-specific versions. That background led to the certifications we have, and we're going for more -- and that will hopefully be announced this year. How important within the walls of Novell is it to be the first to market with an enterprise server that's based on a 2.6 version?
When you look at our partner ecosystem -- IBM, Intel, HP, SAP, Oracle, Fujitsu-Siemens -- they all are working on 2.6; 2.6 is clearly where the evolution is going, so we feel that we are making the best decision possible for ourselves and our customers going down that route. Whatever else might have been possible, it will be considered inefficient and not really a sound business decision.

FEEDBACK: What is the biggest implication, positive or negative, of Novell's acquisition of SuSE Linux AG?
Send your feedback to the SearchEnterpriseLinux.com news team.

Dig deeper on Linux news and updates

0 comments

Oldest 

Forgot Password?

No problem! Submit your e-mail address below. We'll send you an email containing your password.

Your password has been sent to:

-ADS BY GOOGLE

SearchDataCenter

SearchServerVirtualization

SearchCloudComputing

SearchEnterpriseDesktop

Close