Amsterdam, October 15, 1999First, a few words about the conference itself. I was very pleased with the fact that the organizers of the conference - LogOn Technology Transfer GmbH - invited members of various Linux User Groups to participate in the conference at no charge. I was, however, rather disappointed that I had to shell out 20 NLG for the conference proceedings. They could do us a little favour and give the conference proceedings for free as well. Despite this the conference was very well organized and I am glad to have participated in it.
One thing that kept us annoyed was the microphone. It kept turning on and off all the time. We even had the joke that it was running Windows CE and kept crashing...
Scott made a well-balanced presentation that focused on both challenges for Linux community and the opportunities for business development. He pointed out that Linux just started to take off really big time but it already has witnessed a huge success. He is convinced that Linux is a new opportunity with limitless opportunities.
Linux insofar does not present users with all the capabilities they might want but we know that in the Open Source model the things move quite fast and the improvements are abound. In this respect Scott shared with us his vision of Red Hat being the ``quality control funnel'' for the code produced by the Linux community.
Linux is now positioned in the middle-level market, that is not yet widely accepted as a platform for either embedded systems or the supercomputers. Linux, as opposed to OS like SCO and NT has huge capabilities and it will expand both ways - to the bottom of the embedded and to the top of the high-end systems.
Linux is taken quite seriously by the heavy-weights (like IBM, Oracle, etc.) and this is the proof that Linux is taken seriously and being accepted by the market.
The Unix world was segmented by the proprietary extensions and many people are wary of the possibility that this might happen to Linux. Scott argues that the Open Source development model helps to avoid segmentation and, on the contrary, makes the systems more compatible and standard. Therefore Linux will avoid the segmentation that plagued the proprietary Unix platforms.
When we talk about Linux in the enterprise we stop talking about its superior technology. The business does not care if the OS is neat and fancy. The requirements are quite different: stable, predictable, flexible, scalable system. Here is what the companies are, in his opinion, expecting:
During the panel discussions Scott pointed out that Red Hat is committed to the Open Source development model and will always be true to the Linux community.
Laurent first reassured the Linux community that Borland views the Linux OS as being a mature operating system - stable and well-scalable. Since Linux becomes more mainstream Borland cannot stay on the side and has to move into this new for them market. Borland is not committed to any one particular platform.
The point of action that Borland takes is in moving Linux from almost purely server environment onto the desktop. Borland makes a lot of effort in the joint marketing of the Linux products (strategic alliance with Corel) to accelerate the seeding of the Linux market.
The Linux - related products from Borland are developed in the project code-named ``Kylix''. The project is headed by Michael Swindell. They took a number of surveys from the developers in the Linux community with not-so-staggering results. KDE seems to be the most preferred GUI environment with GNOME at the second place and the Linux on x86 platform is still the most popular port.
Further on Laurent concentrated on explaining the strategy of Borland to cover the distributed applications market and why they think it is so important. He pointed out that middleware is not always bad and sometimes can help to develop faster and more reliable applications. This is the point for their VisiBroker product which supports CORBA 2.3. Other two products were InterBase - SQL relational database and JBuilder - Jave development and application environment.
Borland is now porting their C compiler - bcc - to Linux and we saw a demonstration of the alpha version of it. As a demonstration Laurent compiled xgalaga from scratch with gcc first and then with bcc. Compiling with bcc took about twice as short time. Some of that may be attributed to files being cached in the file systems buffers on the second run but still the effect looks very nice. Unfortunately, they did not provide the bcc on the demonstration CD so I could not try it out myself.
This presentation consisted of two parts, one being concerned with how Linux is doing in the marketplace and how it can help the businesses to struggle with their low budgets and the second part being about Informix products.
The total cost of ownership (TCO) consists of the following major costs:
From the analyst perspective the technology which is represented by the GNU/Linux operating system is widely accepted. There is a big threshold for every new technology to overcome between being accepted by ``early adopters'' eager for new technology and being accepted by the majority. Linux has passed this threshold and is now mainstream.
Linux is highly dependent on the Web and the other way around. They are becoming unseparable. This means that Linux will grow as fast as the Web grows. Web at the moment is the fastest growing technology ever in the history of the mankind and Linux is an excellent platform for the Web.
Informix takes their tools to the Web too. The Informix database is mean and lean and supports practically anything you can think of that you might need. Moreover, Informix database is object-oriented which means you can create your own types of objects and methods to access them. I was rather impressed by the capabilities and the small size of the Informix database but I will not give all those details here. The important thing is that they have it for Linux and you do not even have to tell your boss you moved the database from NT to Linux - they won't notice.
Mark Daly made a very technical presentation about their products. He explained why and how they build the middleware layers that are used in the software projects by their clients. They are focusing currently on enabling their clients to build and manage e-Business applications fast, easy and cheap. They are the market leader in the transaction server market. The main products are Application Servers - BEA Tuxedo and BEA WebLogic - and Application Integration - BEA eLink. Tuxedo and WebLogic are available for Linux now.
The strange thing I noticed is that they announce that their products are available for RedHat only. I heard that before and it keeps me wondering why...
Jurgen Geck explained the background of the company and the current status. Currently SuSE employs about 200 people and is aggressively expanding worldwide. The main focus of the company remains to expand further worldwide as a service-centered company.
SuSE pays a lot of attention to the consistency of the distribution and testing. The automated distribution build system allows them to reproduce every single CD they put on the market. They have a number of people who are specialized in particular areas and are responsible for particular packages. Those people keep in touch with Linux developers around the world.
SuSE distribution is available in English, German, French, Italian, Portuguese and Spanish. SuSE is working currently on the support for multibyte languages like Japanese. They want to provide general Unicode support so that they can provide the distribution potentially in any language.
SuSE support (besides the 60 days installation support you get when you buy it) is divided into three categories:
SuSE is providing Linux training and certification programs as well. They do not seem to do that by themselves though. They have partners around Europe that do the actual training.
SuSE is not limiting itself to working only with Free Software community but they do and will work with commercial software. They do not want to limit the possibilities and therefore they will keep distributing the software which does not satisfy the Open Source criterias. One point was left unclear though is whether SuSE will keep all software they write themselves in the Open Source or not.
I will not describe this one in detail. As you probably already know Oracle is, in fact, an operating system in itself. All it needs is to boot up an then it works using the raw computer resources. They now ported to Linux but I suppose it was no more than writing a ``boot loader'' that could kick off their own kernel. So I do not see this as anything interesting or outstanding. In fact, Oracle does not deny the fact they can run without an extra operating system and even have plans to introduce an ``Oracle machine'' in the future that will run nothing but Oracle and Oracle only. (No surprise Oracle is soooo huge...)
Dell started to market PCs with Linux pre-installed some time ago. They explained their vision of the company and their idea of eliminating the retailers from the chain between the producer and the client. I was surprised to find out that the machines you buy from Dell are actually produced after you make an order. Of course, they keep a bit of stock for the most requested configurations still.
Dell demonstrated a couple of desktop machines running Linux (RedHat as usual) with KDE. They were surprised to find out that not everybody like RedHat and some people would even prefer to have FreeBSD instead so I guess they might be introducing some other distributions on their machines in the future. Did you know Titanic was done on Dell machines running FreeBSD? I always thought it was Linux...
Dell can provide a wide range of hardware - from desktop to rack-mountable servers. You can even buy a Beowulf cluster from them. They are actively negotiating hardware manufacturers to provide drivers for their hardware for Linux (like they do for Windows) so this is good for us, I guess.
Dell provides support for hardware and the support for software (Linux, that is) comes from RedHat and LinuxCare.
The panel discussion was held after the presentations but, unfortunately, it was rather short since some of the speakers had to leave for the airport.
I picked up a few interesting moments from the discussions:
That's it, folks! I would like to request you to send questions, comments and corrections to albert at tigr.net. Please, note that these are my personal notes and I may be seeing things from a different perspective and may have misuderstood some things alltogether. So, no flames, please.
Albert ``Tigr'' Dorofeev
17 October 1999
This document was generated using the LaTeX2HTML translator Version 98.2 beta6 (August 14th, 1998)
Copyright © 1993, 1994, 1995, 1996,
Computer Based Learning Unit, University of Leeds.
Copyright © 1997, 1998, Ross Moore, Mathematics Department, Macquarie University, Sydney.
The command line arguments were:
latex2html -split 1 -no_subdir -no_navigation Linux-at-work-1999.tex
The translation was initiated by Albert Dorofeev on 1999-10-18