Red Hat is one of the prime participants in a movement that has channeled hundreds of programmers' collective energies into a product, a version of the Linux operating system. The company, which has become the top seller of Linux software and services, employs several top-ranking Linux programmers on its staff and probably has more control over the software's direction than any other single company.
Naturally, there's no shortage of new technologies bubbling up from the company, ranging from software that could standardize the way cell phones switch on to a program that holds records for delivering Web pages as fast as possible.
But Tiemann, winding up his first year as Red Hat's CTO, is focused instead on the Red Hat Network, a system the company just launched to make sure technology doesn't get ahead of the people who use it.
"We feel that the technology industry is at this point where the technology has exceeded a lot of people's capacity to take advantage of it," Tiemann said in a recent interview.
Naturally, Red Hat hopes to make some money off the service by offering it as a subscription service that will automatically update Red Hat software and other Linux programs offered by Red Hat's business partners, Tiemann said.
Tiemann talked about the Red Hat Network, the new version 2.4 of the heart of Linux, and a host of other matters in an interview with CNET News.com's Stephen Shankland.
CNET News.com: With the acquisition of C2Net, Red Hat is expanding beyond just offering Linux to offering higher-level software for delivering Web pages. Are there other areas of open-source software you expect to expand to? Are you thinking of doing something with Sendmail email software or PostgreSQL database software?
Tiemann: I think that before we go and try to expand in too many additional technical directions...we need to address the most pressing problem today, which is a critical shortage of IT (information technology) resources.
The Red Hat Network is really designed to substantially simplify the administration of the Red Hat software. When you look at the IT resource issue, even if Silicon Valley lobbyists are successful in getting all of the H-1B visas that they can ask for from Congress, there simply aren't enough people to fill the jobs that exist today.
Our goal is to focus our efforts to make Red Hat 7 the easiest solution on which to develop, deploy and then ultimately manage.
At Red Hat you probably have more influence over Linux than most companies because you employ a host of open-source programmers, but obviously, a lot of the Linux development is outside of your direct control. How does that work? Can you plan on having this or that feature arriving by this or that time?
No, you definitely can't, you cannot make those plans very effectively. But at the same time, what you can expect is that open source is going to continue to disrupt the current software industry.
If you look at the macro trends, the cost of hardware is trending toward zero. The cost of network bandwidth is trending toward zero. And the cost of software is trending toward zero. Open source is accelerating those trends. And we actually see it as beneficial, largely because our position is one of providing services.
Did you initially plan to ship Red Hat 7 with the 2.4 kernel?
That might have been something that we had hoped for, but at the same time, which kernel ships was not really that important to us. If it were very important, we would have waited. But from our perspective it was much more important to deliver the Red Hat Network functionality.
The 2.4 kernel I think is going to be nice. And by the way, if you are a registered user of Red Hat Network, then when the 2.4 kernel has (passed quality-assurance tests) and been certified by Red Hat, your system will be 2.4 kernel-ready. Red Hat 7 ships by default with 2.2.16. But we actually also do ship with it a preview version of the 2.4 kernel.
The 2.4 kernel is better able to harness the power of high-end servers with multiple CPUs. What does that mean for you?
The 2.4 kernel is definitely going to open up some new opportunities. According to our kernel people, the 2.4 kernel should have clear sailing in terms of scalability to 32 processors and beyond, which is far ahead of where people thought Linux was going to be as of even 12 months ago.
Twelve months ago, people really thought it was designed as a single-CPU operating system and it was not going to make it past four processors with any sort of grace. And now we're showing dramatically good scalability on eight-ways. We're just looking forward to when companies are going to start building 16- and 32-way systems.
In what kind of tasks do you see the benefits of that multiprocessor performance?
Certainly for Web serving and application serving and other kinds of e-commerce solutions, where the name of the game is high throughput, those tend to be probably more friendly to scalable processing than some of the back-end database codes, which are notoriously challenging in a lot of ways.
How do you see your relationship with the open-source community in the future? Are you going to keep on trying to hire open-source gurus, or do you feel like you have a sufficient stable now?
I think Red Hat is extremely happy with the quality of people that we've been able to attract and retain at our company. The open-source community is in this relatively unique position compared to the proprietary community in that the open-source developer has complete freedom of choice about what they work on and who they work for.
So I think we will continue to hire the people that we need to get our job done, but I don't think it's necessary or necessarily even proper to try to hire the entire open-source community. That's not our strategy, and that's not our plan.
Do you have plans to capitalize on your dominant market share position by charging business partners fees to be included on the CDs Red Hat distributes?
We have software certification, hardware certification, which includes things like peripherals and computer systems, and also embedded certification. I think over time what we're looking to do is to really increase the number of certified programs, applications and devices to make manageability easier. Right now it's fairly informal, but I think that as commercial partners do certify, it's going to make it easier for people to deploy and easier for us to help others support their systems.
So are you looking at some kind of deal where the software that's distributed on Red Hat CDs is going to be certified?
Yes, it's not going to be overnight. But I think that our goal is going to be to increase the percentage of certified solutions on the Red Hat CD.
So could that drive some revenue for you guys?
Yes, but I think it's a mutually beneficial thing. There is revenue associated with certification, but at the same time I think it saves money for the (companies) getting their programs certified, because then Red Hat becomes able to support that software.
What new technological advancements are on the horizon for Linux?
We won a (Comdex) show award for Red Boot. It dramatically expands the opportunity for Linux in the embedded space and it also provides I think a very easy-to-understand model for how we can do remote manageability.
With Red Boot...you can use eCos as a platform for booting Linux. You can port this environment to every kind of CPU which is supported by eCos. That dramatically simplifies the process of porting and maintaining and most importantly upgrading Linux. So for example, if you have a system which is built with Red Boot, then you can scrape off one version of Linux and put another one on top and that simplifies the remote management problem.
And what about Tux, the Web server that's built into the Linux kernel?
Tux is a Web server that we built. It's actually integrated with the 2.4 kernel, and it delivers world record SpecWeb (testing standard) performance. We set four records: single, dual, quad and eight-CPU records. One of the things that we see is being able to use Tux as an Apache (Web server software) accelerator.
Do you have any longer-term plans for Tux?
Tux is so much better on identical hardware--two to three times faster than (Microsoft's competing) Internet Information Server. There's not much to be gained by trying to be three to four times faster. So what we're doing is we're going to figure out what is the next problem we need to solve. And it keeps coming back to manageability.