Slashdot Mirror
Interview: Larry Augustin Finally Answers
1) THE question
by Nicolas MONNET
Much like most net-related IPOs, VA's valuation is impressively high. Much higher than bigger and older companies which have proven to be able to actually MAKE money. So let's assume that this valuation actually means something. How do you see VA Linux fulfilling this promise, IOW, what will it be like when it will sell as much as its stock price mean? Will it compete directly with Compaq and the likes on the server market?
Larry:
We're building a serious company here and we plan to grow significantly, and I think our valuation reflects that. I think we are undervalued compared to some of the other companies out there based on the potential of Open Source and our target markets. The market opportunity is huge.
We sell systems based on Open Source software and provide significant amounts of consulting services to companies building major sites on the Internet like eToys, Akamai, DoubleClick, and Starmedia.
As those customers grow, we grow with them. Intel estimates that only 5% of the servers needed to run the Internet by the year 2003 are deployed today. That's a huge market for VA Linux. That's where investors see the big potential.
But you have to understand more about what we offer those customers. On the surface, people think we're a hardware company. We're not - we're an Open Source company. We work directly with those customers to solve their problems using Linux and Open Source. We offer some of the most powerful high-density rackmount servers for the Internet server market as part of that solution. Since you mention Compaq, in at least two instances I know of VA engineers have rescued large installations using Compaq hardware. Although we offer the best hardware for Internet servers and will continue to do that, that's only a fraction of what we do. We have a tremendous amount of expertise in Open Source, and it's really that expertise that we offer to our customers, and you can see some of this in our announcement this week with HP. The systems side of the business really just gives us a solid platform on which to build a known, stable, and supportable solution for our customers.
So the one thing underlying all of our business is our commitment to Open Source. The computer industry is changing; Open Source is emerging as *the* way to develop software in the next decade. The companies that figure out business models that work around Open Source will succeed. At VA we will be the leader in Open Source. I've built a company with a commitment to Open Source at its core. I don't want us as a company to have any incentive to develop proprietary software. I'm building a company that succeeds if Open Source succeeds. That creates a business model where supporting Open Source development (through initiatives like SourceForge.net) are absolutely critical to us. We will be Open Source pure.
2) IPO Question
by jezzball
I know how you choose the people you picked as friends and family (being one of them), but were there any specific HOW-TOS you looked through? Did you weight projects? Is there a place where you've amassed this information for people to look at?
I was quite surprised myself at being picked, as several other people who I felt have contributed more than me weren't.
Larry:
Sometimes going second has its advantages. One of the pieces of feedback we had from the Red Hat program was that people in the developer community wanted to understand how their names were picked.
From the beginning we decided to create a process that was as mechanical as possible in extracting names of contributors Open Source projects in general and Linux specifically. Chris DiBona and Chip Salzenberg assembled a set of custom scripts that scanned for developer names. It's best if I just leave the description of that process to Chris.
Chris:
There is a lot more to this question, so this is a long one...
When it was clear that we were going to go public and we wanted to do a community program, Larry and Todd (Our CFO) asked me to start working on putting together a list of folks to target.
Since I got word pretty early (before the initial S1 filing). I wanted to plan for as much flexibility as possible, as I knew that while I would probably be in charge of administering the program that final say on who we wanted to help would be not just mine, but on the shoulders of Larry and others (like ESR, who is on our Board).
So I first got mirrors of Sunsite, gnome.org, kde.org and many other software repositories. I wrote a perl script to rifle through them and un tar.gz the source code therein. It worked out to about 11 GBs of files. I then ran an e-mail scanning script that Chip Salzenberg cranked out for me against the archives to pull e-mail addresses.
This netted about 17,000 e-mail addresses. Each e-mail address was stored with a reference to all files it existed in.
I then went through the list by hand and got rid of bad addresses. This included things like lug addresses, root@ addresses and corporate info addresses. I also removed known silly addresses (foo@bozo, billg@microsoft.com) and straight errors that got through (1024x768@75mhz and such). This brought the list down to 12,000 e-mail addreseses.
I set these addresses aside and began accumulating contributor lists for various projects including gnome, kde, cpan, debian and others. These when concatenated with the top 1000 addreses from the scanned addresses amounted to about 3500 e-mail addresses. Since our orignal target was to provide 2700 people with an opportunity to participate, we felt 3500 was an appropriate target.
We were wrong. About three days into the program, we realized that our returns were much less than anticipated. At this time I scambled to include more people, I took the next 1000 from the scanned addresses, the beta developers from SourceForge (another 950) and a bunch of people were added anyhow because we simply missed them (about 200 or so) We also realized that due to a glitch in the scanning process a chunk of the kernel credits list was missed and I added those folks and made sure they were notified. For the most part if somone had been in the kernel list, they had already e-mailed me personally and I had already added them.
Our selection process (including Sourceforge) meant that Open Source developers on other platforms were allowed in as well, which was cool by us as we feel open source benefits all platforms.
In the end we got to include about 1500 people from the community, when we realized that we were going to have more shares available in the community pool, we reallocated to allow people to recieve 140 shares. At this time our price had gone up to $30, so we allowed people to take part at the 50 share level if they still wanted to participate and didn't have the money for the full 140 or 100 shares. Someone on Slashdot suggested this, BTW.
Everything went very smoothly, and I'd actually like to thank everyone in the free software community. Why? Well, as many of us now know, the VA IPO was very hot, we had an astonishing number of people calling, e-mailing and visiting asserting that they were a community member, and we took all of them seriously as that's the right thing to do.
The reason I want to thank everyone is that, 99.9% of the time, we were able to tell right off if someone was legit if they were nice. For the most part, everyone we included in the program was nice, and everyone who was trying to get into it fraudulently was a jerk. It made it really easy, so thanks, I appreciate it and the folks at the Underwriters said that the folks in the community program were just about the nicest people they've had to deal with in any designated share program they can remember.
Also, to those who have done things in the free software world and were not given an opportunity to participate, I just want to say thanks for your hard work. We do appreciate it, but we were working with a limited number of shares and we had to stop somewhere, but this doesn't mean that we don't appreciate you or what you've done.
As a last note, we are -very- sorry we had to send such a huge e-mail out to everyone. A megabyte is a really big e-mail to get unexpected. The SEC was very clear on this, we were given the choice to send it out with prospectus and forms in a PDF format or not offer it at all. And since for 90% of the list all we had was e-mail address, this meant spamming a lot of people. In the end, I personally decided to do this, as I felt it would have been a greater error to not include free software developers in our IPO. So in essence I had to decide what was worse, spamming or exploiting the community. I chose to spam rather then exploit. I still feel bad for it, but I do feel it was the appropriate thing to do.
3) Prices
by FascDot
Every time I need to buy a computer (for self or business) I check out VA Linux Systems. But your prices are always $500-$1000 higher than even the Microsof-tax laden goods from someone like Gateway or Dell. Is this all due to volume discounts or is there something else at work?
Larry:
First of all it is not our intention to be the lowest cost source of a box. We're not just selling a box. Typically we're selling a great deal of service and support around the box, and in the end the customer gets a better value as a result.
But if you're just comparing box prices, it also depends on what you are looking for. We do competitive pricing comparisons weekly versus Dell, Gateway, and many others. We work very hard to make sure we are comparing similar configurations. We have some configurations in the server space where we consistently come in lower because we have products that are better suited to the customer. In other cases, we have configurations that are consistently higher. For example, our StartX MP dual processor desktop system has easily a lot more cooling and power than anything else in its space in the market. Frankly, the chassis and power supply are over-engineered for a dual Pentium III desktop. As a result, we look a lot more expensive when comparing products in that space; the quality is apparent until you open the system up and do things like thermal profiles inside the chassis.
Also, I don't think people realize that we do a tremendous amount of engineering on our products. Every disk drive is flashed to a known, qualified firmware level. We regularly work with disk drive manufacturers to fix firmware problems. We need to do a lot of qualification that a Windows supplier has already done for them. For example, we had an incident in November where we attempted to qualify a major name-brand memory supplier under Linux on one of the Intel server motherboards. We expected it to be easy because Intel and the memory manufacturer had already qualified the hardware under NT. Well, Linux is not NT. We could make the memory fail under Linux. In normal operation the error might show up as a system hang once every few weeks. That's unacceptable to us and our customers. We log thousands of hours of testing of every component and configuration before we qualify it under Linux. No one else does that.
4) The next step
by DarkRyder
Now that the IPO period is over, what directions do you see VA Linux moving in? Are there projects that have been placed on the back burner until the IPO could 'make it possible'? If not, have projects been undertaken, or put into the planning stages, that wouldn't have been if the IPO hadn't taken place? Or is this just a calculated business move, with no plans beyond simple, steady growth as a company behind it?
Larry:
As we said in our prospectus, we'll be using the net proceeds from our IPO to fund operating expenses including broadening our professional services division, hiring more R&D personnel, and expanding our sales operation worldwide (starting in Europe). Another important benefit from the IPO is the credibility it's given us as a company. A lot of prospective customers who might not have considered us before the IPO now look at VA Linux Systems as a company they know will be here over the long term. We're publicly accountable now, and we're focused on execution. It's a responsibility all of us take seriously.
5) Plans for other platforms?
by imac.usr
Specifically, the new PowerPC motherboards that some companies are planning to build...is VA planning any systems around these boards, or do you see a market for platforms other than i386 in your company's future?
Larry:
We're always looking.
6) Your view of Linux and the open source movement
by God I hate mornings
I've been watching with keen interest both Linux and the open source movement with great interest for the past few years. Now with anything Linux being the darling of Wall Street, it would seem all one has to do is mention Linux and watch the rush of money.
My question is this: How do you, personally, view your fellow Linux distributors? Do you see them as competition to be toppled? Or do you view them as competitors in a worthwhile cause?
Larry:
We are distribution neutral. I believe there is one Linux. We have pre-installed various distributions over time, and we're committed to supporting standards and customer preferences on this issue. Most of our customers don't really care too much; in internet infrastructure implementations, most people have very specific loads customized to their needs. We help them with that customization as part of our added value, and we pre-install the results. We often start from Red Hat, but based on customer demand we have also started making plans to offer standard Debian and SuSE based loads this year.
I can't remember actually competing with a distribution company for customer business. We typically partner with them. We put our competitive focus on counter- or quasi-Linux companies such as Microsoft, Sun and Dell.
7) Joint marketing effort?
by Floris
Watching all the little companies (and the not-so-little companies) that build and support Linux as an OS, I have been wondering about when we will see Linux commercials on commercial TV?
Some CEOs of Linux-related companies have said that their goal right now is to "grow the Linux market" rather than competition with the other vendors. The main competition would be Microsoft rather than each other.
However, Microsoft has a massive marketing machine at its disposal. The Linux vendors dont have nearly its marketing muscle. Are there any plans for a joint marketing effort to counter that?
Larry:
For practical reasons, we're not planning any TV ads. Most of our sales are to folks that buy based on information from other sources.
8) VA Linux Software Patent Intentions
by Aaron M. Renn
In your S-1, you mention that you have applied for a patent. Here is the relevant paragraph:
"Our systems consist primarily of commodity hardware components in combination with the Linux operating system. While we have developed some proprietary techniques and expertise, most of our activities and systems are not protectable as proprietary intellectual property and may be used by competitors, harming our market share and property, we generally enter into confidentiality or license agreements with our employees, consultants and corporate partners. We have also recently commenced a patent program and to date have filed one patent application. In general, however, we have taken only limited steps to protect our intellectual property. Accordingly, we may be unable to use intellectual property to prevent other companies from competing with us. In addition, we may be unable to prevent third parties from developing techniques that are similar or superior to our technology, or from designing around our copyrights, patents and trade secrets."
I realize you will not want to disclose the details of your pending patent application, but I am interested in knowing whether or not it is a software patent. And, what VA's policy towards software patents is, in general. Will VA apply for software patents? If so, how will they be licensed? Will they be made available on free software terms to free software developers?
Larry:
We have applied for a patent on the techniques we used in the design of our FullOn 2x2 high density rackmount server. We have some unique packaging technology there.
I personally believe the patent process with respect to software is broken. However, it has not been a priority because we have not applied for any software patents.
9) Intel influence behind the scenes?
by RelliK
A while ago VA Linux Systems (then called VA Research) used to offer Alpha systems. But soon after Intel's investment in VA, the Alpha systems disappeared. Last I checked AMD systems were not offered either (and don't tell me there is no demand for either Alpha or AMD). Intel is the *only* choice.
My question is: being a Linux company, will you continue to offer people a choice or will you, like Dell, become a slave to Intel?
Larry:
We dropped Alpha well before the Intel investment because we didn't have the resources to support Alpha and Intel on Linux. However, we're always exploring all of our options.
-------
Amazon can turn a profit any time they want... all they have to do is turn off the marketing machine and stop spending so much on infrastructure.
But the reason they have won is that they won the branding war (which is the only war worth winning). They knew that they had to spend giant dollars to rise above the noise of every clown598.com.
Say what you want about Amazon, but they have succeeded brilliantly in diversifying from books. In five years Amazon will be the Wal-Mart of the web.
--
I can understand the difficulty of supporting both x86 and Alpha, but support for the AMD (being x86) is much simpler.
The lack of AMD support would seem to be purely due to intel's influence.
Larry wrote:
:)
"But you have to understand more about what we offer those customers. On the surface, people think we're a hardware company. We're not - we're an Open Source company. We work directly with those customers to solve their problems using Linux and Open Source."
...that's just not true, at least not in the case of the *entire* list of companies he mentioned.
Specifically, at StarMedia, where i'm chief systems guy, we've got hundreds of VA's boxes racked up around the world... we *love* the hardware, and love VA's ability to get us another truck full of servers delivered really quickly...
...but we don't run linux on *any* of them. Like several other companies who've been challenged to scale *way* up, we run FreeBSD. On VA "linux" boxes.
i (and other members of my team) kind of resent outside companies talking about how we need their consulting help... it *is* kind of fun to get a "sorry, we don't support that" response from such an "Open Source" company...
...we're not so resentful that we'd stop buying servers from VA (did you get that last order, Larry?), we just want to make sure people know there's devil stickers on top of the penguins in our cage.
Larry, I do not appreciate your false statements that VA Linux provides consulting services to StarMedia.
As Ryan Nelson, my most senior sysadmin, points out, we do love your hardware immensely, as it suits our purposes well. However, the fact of the matter is that we use FreeBSD, not Linux, which is not where VA's business is focused. As such, we have really not had much involvement with VA outside of ordering hardware, as your team does NOT have the FreeBSD expertise that we require.
In fact, perhaps the only consulting that has taken place was with MY team consulting YOUR team on alternate hardware components to replace those pieces of the host which have no supported drivers under FreeBSD, thus enabling us to continue to use your hardware to meet our needs. Yes, we were offered additional help in developing a FreeBSD driver for the unsupported hardware, for a serious premium.
If you were truly committed to Open Source, you would offer to develop drivers for other popular free Intel-based Unices without any type of monetary demands. This would, in the end, help you as well, as members of other communities outside the Linux community would turn to you for their hardware needs, thus driving your business further among competitors.
Don't get me wrong, I love Linux, and in fact use it at work and at home for my desktops. However, it should be noted that for large-scale operations, Linux does not always meet the requirements, and I would think you'd like to corner the market on these large-scale operations as much as possible.
Sincerely,
Adam Firester
He feels they're UNDERVALUED?!?
.com, software, hardware and tech company are grossly overvalued. Really. There used to be a day when companies were valued based on their actual performance. Amazon suceeded in corrupting that definition so that it included their future performance.
No. VA Linux, just as every Linux,
Now, every company out there seems to believe that they should be valued based on what they think they can do rather than what they are doing now.
It seems that IBM is about the only tech company that's valued anywhere near the levels it's revenue would reflect if it were in any other industry. They actually trade a lower PE ratio than General Electric.
I mostly liked everything else he had to say, but due to the order of teh question, i was set off by him. Build a company. Get the sales. Get the PROFITS. Then you'll get your valuation. Any other order is just insane.
This was beautiful, and totally worth waiting for. If Larry had rushed things, we might have gotten short, coherent answers. Fortunately the man is a perfectionist, and we got a total dodge of the AMD question, a lot of vaugeness and 135 repetitions of the phrase 'open source'.
Great.
--Shoeboy
We sell systems based on Open Source software and provide significant amounts of consulting services to companies building major sites on the Internet like eToys, Akamai, DoubleClick, and Starmedia.
If you're trying to endear yourself to Slashdot readers, perhaps you shouldn't emphasize your ties to eToys and DoubleClick this month. Eh, Larry?