Thursday, March 30, 2006

Secret Society Success

The recent Secret Society party was a huge success! Allan already posted a bit on it. We had a bunch of folks interested in how open source is going to impact the network infrastructure market. That included users, engineers and technologists, business people, finance people, analysts, press, and just about everything in between. Basically, it was a great cross section of the world. The conversation ranged from features and functionality of XORP and the OFR, to open source business models, to what's Allan's kick with vanilla ice cream (I'm a strawberry guy, myself), and why I hate fax machines. All-in-all, it was great fun.

I think the best part for me was when Simon Crosby, Founder and CTO of XenSource, told me that he loved our web site. After scrambling to get it up and finished just as the Business 2.0 article was hitting newsstands, it felt nice to have the work recognized. I wanted something simple, non-corporate, and straight-to-the-point. Simon said we achieved that. Thanks, Simon! Props to Katie Bush, our graphic designer, and Matt Tucker, our web programmer; you guys made it a reality.

Hopefully, many people are now sipping coffee from their brand-spankin' new Vyatta logo coffee mugs, making their colleagues green with envy.

If you were wondering who was there and what we all looked like, I posted all the photos to Flickr. You can see some folks clutching their mugs. Enjoy!

If you missed this Secret Society meeting, please try to come in Vegas. We think the event will be the evening of May 2, but we'll get back to you with more info as we lock things down. We're growing, getting bigger every day. Hopefully, Interop in Las Vegas will afford some international people the opportunity to take part face-to-face.

Friday, March 24, 2006

Bouncing emails and spam blacklists

Earlier this afternoon, I sent out invites to the next Secret Society party taking place on Tuesday. (You're welcome to come. Drop on by.) A couple things struck me during the course of this exercise.

My first question was, why do people give us bogus names and email addresses? The registration form specifically states that registration is optional and has a big button labeled "skip it." We're trying to be as courteous as we can here. Please do the same and either give us real data so we can keep in touch, or skip right on through if that bothers you. We only want to communicate with people who want to hear from us.

The next issue was handling all the bounces. Quite a few of the bounces were because of various mail servers using spam blacklists. It turns out that the particular address block in which our mail server resides was once home to a big spam company. Grumble. Of course the spamming company is long gone, but we're still suffering. We'll work with our ISP to get things moved around and try to get off the various blacklists, but it's really painful. In the era of great Bayesian filtering, it amazes me to no end that people still use blacklists as a binary spam test. I mean, sure it should be an input to the Bayesian filter, but it's just one input. This isn't the first time I have had this problem happen to me. Previous companies encountered the same thing.

Such is life in a new small company...

Industrial fax support

I have already expressed my general disgust for fax machines. We finally got our fax line up and working.

The other day, one of our VCs moved their offices and told us that we could have some old office equipment. Among the scores were some heavy-duty HP printers and a large fax machine. Now we can fax with the best of them. The possibilities are endless...

Wednesday, March 15, 2006

Pictures from ICSI Bears

A couple weeks ago, Allan mentioned presenting at the ICSI Bears conference. The International Computer Science Institute (ICSI) is the home for the XORP project. The Bears conference is ICSI's way of showcasing some of its research areas and demonstrate how they are being used by industry. I finally got around to uploading some pictures.

Allan presenting the Vyatta story:

Pavlin Radoslavov (left) and Atanu Ghosh (right) are two of the XORP contributors. You'll see them posting to the XORP mailing lists quite regularly:

A few loyal Bay Area Secret Society members showed up to support Allan. From left to right: Mark Culpepper, John Shaver, Peter Wohlers, and Mark Weingarten:

Vern Paxon (left) and I, Dave Roberts (right), talking about some of Vern's work on intrusion detection and prevention. Vern is a former chair of the IRTF, the research side of the IETF, and the author of the popular Flex lexical analyzer construction tool:

Tuesday, March 14, 2006

Secret Society Party, March 28, 2006

When we first started Vyatta (even before it was called Vyatta), we gathered a group of friends together to let them know what we were up to and get their input on features and direction. We jokingly called that group the Secret Society. Since we had no offices for the first several months of the company's existence (not until February 2006, to be exact), we held these events in restaurants, bars, and our VC's offices. (The least you can do for your VCs after funding is buy them drinks with their own money.)

When we finally went public in February, we talked about whether we should continue with the Secret Society or not. We agreed that the events were more than just a way to get feedback from people and also more than simply a party, some sort of fusion of the two that seemed pretty important to our culture to keep.

You see, open source, by its nature, is collaborative. In contrast to a closed-source company where the only contributors sit within a 10-cube radius of each other, open source seeks out the best talent, where ever it can be found, globally. Allan mentioned Joel Krauska the other day. Joel lives in the San Francisco Bay Area. But Hasso Tepper has been contributing bugs and comments for more than a month and is a big XORP and Quagga contributor. Hasso lives in Estonia. Kristian Larsson, also a big XORP contributor, has given us plenty of feedback. Kristian lives in Sweden. Eventually, we'd like to connect names with the living, breathing faces behind them. The Secret Society events give us a mechanism to do that.

Yes, I know what you're saying, because some of you said it about 10 nanoseconds after I sent out an event invitation last night: "Dave, thanks for invite, but you guys are in San Mateo, California, USA. I'm here in [insert distant location]." Well, that's true. We can't get everybody in one place at one time, but eventually we see Secret Society parties going on the road (with one already in the planning stages that I'll tell you about in a second), as well as local "chapters" springing up.

Whew! So if you have read this far, you're thinking, "Get to the point." Here's the point:

We'd like to invite y'all (as our graphic designer, Katie says) to the first post-launch Secret Society event. The details are:

Location: The Vyatta Bunker, San Mateo, CA, USA.
Date: Tuesday, March 28, 2006
Time: 5:30 PM PST

We'll have drinks and food and a little bit of Vyatta schwag. If you think you might be interested, please drop me an email so I can get an approximate count for food/drinks. My email address is "dave" at vyatta.

If you aren't yet a member of the Vyatta Secret Society, joining is as simple as filling in the registration form on our site.

Now, I'll say that we're also planning a Secret Society event at Interop, Las Vegas, in May (show dates are May 2 - 4). If anybody is going to be there and is interested in attending, again, send me an email at "dave" at vyatta so I can get an approximate count. I'll send out details to the Secret Society mailing list as we get closer to Interop.

Finally, let me conclude by saying that I think people who see open source simply as "free software" miss the point. Open source is about setting the code free so that people, where ever they are located, can shape its future. It's that participation that creates community. And true community demands more than people typing on keyboards around the world. It demands people coming together, in person, looking in each others eyes, and agreeing to move the world forward. I hope to see you here.

Build 1299 released

We, Vyatta, just released build 1299 of the OFR last night. This build concentrated on fixing various high-priority protocol (BGP and OSPF) bugs.

The new .iso image is available from the download page and the documentation is on the wiki. If you have already registered for the Vyatta Secret Society, you can skip through the registration page.

The announcement is available in the vyatta-announce archives. If you haven't yet subscribed to the mailing lists (particularly vyatta-announce), then you should see the mailing list instructions on the wiki.

Tuesday, March 07, 2006

A graphical look at Slashdotting

We took a look at our bandwidth graph from our colo provider this morning. This is what it looked like. In this graph, time is increasing to the left, so you can see bandwidth from some of the activity last week and the quiet period of the weekend right before the Slashdot storm. We had capped our bandwidth at 10 Mbps and up through this morning we were using all of it. What's amazing is how fast it starts.

We had some people ask us about using Bittorrent for distributing future images. We didn't have time to get things setup this time around, but we'll be ready next time. Hopefully, nobody experienced excessively slow downloads from our site.

Monday, March 06, 2006


Robert just called me. He starts right off with "I'm driving in my car, but my buddy just called and told me we're the #1 story on Slashdot. Can you open a browser and confirm it?" I gulped. I had a feeling this was going to happen at some point. Sure enough, we are. I feel honored.

Fortunately, the server seems to be holding up well. We debugged the previous round of crashes and implemented some fixes.

Come one, come all! We love ya!

Friday, March 03, 2006

Open Season On Open Source?

I spent some time this afternoon reading Sarah Lacy's article "Open Season On Open Source?" at BusinessWeek Online. The article discusses the trend of closed source companies (namely Oracle) buying open source companies (namely Sleepycat and rumors of JBoss). Lacy writes: recent weeks the open-source community has been thrown into tumult. Software giant Oracle Corp. (ORCL ) has acquired two small open-source companies and is in negotiations to buy at least one more. Many experts believe this is the beginning of a broader trend in which established tech companies scoop up promising open-source startups. While the validation is thrilling for Galstad and others in the community, it's also unsettling. Many young idealists who set out to create an alternative to the tech Establishment now find themselves becoming part of it. "When your main goal is to turn a profit, you start to lose some of the things that made open-source projects thrive," Galstad says.

When I read that, I had to sit back and think about it. Is that really true, or just romantic thinking on somebody's part? Are profit and open source ideals really in conflict?

Lacy continues:

Galstad is one of the people feeling the tug [to accept outside money]. He says he has received dozens of unsolicited calls from venture capitalists interested in taking a stake in Nagios. But he isn't tempted. He figures that if he takes venture money he'll have to start looking for a way to cash the investors out, probably through a sale. That could drive him into the hands of a big software company, where he may not be able to pursue the projects he wants. "Once you incorporate, you get shareholders who want to see their investment turn a profit, and all of a sudden the goals and ideals of the project are going to change," he says.

I thought about it some more.

Eventually, my conclusion was that no, profit and open source ideals aren't fundamentally in conflict. Maybe that's just situational ethics since my salary is being paid by a commercial open source company, but let me argue that I'm not that much of a sellout.

Sure, I agree with Galstad, quoted in the article, losing control of your work environment and being forced to work on projects you don't enjoy sucks. But that's more a matter of maintaining control of your work environment, unrelated to whether you're working on open source or not. If you want to be a one-man-shop, in total control of your destiny, you can do that whether or not you're writing open source. Conversely, I think that open source and the profit motive aren't necessarily in conflict.

The fundamental issue for open source is how people can give away code while avoiding going hungry. (We've been getting a lot of those questions from everybody lately about Vyatta. While we have a plan for that, we just aren't sharing it yet.)

As an individual, it's more difficult to write open source and avoid starvation. If you, yourself, spend all your time coding, it's very difficult to make a revenue stream off your work if you're giving it away. If you were to try to monetize it yourself, you'd be constantly stopping your development to answer support calls, provide training, or whatnot. So what do you do? Well, if you want to eat, you find another guy to answer those calls while you do the coding, and you form a small business. And as soon as you do that, there's a profit motive involved. If you do your jobs well, people will love your software and your business will grow. You'll be able to hire more coders and more support guys to help with more features, etc.

This is how Linux worked. Now it's a multi-billion dollar ecosystem supporting all manner of businesses. But all during the development of that ecosystem, people still had to eat.

I think the key point to keep in mind with open source isn't whether there is a commercial interest involved in its production. The key point is whether the code is open and free and whether there is a healthy community thriving around the code-base. If the code is free, then the community always has the ultimate power. They can fork the code, develop derivative versions, etc. As long as the commercial entity is providing value, they should be encouraged to earn a profit and support the work.

I think one only needs to look as far as Red Hat for a good example. By most accounts, Red Hat is a great open source citizen. Its coders work on a variety of open source projects, including the Linux kernel, and contribute all that code back into the community at large. Red Hat is also a public corporation (of which I'm a shareholder, in fact). Personally, I love the fact that Red Hat make a profit. I think they provide a good service for their customers and they use that revenue stream to fund further open source development. If Red Hat ever becomes a problem child, it's code can be incorporated into other distros immediately (if it hasn't been already).

So, back to the question: is open source in conflict with the profit motive? I say not, since the essence of open source is choice and the ability to innovate on the open codebase. As long as that's preserved, let anybody and everybody supply additional goods and services to the community, earning whatever customers are willing to pay. The bad businesses will die; the good ones will thrive. And the good ones will plow that revenue back into the creation of more great open source software.


The Vyatta crew has been hunkered down in the San Mateo bunker the past few days watching the web stats ratchet upward. Remeber that old dot-com Superbowl commercial with the e-commerce site designers looking over each other's shoulders as their web site went live, registering first a single order, then a few more, then an eventual torrent of traffic. It's been a bit like that for us the past few days.

One of the advantages of picking a slightly obscure name like "Vyatta" when you create a company is that you have virtually no hits on Google. Before we registered the domain name, the top hits were for a Sanskrit dictionary defining the word. This means that as the company grows, you can use Google as a rough indicator of how effective you're being at getting the message out and growing the community.

Let's put some real numbers around this. Last Friday morning, February 24, searching for "vyatta" on Google yielded 157 hits (which was actually up from the something like 120 we had earlier in the month). I just checked this afternoon and Google is telling me 76,500.

Now for reference, searching for "linux" returns 433,000,000 hits. But 76,500 isn't bad for a single week.

Thursday, March 02, 2006

Product release cycle decoded

Allan came in this morning and told me about his odyssey working with one of our systems that he thought had crashed. Upon reflection at lunch, I started to think about the phases of product maturity. Now you have to understand that I have been doing engineering and product management at many companies, from small to gigantic for more than 15 years. I wake up with night sweats muttering "alpha, beta 1, beta 2, release candidate 1, release candidate 2, FCS."

The reality is that those terms are all code words, shorthand if you will, for the following:

  • Alpha -- It crashes when doing the most basic tasks, or every five minutes, whichever comes first. When interacting in a large system, if something happens, it's definitely our fault.
  • Beta 1 -- Somebody, long ago, came up with Beta 1 to fool management into thinking we're making progress. It's really still Alpha and thus it's still our fault.
  • Beta 2 -- It finally doesn't crash when doing the most basic tasks, but still crashes at least daily. It's probably our fault.
  • Release Candidate 1 -- Things are finally starting to stabilize. Maybe it's not our fault.
  • Release Candidate 2 -- This is here to fix all the stupid stuff that inevitably slips through RC1. It's probably not our fault.
  • FCS -- Ship it. What bug?

Allan's experience suggests we, Vyatta, are crossing that line between "probably our fault" to "maybe it's not our fault." That's a good sign. Progress happens in slow, plodding steps.

Fax machines are stupid

If you'll recall, when Allan and I had our Office Depot experience, we were returning an all-in-one printer/scanner/copier/fax product. Instead, we bought a standard network-ready color laser printer, and a really, really, really cheap all-in-one printer/scanner/copier/fax, just to serve as a low-volume copier/scanner/fax machine. (Aside: the two machines were actually $1 cheaper than the expensive combo machine.)

So, of course the inevitable day arrived when I had to send my first fax. Like fax users everywhere, I dutifully printed out the document I wanted to fax, signed it, made a cover sheet, then stood in front of the fax machine trying to figure out how this thing worked. Does the paper go in face-up or face-down? Do I have to dial it before I put the paper in, or after? Do I dial a "9" to get to an outside line?

After about 10 minutes of being thoroughly unsuccessful, Allan walks by and informs me that he doesn't think we have the analog line yet installed to connect the thing to the PSTN. Doh!

I think fax machines are about the dumbest pieces of office equipment imaginable today. Do-dads like paper clips, staples, and Post-It's have earned iconic status, finding their niche in the office ecosystem forever. Fax machines, on the other hand, are Neanderthal throwbacks to the period before the modern Internet. The problems with fax machines are numerous:

  1. Why do we always print out whole documents and fax around unchanged sheets of paper all the time? I have gotten into the habit of just faxing signature pages, trying to do my little bit for the environment.
  2. Why doesn't the fax machine generate the cover sheet automatically? It has the local number, the remote number, and the number of pages. I suppose you'd need to tell it the name of the person to deliver it to, but for routine faxes between two people, that could easily be on a speed-dial list.
  3. Isn't email simpler? Given that virtually every page that gets transmitted between fax machines originates in a computer first and is then printed, can't we just start using digital signatures to sign documents and then email them?
  4. Fax machines frequently have to have their own dedicated lines. They are analog devices, so your alternatives are to buy a dedicated analog interface card for your PBX (expensive even with Asterisk if you have mostly digital handsets), or get a dedicated analog line from the PSTN.

Of course, we went the route of getting the dedicated PSTN line. And of course we're still waiting for that to be installed...

Fax machines are stupid.