I'm really happy for nginx to get the recognition that it deserves and I'd love nothing more than see the people behind it being able to make a living doing nothing but nginx.
But when I read about funding, about founding a real corporation and everything else that's going on, I'm afraid that at one point, nginx will do a MySQL and become open-core.
It's not that I'm not willing to pay for software in general, but I would love to have flexibility when I'm chosing my infrastructure. If the need for an additional server arises, I want to just start it and not matter about buying licenses and entering keys.
This is one of the main reasons, I prefer my infrastructure to be open source.
"But, open-core is also open source", you might say, but none of these projects ever end up getting a good community behind them. Many talented people would not sign a CLA and they would not want to invest their time into something that's "tainted" by a "better" enterprise edition.
Also, often times, submitted patches would not get accepted because the submitted feature is conflicting with an enterprise-feature - maybe even just one that's planned for somewhere in the future.
So in the end most of these open-core projects are exclusively developed by the same people who do the enterprise version and have a vested interest in selling enterprise licenses, degrading the "open" edition to nothing more than a trial version.
I would hate to see this happen to nginx. Not because there are no open alternatives (there are), but because nginx is elegant, fun to work with, fast and, above all, stable.
Let's hope that a pure support-based revenue model is going to work out for them.
(edit: nginx is released under a very liberal license, so it's entirely thinkable that, nginx would be OpenSSH'd: OpenSSH AFAIR started as a fork of the last Free version of ssh. And now just look where OpenSSH stands compared to non-free ssh)
I would hope most of us would want Igor Sysoev to at least put $10-$100m in the bank. He has easily created that much value over the past few years for hundreds if not thousands of companies.
If it turns out that nginx becomes less usable, or the free version becomes too expensive, a competitor will crop up. But we should want people who are gods of open source to figure out business models that enable them to become richly rewarded, and not just make a living. Maybe that's open-core, maybe it's a talent/showcase acquisition by Google or Facebook (as opposed to Oracle), but we should want these guys to become quite wealthy to inspire people.
nginx is one of the best pieces of software ever written.
Igor deserves a truckload of money, and it's good to see him going out on his own instead of being acqui-hired and never seeing the light of day again.
I feel the same way. Nginx has been the little ngin that could, supported by a bunch of talented hackers dedicated to bringing something to the table that the big-time servers, even the open source ones, couldn't. Nginx's history of being the lightweight, faster competitor to Apache (a shining example of success in the open source world) practically defines its identity.
I just hope that nginx doesn't sell out. I don't think it will, because I can't see what it has to gain for becoming exactly like its competitors when it already has a substantial user base. Of course, as you stated, dividing the userbase between enterprise edition users and free edition users could cause some software development political issues that could stifle progress.
Nginx is one of those rare products where I felt no need to upgrade from the now-ancient 0.7.6 version because it just does everything I need well. Truly an amazing little piece of code, on par (in my books) with venerable oldies like qmail.
It's worth keeping an eye on the current development version goes stable. 1.1.4[1] introduced preliminary HTTP/1.1 reverse proxying which means persistent connections to the backend and thus less overhead.
That is also true. You can use that version for as long as you want! The company I work for uses a newer stable version of nginx in production, but I think I actually make use of a directive that was introduced after the 0.7 series. Even so, as a small company with only a few customers, we're okay sticking a little closer to the edge. At this point, there's no compelling reason for us to upgrade past where we are (I think 0.98) unless I want bug fixes for things bugs we haven't seen.
I almost agree, i.e. second-level-legacy 0.7.x line is good enough for most of web needs, so indeed there is no really strong urge to upgrade to the current stable, though some nice features came later. What I do not agree is keeping really old 0.7.6 instead of upgrading to safer and more stable 0.7.6y (and if you do, going to the latest 0.7.69 is obviously the best solution).
Upgrading to at least the latest minor version (0.7.6y as you say) may indeed be a good idea since there have been some security vulnerabilities discovered e.g.
My computer club has used nginx since it first came out in 2004. We also had a bit of contact with Igor to get it running on Solaris. Since then it has been a very reliable piece of web infrastructure for us, and I wish Igor and the rest of the team the best of luck. I have high hopes that really great things will come out of this.
> (edit: nginx is released under a very liberal license, so it's entirely thinkable that, nginx would be OpenSSH'd: OpenSSH AFAIR started as a fork of the last Free version of ssh. And now just look where OpenSSH stands compared to non-free ssh)
Wasn't nginx imported into OpenBSD base not too long ago? I imagine worst case scenario they just keep maintaining their version (like they did with apache).
However, I am hopeful for Igor. I have used (and benefitted from) nginx quite a bit. I wish his endeavor the best!
If that happens it is pretty safe to predict that nginx will be forked immediately. Once it gets forked it can get community support without anyone worrying that they are supporting a paid version for free, and the enterprise people will have absolutely no control over features. Mysql got forked. And if you say that the mysql forks do not get sufficient community support (actually I do not know if that is the case) the reason is more likely the presence of a better sexier alternative (postgre) than anything to do with Oracle.
It is very hard to lock down free software if it is something popular or needed.
I don't trust this model, since when building a company that's based on open source software you have just two business models available:
1) make money from services.
2) make money from products.
If you pick "1" you can do a good, very small company, to get your bills payed, but service companies are hard to scale, and unlikely to grow their value in a short time. This is not what your VCs have in mind, basically.
So you end doing "2" if you are VC-backed. With "2" the only way to get money is to either close part of the software, or to find a different way to provide some product value to the user if the use will pay. And this will make the project weaker one way or the other in the end.
That is what I'm doing with Redis. If the big company funding the project is as polite as VMware you end with a lot of freedom but the money to just focus only on your project (and not just you, VMware just hired Pieter Noordhuis for instance).
Only disadvantage is that the lead developers are unlikely to get rich, as they are payed to work with a good salary but this is not like a big exit for a startup. But my grandfather and my father always got payed to work, so I trust this model, and I'm not seeking richness, so it's the perfect model for me, but not for everybody.
FYI, the correct spelling in this context is "paid"; the spelling "payed" is only used in a couple very obscure cases which you'll probably never encounter.
(I'm only mentioning it because I'm pretty sure you're not a native English speaker, and I figured you might like to know.)
> service companies are hard to scale, and unlikely to grow their value in a short time. This is not what your VCs have in mind, basically.
It depends on your services I suppose, but I don't think it would be especially hard to generate revenue by selling Nginx training and support. A conference, seminars, books, support contracts, etc.
Sadly just having an official company behind it will be enough to get approval to use Nginx in a lot of enterprises--sign a fat support contract so you can cover your ass and not think about it again.
>Sadly just having an official company behind it will be enough to get approval to use Nginx in a lot of enterprises--sign a fat support contract so you can cover your ass and not think about it again.
I foresee that plugin writers and source contributors will slowly creep towards the consulting and support model, while the
'community' or 'core' version users will get scarcer documentation. That's fine by me, but signing that support contract will probably mean paying 2 grand for getting a tailored nginx.conf and an SOW document.
That's a good business opportunity overall, but at what cost to the community ?
There are hardly any source contributors to begin with. Igor is notorious for being very slow to accept patches for anything but simple bug fixes, it's only recently that other people have gotten access to SVN, or that the SVN was even public.
Meanwhile the nginx wiki is completely community run and maintained. The nginx company would cripple itself if it ostracised the community behind it.
PuppetLabshas found that #1, being a services company, really sucks. There's less of an impetus to do a good job at maintaining your community since you're directly competing with them. The realization that their documentation is somewhat sub-par, and that they haven't really built the community that Chef has is why they're shifting to being a products company with their hosted offerings.
Perhaps Nginx can learn from that, and maybe build out some hosted products, like selling self-scaling nginx ami clusters on ec2.
Webservers are a hard business to make money in. Look at Zed Shaw, he built the app server that made Twitter possible, and I'm quite sure he never saw a dime out of it, though it increased his reputation. Likewise Igor hasn't seen money out of nginx yet, but he's built software that also helps enable hundreds of millions of dollars worth of transactions per year of business. He deserves the opportunity to profit on the technology he's shared with the world for free.
nobody can stop you from doing a bad service company, but no one can stop you from doing a bad product company as well IMHO. The difference from an open source point of view is that the community can write better documentation, but can hardly fight you if the open source main developers are trying to push forward a business model where part of the code is closed, or there are commercial plugins that are very useful, and so forth.
With "2" the only way to get money is to either close part of the software,
Or, as in our case (x264), you keep everything open, but open source version is GPL, thus limiting its possible use in proprietary software packages, whereas the commercial version is not (despite them being line-for-line identical).
This only works if the project's license is copyleft, of course.
That's what contributory license agreements are for: any project that plans to follow such a route needs to get all their developers to sign one in order to get the rights.
In my experience this was some effort, but not overwhelmingly so, IMO. This would probably be more difficult for a larger project, but other projects have done similar things (e.g. VLC's planned relicensing of libVLC to LGPL).
1. create something of value, but serious value
2. generate traction, but in a big way
3. go after capturing some of that value.
I like how they move to 3 only when 2 is fulfilled, like half the internet uses their product.
This seems to go against the current trend of first deciding to "do a startup", then figuring out what to do. Reminds me of something PG wrote in an essay about the craigslist way and staying "upwind" of profits.
I think this is great news. I remember the excitement of first discovering Nginx back in 2006 and realizing that the world's most popular web server (Apache) was based on inferior technology. Here was this unknown proxy server written by one guy in Russia, but inside the well-structured, completely undocumented code base was the power to serve almost all the world's web traffic. And he was always more than happy to respond to dumb questions on the mailing list and incorrect patches to his personal account.
"Going corporate" will make Nginx feel like less of a hacker's playground, but I think that's OK. Like the Linux kernel in the late Nineties, Nginx has grown up and is stable enough for huge companies to rely on for critical functionality. Now that Nginx has a marketing budget of more than $50, it will be installed on many, many more servers and generally make the web faster and more reliable for everyone. That seems like a great outcome to me.
"Today NGINX powers over 40,000,000 domains on the Internet"
I'd love to know where they got this stat from. I've indexed 130 million domains and 3.1 million of them use Nginx. Don't get me wrong that's an amazing total but if by domains they mean registered web domains (example.com etc..) then I'd love to know how they came to it.
This is really awesome... I didn't even know they had a real company, I thought it was just another open source entity. Can't wait to see the awesome software they produce!
But when I read about funding, about founding a real corporation and everything else that's going on, I'm afraid that at one point, nginx will do a MySQL and become open-core.
It's not that I'm not willing to pay for software in general, but I would love to have flexibility when I'm chosing my infrastructure. If the need for an additional server arises, I want to just start it and not matter about buying licenses and entering keys.
This is one of the main reasons, I prefer my infrastructure to be open source.
"But, open-core is also open source", you might say, but none of these projects ever end up getting a good community behind them. Many talented people would not sign a CLA and they would not want to invest their time into something that's "tainted" by a "better" enterprise edition.
Also, often times, submitted patches would not get accepted because the submitted feature is conflicting with an enterprise-feature - maybe even just one that's planned for somewhere in the future.
So in the end most of these open-core projects are exclusively developed by the same people who do the enterprise version and have a vested interest in selling enterprise licenses, degrading the "open" edition to nothing more than a trial version.
I would hate to see this happen to nginx. Not because there are no open alternatives (there are), but because nginx is elegant, fun to work with, fast and, above all, stable.
Let's hope that a pure support-based revenue model is going to work out for them.
(edit: nginx is released under a very liberal license, so it's entirely thinkable that, nginx would be OpenSSH'd: OpenSSH AFAIR started as a fork of the last Free version of ssh. And now just look where OpenSSH stands compared to non-free ssh)