Some of the industry's smartest prognosticators, like Redmonk's Stephen O'Grady and inveterate free software advocate Glyn Moody, have questioned the likelihood of a billion-dollar open-source software vendor. But in its most recent earnings call, Red Hat, the industry's leading open-source company, promised to surpass this goal …
Open Source as a profitable model?
I'm not sure why the author thinks that Open Source can't be an effective model for creating a billion dollar corporation and that RedHat is a one off.
First, I have to disagree about RedHat being a one off. Sure other companies can get to a billion dollars revenue point. However, like Michael Jordan or Lebron, the number of companies that can achieve this level of revenue are few.
In part, its timing and its what the company has to offer.
If you have the right product/service mix along with timing of market entry, its possible.
Cloudera has the potential of hitting this level because they gained market presence and established their dominance before others entered their niche.
In addition, one has to gauge the level of disruption the technology creates. Linux was a disruptive technology. Hadoop is. CRM? BI? Not so much.
There also has to be a compelling reason why one is willing to pay for support while the actual core code is available for free.
Find the right mix of these and you'll have another billion dollar company in the making.
But isn't the biggest reason that Redhat is so far out in front, is because they have a larger product portfolio? As you said, MySQL was growing quite nicely. If MySQL had 9 other products besides, just a database, they be at $1B too. Redhat was smart to acquire all sorts of components of the stack. And the value of a full stack is greater than each individual piece. And Redhat is still growing their stack. Their devs have their fingers into the AMQP (messaging) standardization, which is a huge enterprise cash cow.
Canonical doesn't have enough pieces in house to compete. It doesn't seem like they even want to compete actually.
Novell was split between developing their Linux stack and keeping the Groupwise & Netware legacy alive, and failed at both.
It would be nice if this story compared the miniscule peanuts earned by Mandriva, and the relatively few commercial-acting Linux distros. Actually, nix that. Searching on Mandriva, I found out I'd been more than just a few months behind. More like 10 months behind:
Looks like RH have virtually NO competition from other Linux other than possibly Canonical funding Ubuntu into or deeper into corporate space.
In general you make some good points. However I do take issue about the bit of 'free loaders like CentOS'.
When RH changed their patch release structures it wasn't to get at CentOS or Scientific Linux. This was aimed right at Oracle. AFAIK, CentOS has come out and said it won't matter to them.
Again AFAIK, CentOS does contribute patches upstream and as I understand it, have a fairly decent relationship with RH.
Oracle on the otherhand is the freeloader here with their so called Unbreakable Linux.
On Commercial terms, the RH model as you said, shouldn't work. Others have tried and failed yet they go from strength to strength. Why? You did hit the nail on the head. The 'Peanuts'.
I use some very closed source software on Linux every day. It costs us $60K a year in software maintenance. Paying RH $1K is neither here nor there in comparison.
… is still around (now running LSE, for example). And I am glad it is (although I work for RH).
"More recently, the company has made changes to how it distributes source code in an effort to bloody free-riding competitors like CentOS."
I disagree. This will have almost no impact on CentOS at all. CentOS and Scientific Linux make essentially no changes to the RHEL sources and don't have a support based business model. How the patches are bundled in the rpms is just about completely irrelevant for those distributions. It DOES, however, affect Oracle's ability to support their RHEL build. They have a support based business model around it and are commercializing their freeloading. I rather doubt anybody in the open source community will shed many tears over that having become more difficult.
Centos wasn't the target
Oracle on the other hand, they were. When they take Red Hat's work, and then rebrand it, before reselling it for less, that's what ate into Red Hat's margin. Not Centos. It's not like there aren't other free versions out there.
If they really wanted to do it, the option would be to stop distributing the SRPMS to all and sundry. They're under no obligation to give them to anyone but their customers.
The Company I work for uses a lot of RedHat products, and I have regular contact with representatives from RedHat. They have said more than once that CentOS is viewed quite favourably, and I would say why not? Small companies might go for CentOS, as there are no immediate costs apart from employees to support it, but if said company grows, so will their server infrastructure, and then things start getting more complicated. At this point, they will seriously consider proper support for their servers.
CentOS is RedHat re-packaged, and a transition from CentOS to RedHat as a company grows would be very easy. Any staff dealing with CentOS will have no problems dealing with RedHat and vice versa. They essentially feed into each other, and as noted in the article, RedHat is much more interested in large companies than they are in small ones or individuals.
RedHat view Oracle as both good and bad (they call Oracle a frenemy). Many companies use RHEL with Oracle databases, and so RedHat provides support for this in a big way. However, Oracle has also made very hostile statements toward RedHat and followed through on those statements. They re-package RHEL and sell its support for 1/2 of what RedHat does. But as the saying goes, you get what you pay for. Oracle support is not know for being good. Oracle is actively trying to get people away from RedHat support onto Oracle support with their take on RHEL.
This, I believe, is the background for making things harder for competitors like Oracle and SuSE, but more specifically Oracle. SuSE at least puts source code back into the community.
Mozilla no comparison
They live and die with the grace of google, that with the "scented search" revenue being their mainstay and all that. red hat has a much more solid base of revenue. And more wiggle room for growth and so on.
I don't think red hat targeted CentOs, and the CentOs guys don't seem to care much. It's oracle and novell who're undercutting red hat on their own business model with their own support materials. That's what drove the change, as reported by Cade Metz right here on el reg. CentOs simply isn't in the support business. People "switching" from paid to unpaid means they lose support, it's as simple as that. Matt, are you asleep or something? Did you forget to pay attention to el reg?
You know, it seems to me that oracle, re their kicking of hp in the nuts, is leaving some market opportunities for a red hat style support model. For example something based on PostgreSQL. hp might even help with the getting started. If you do run with that idea, I'll take a couple percent shares as thanks, thanks.
Appreciated the article
I'm not too familiar with the open source corporate space and found it a good primer on the company.
It looks like Red Hat know what they're doing. Over the past year, their share price has risen significantly faster than the rest of the market, it's on par with Apple.
But I'm looking at the financials, and their P/E ratio is sky high, in the 80s, how is that sustainable?
The obvious counterargument is...
If Red Hat is a "one off" because no other Free Software companies have achieved anywhere near the same level of success, and this is therefore an inherent failure of Free Software which Red Hat has miraculously managed to overcome ... then surely that means proprietary software has exactly the same inherent failure, because Microsoft has a 90% monopoly on the desktop, and no one else has ever come close to achieving the same thing.
Hmmm, I think I may have exposed a slight flaw in Matt's logic.
The reasons companies succeed or fail are many and varied. Microsoft's success was the result of decades of dirty dealings that secured a monopoly. Red Hat's success was simply because it's a competent business, and because it specifically targeted a segment of the industry not locked in to Microsoft's monopoly. In other words Red Hat chose a battle it could win. Other vendors do less well, irrespective of their licensing model, because they're either less competent or foolish enough to target Microsoft's impenetrable monopoly.
It's not exactly rocket science.
Of course, Microsoft's monopoly is now being seriously challenged via the back door. No vendor stands a chance of tackling Microsoft's desktop dominance head-on, but they may not have to, since consumers now seem to be engaged in a bit of a paradigm-shift to mobile devices and Cloud services, making the "desktop" less and less relevant, and thus Windows less relevant too, along with the proprietary licensing model it promotes.
Microsoft has demonstrated time and again that it clearly isn't ready for that paradigm-shift, that is has trouble keeping up with rapidly emerging technologies (and thus markets), and that it wields no significant power outside the desktop. Companies like Red Hat and Google ARE ready, though, and have been for years. The primary reason for this is Free Software, since it is inherently flexible enough to rapidly adapt to changing conditions, both in technical and licensing terms.
This then represents another opportunity to circumvent Microsoft's monopoly. The question is, will anyone (apart from Google) attempt to capitalise on it?
Red Hat didn't succeed BECAUSE of Free Software, it succeeded in SPITE of Microsoft's monopoly, USING Free Software. It's a subtle but important distinction. The licensing model is irrelevant when one is faced with a closed-shop market. One either supports that market, foolishly opposes it, or finds a way around it by exploiting a niche segment, then working to expand that niche.
That isn't a reflection on licensing models, it's simply a reflection on the corrupt state of the desktop market.
Paris, because she knows a good opportunity when she see one.
You gotta love naive capitalists. Tell us Matt, why are Billion dollar companies a good sign? You'd think that since free and open source represents a more distributed model of development and choice in support, that you'd naturally have a more distributed market place.
This is perhaps more interesting then journalists who have been spending way too much time speaking to CEOs and economists and much less time speaking to small businesses.
Why? Are you daft?
Only companies that make money can provide those things called jobs. Talented developers and DB people sitting around working for free doesn't help anyone in the long run.
Oh, but I disagree :)
Redhat has taken an awful long time to get to the $1bn mark. Novell could have made SuSE into $1bn but they took some funny turns because of an over-focus on the US market with what was an essentially European product. Alfresco making it to $100m on a single product tells you a multi-product company can make the $1bn no problem. And these are both business-focused companies, while any mobile phone company can tell you the consumer market is where the monies are at.
Frankly, look at it this way; I served on a panel at the IT Leaders Forum (Windows 7 migration, and yes I was in a minority) yesterday and watched a lot of people committing to spending a lot of money and time to keep their companies locked into a monoculture; one has to assume their households are equally proprietary in their computing solutions. Awareness of the options remains absent in mainstream UK; once awareness rises, the opportunities to provide services will be manifold.
Most open source business is through mixed as opposed to pure plays
I suspect the business done by selling open source support is a small fraction of the business done where software is a cost centre as opposed to a profit centre. E.G. hardware vendors don't sell hardware without software, and for bespoke software and service contracts it is cheaper to reuse existing available software than create what is needed for the customer and contract entirely from scratch. Open source software is by definition more readily reusable than closed source, due to contractual and transaction costs and licensing impediments.
IBM does many times the business done by Red Hat on the back of open source software. But IBM has a much more diverse business than Red Hat. Both IBM and Red Hat have excellent reputations as free software community code contributors.
You're A Winner
You appear to be one of the few people commenting that grasped the fact this article was about what Matt called "pure-play" OSS companies.
I've never understood the viability of a pure OSS model, but I can definitely understand OSS as an incremental addition to a traditional model.
CentOS and RHEL
"an effort to bloody free-riding competitors like CentOS"
Red Hat loses money to CentOS in two ways: people too cheap to pay for RHEL (and hosting providers would be the top of that last) and in training revenue (because so many RHEL sysadmins have learnt their trade via CentOS rather than Red Hat branded certification training).
I don't think that hosting providers are a market Red Hat is interested in. They'd have to price their product so low that it may not be worth Red Hat's time. Nor would Red Hat be interested in providing the sort of support dearest to hosting providers (eg, just how hot can I let this server get).
The training argument cuts both ways. The ready availability of deep RHEL/CentOS expertise gives Red Hat a significant advantage. CentOS also allows for zero-paperwork, zero-approval projects in large organisations, which can lead to uncontested RHEL sales when that skunkworks project succeeds.
Finally, CentOS helps RHEL users. Red Hat has no interest in supplying niche software. Niche software developers can't afford RHEL. So RHEL customers with niche software needs can meet on the CentOS middleground: the developer using CentOS, the customer using RHEL. There are quite a few CentOS-based RPM repositories for various fields of science, electronics etc.
- Tricked by satire? Get all your news from Facebook? You're in luck, dummy
- Google straps on Jetpac: An app to find hipsters, women in foreign cities
- Updated Microsoft Azure goes TITSUP (Total Inability To Support Usual Performance)
- The Return of BSOD: Does ANYONE trust Microsoft patches?
- Munich considers dumping Linux for ... GULP ... Windows!