Why There Will Never Be Another Red Hat: The Economics of Open Source (2014)

https://a16z.com/2014/02/14/why-there-will-never-be-another-redhat-the-economics-of-open-source/

by jadk157 • 8 years ago

153 74 8 years ago

a16z.com

de_watcher• 8 years ago

(2014)

koolhead17• 8 years ago

Reading his post makes me feel for Docker. :(

rumcajz• 8 years ago

The article argues that Red Hat's business model was selling a "slightly better version of Linux".

IMO they were rather in the business of being a professional scapegoat. Someone to blame when that scary open source thing goes wrong.

organsnyder2• 8 years ago

That's how they were used at my last employer. Yes, their OpenShift platform is very good, and they have a lot of good products in their ecosystem, but the biggest value they provided was a vendor to blame if things went south.

Maybe Red Hat and IBM aren't as different as we like to think.

freehunter5• 8 years ago

Exactly. The best reason to hire IBM is so you have a company who is big enough to sue when something goes wrong.

Angostura• 8 years ago

See also "a single throat to choke"

mixmastamyk2• 8 years ago

I like to think of it as liability insurance.

setquk• 8 years ago

I never understood that mentality.

Not once have I had an insurmountable problem with any Linux platform.

The same is not true with Microsoft, Oracle and IBM from experience, even with paid up top tier support. Even 60,000 employee defence companies have no real power. I've seen someone forced to throw a registry fix out to 5000 desktops (over hundreds of disparate sites) because MSFT wouldn't fix a fucking bug.

phendrenad2• 8 years ago

Buy Red Hat
Something goes wrong
Blame Red Hat
Keep your job

Sounds about right.

smackay• 8 years ago

The article touts the business model as the underlying failure of Open Source. However I think this is missing the forest for the trees. If you add up all the value that Open Source delivered by giving millions, yes, let's go for millions, of developers access to high quality tools that previously were only available as expensive, proprietary products then clearly the model has been a runaway success. The only lament would appear to be that the value was not captured by a single corporation.

I think the pendulum is now swinging back to the pre-PC days of large corporations, centralized control of technologies where only groups with large resources can participate. However that will only be a temporary phase until the next generation of mammals are able to come up with the tools to create competition once more and the whole cycle will repeat.

awkward• 8 years ago

It seems to be closer to saying that the Red Hat model - essentially being a professional scapegoat, with some consulting on the side - hasn't succeeded to keep up with the returns venture capitalism expects.

The flip side is that companies that take over the operations part of the value chain and not just corporate liability are fully able to compete with closed source companies, and that kind of hosted service is where corporate open source is likely to go.

vinceguidry• 8 years ago

The article didn't claim that OSS wasn't successful in the broader sense, just that in the business sense, the Red Hat business model has failed. It further points to a model that does work, SaaS.

dragandj2• 8 years ago

The article /does/ discuss that, and mentions that Open Source in the sense of software technology has been a huge success, but also that as a /business model/ is far, far, behind the proprietary models. It also mentions that there are another OS-based models that are successful, such as SaaS, cloud, etc.

pjmlp• 8 years ago

Basically models where having partial access to the source code still doesn't provide a way to clone a product at zero cost.

apatters• 8 years ago

I think this is a great article with a provocative title. I don't know whether there will ever be another Red Hat or not, it's a meaningless question. But if you want to build a business on open source, this is a critical article to read because Red Hat's model is a hard way to do it and there are other models out there.

cyphar6• 8 years ago

Red Hat is a fantastic company, and a pioneer in successfully commercializing open source. However, beyond Red Hat the effort has largely been a failure from a business standpoint.

SUSE started around a year before RedHat, in 1992. Sure, we haven't had the same success and growth, but I would disagree that RedHat pioneered the model. I wouldn't even argue SUSE pioneered it (arguably it was Cygnus in 1989 -- though it was eventually acquired by RedHat in 2000). And even earlier you could argue that RMS pioneered it with his distribution of GNU Emacs in the early 80s (which had a description of the business model as the final chapter in the manual).

[I work at SUSE and am obviously biased, but it is a little bit annoying that people always ignore that we still exist...]

falcolas• 8 years ago

it is a little bit annoying that people always ignore that we still exist

Count me among them. I wasn't even aware that the SUSE Linux project was still a thing, let alone one backed by a business.

Marketing might not be a terrible idea at this point in time - the best product in the world won't sell if nobody knows about it.

beatgammit• 8 years ago

It's very popular in Europe, RedHat is very popular in the US.

I think the IBM acquisition is a fantastic opportunity for SUSE to start their engines here in the US. They have a fantastic value proposition: start with openSUSE Leap for free, then upgrade to a paid SUSE contract (can be done on place). This should appeal to CentOS and Fedora users who consider using RHEL, but don't because it's too hard to switch. They can also appeal to people who don't need the IBM integration but feel like they're not getting as much attention as before (priorities will likely change at RedHat). They can also appeal to developers with their rolling Tumbleweed release (which can also convert to openSUSE Leap/SUSE).

If SUSE had publicly traded stocks, I'd consider buying in right now as they have a really good opportunity over the next few years to steal customers from RedHat/IBM. They're big Linux contributors (Greg Kroah-Hartman used to work on the project), so they are _very_ attractive.

notacoward• 8 years ago

I think you're being a bit over-sensitive. They said a pioneer, which does not preclude other pioneers. Also, what's the bar for "successfully" commercializing open source? In a world where anything less than a unicorn is considered failure - and bear in mind the OP is from a promoter of that ethos - Suse's trajectory might not qualify. Even if it did, it would be one of the few exceptions to the general rule that commercializing open source largely has been a failure. Open source as a development method is alive and well, but as a business in its own right not so much.

P.S. Eight years at Red Hat, ending a year and a half ago. Glad not to be under IBM's inept thumb as my erstwhile colleagues soon will be.

erikb2• 8 years ago

Is anybody still using Suse besides some big three letter corps?

cyphar6• 8 years ago

Yes. https://www.suse.com/c/success/

NikolaNovak3• 8 years ago

FYI/FWIW, I think they were preferred for Canadian Government / Shared Services Canada deployments until a few years ago

madhadron• 8 years ago

I was delighted to find that SuSE has started using btrfs to do the same root filesystem snapshotting that Solaris does with ZFS! Suddenly it was back on my radar after not paying much attention to it since the 1990's (I went into Debian land and stayed there, pretty much).

h1d3• 8 years ago

You need to read more into the news. It's actually a reason to get unexcited about.

https://news.ycombinator.com/item?id=14907771

freehunter5• 8 years ago

Purely anecdotal and I've never worked for or on a team whose main responsibility was Linux administration, but at two places I worked in the past, our Linux guys didn't like SUSE. It wasn't a technical problem, and more marketing could probably help fix the problem.

I have a SUSE Linux certification, and at two companies folks from the Linux team commented that SUSE makes them uncomfortable because they're unsure of how to pronounce the name. It's like GIF or Porsche but worse. Half of them pronounced it Suzy, like the person's name. Some pronounced it closer to Zeus with a -e at the end. Some pronounced it "Soos" like it has a silent -e. They preferred Red Hat or Ubuntu not because of technical superiority but because they were embarrassed to attempt pronouncing SUSE. It's likely the only time they heard it pronounced was from coworkers who didn't know how to pronounce it either.

cyphar6• 8 years ago

There was a video explaining how to pronounce SUSE several years ago (both funny and a little worrying that such a video was necessary)[1]. It is pronounced strangely since it was a German name and so you have to put on your best German impression to get the "s" sound right.

All jokes aside, I think most folks would agree that SUSE's marketing really needs to be reworked very significantly.

[1] https://www.youtube.com/watch?v=N4lFHP-UxTk

HeyLaughingBoy• 8 years ago

I get that it's a German company and that they can name it what they want, but I also think it's part of the problem. The only reason I know the correct pronunciation is because I was in a class with a Swiss person who made reference to it.

More generally, it's an example of why I think that, for US companies, "weird" names are a bad idea. People don't like to sound stupid and something as simple as a name that's difficult to pronounce, or sounds odd can limit adoption of your product.

jcbrand• 8 years ago

Ubuntu's pronunciation isn't exactly clear to English speakers either.

People needed to be told that it's Ooboontoo and not You-ban-too (or whatever else people came up with).

kbutler• 8 years ago

Are there still people who pronounce linux lie'-nucks?

And Linus Torvalds[1] originally pronounced it with a bit more ee and oo (almost lee-nooks) than he does now[2] (lih'-nucks).

[1] https://youtu.be/uK0xXFZtJ8Q

[2] https://youtu.be/5IfHm6R5le0

And gnu and gnome and Debian and nginx and SQL and GIF and ...

The industry is full of made-up words with non-obvious or disputed pronunciations.

SmellyGeekBoy• 8 years ago

SUSE makes them uncomfortable because they're unsure of how to pronounce the name.

Amusing the irony in this considering most people pronounce "Ubuntu" and indeed "Linux" incorrectly.

FWIW, it was the first distro I ever tried back in 1999 and I always pronounced it "Suzy" but have heard a few variants. Everyone pronounces different words differently so it doesn't really matter.

h1d3• 8 years ago

Isn't it pronounced 'sues'? I doubt it's the name but it always makes me wonder what Suse's position is. Is there anything specifically better than Redhat or Fedora?

It may have bigger market in Europe but when I heard it sets the default file system to btrfs when others are jumping off that ship, I wonder what their plan is.

bradknowles3• 8 years ago

I’ve been a Linux user and Administrator since late 1998, and various forms of *nix before then.

I’ve known of SUSE for a long time, but it was never the name that stopped me from doing anything with it. I just never used a product or worked on a project where it was used, and so I was never exposed to it.

TIL, I’ve been pronouncing it wrong for probably something close to a decade.

bewo001• 8 years ago

Wasn't there a close collaboration between IBM and SUSE?

wmf• 8 years ago

Back in the early days of enterprise Linux, IBM wanted to avoid Red Hat becoming a monopoly so they backed SUSE (and later Ubuntu) quite a bit.

cyphar6• 8 years ago

IBM uses several distributions. I know that both RedHat and SUSE (and probably Canonical but I'm not sure) have IBM departments as customers.

geraldbauer• 8 years ago

Just a reminder: Never say never. Of course, there will never be another Amadeus Mozart too (in that sense of the word). Greetings from Austria.

bryanrasmussen• 8 years ago

https://medium.com/@alexfclayton/elastic-ipo-s-1-breakdown-1...

laurentb• 8 years ago

an interesting, yet somewhat unrelated thing that came through this article was the market cap's graph of some big names like Microsoft, Amazon,Oracle, VMware and Red Hat. (https://a16zpeter.files.wordpress.com/2014/02/peterlevine1.p...)

it mentions in 2014, their Market cap was: - Microsoft: ~$300B - Amazon: ~$160B - Oracle: $160B - VMware: $40B - Red Hat: $10B

I'm not sure how true this was back in 2014, but the current market cap 4 years later for these companies is as follows:

- Microsoft: ~$833B (177% increase) - Amazon: ~$803B (400%) - Oracle: $181B (13%) - VMware: $60B (50%) - Red Hat: $30B (200%)

it's incredible to see in the same period of time just how massive both Microsoft and Amazon have been able to perform (market wise at least) and specifically the hockey stick market cap of Amazon in that time frame...

mixmastamyk2• 8 years ago

A portion is investment in the stock market in general.

BenoitEssiambre• 8 years ago

The problem of investment in the world of open source software is a difficult one. What company wants to pay for improvements when they know competitors are going to get all the benefits of these investments too. It's one of these bad equilibrium, prisoner's dilemma, kind of situation.

In practice, the funding often comes from those people or companies that more urgently need the improvements or those that are pushed by end users who are willing to pay a premium to avoid lock-in. There is a longer tail of users who free ride and pay nothing for improvements.

Even businesses that support open source, at any moment in the future, have strong incentives to use network effects to steer their customers into more subjugated positions and adopt proprietary non standard or non open systems, or to sell themselves to businesses that will do that.

The coagulation of tech companies in a few large players is a sign that this may be happening right now.

I wonder if there are ways to promote wider investment in open source and open standards technology that benefit more people and get us into better economic Nash equilibria.

zackbloom• 8 years ago

That article never actually explained why Red Hat was successful where others weren't. It got lost on showing how _not_ successful they were, even as it said how exceptional their results were.

robla• 8 years ago

Agreed. My theory: the market will only sustain a finite amount of revenue by open source software creators. That amount grows 20-30% per year as competing business models collapse and as customers learn the value proposition, but there is a ceiling. Red Hat has been brilliant about identifying the market segments that were ripe for an open source player, and entering those markets (often via acquisition).

There might be room for another Red Hat now that Red Hat will no longer be independent. IBM may have just legitimized Red Hat as a top-tier player.

That is, of course, if IBM still has the clout and capitalization to be considered a top-tier player. 20 years ago, it would have been considered funny to say "IBM is great and all, but they're no Apple". My oh my have times changed.

nimblegorilla• 8 years ago

Seems weird to claim Red Hat failed. Even if their revenue is tiny compared to Microsoft or Amazon they are still on a chart measuring market cap by billions of dollars.

dredmorbius• 8 years ago

It's worth nting that the exemplar of the prorietary shrinkwrap consumer softare market is also a one-off: Microssoft.

There is no other pure-play software company of comparable market success.

The consumer personal-computing operating system market resisted incursions by CPM, other DOS vendors, other GUI alternatives, x86 commercial unices, Linux, FreeeBSD (and others), and BeOS. Apple's share remains a scant fraction.

Likewise, the Office applications, and small-business server market remains dominated by Microsoft, especially file, print, mail, and directory services (AD).

Contenders in the software space tend to be small, heavily reliant on professional services (IBM, Oracle, Salesforce, consulting firms), specialised (Adobe, AutoDesk)<, or hardware-integrated (Apple, also Sun/Oracle and some others).

The SaaS market has impinged somewhat; Gmail, Google Docs, etc., but still only marginally.

This from someone whose watched and criticised Microsoft for decades, and strongly prefers Linux. I've not had a Microsoft desktop personally or professsionally since the mid-2000s, and even then it was under protest and largely avoided.

patmcguire• 8 years ago

Elastic, Mongodb, and Docker all have this business model. I guess the difference is in their case they developed it first and then open sourced - they've got a much more credible claim to superior support.

tschellenbach• 8 years ago

How is Docker doing?

Elastic and MongoDB also offer cloud hosted versions of their tech.

jondubois• 8 years ago

> We had made the product so easy to use and so important, that we had out-engineered ourselves. Great for the open source community, not so great for us.

I can relate to this. The other sad truth is that if you do not 'out-engineer yourself', then your open source project will likely not even be able to get any users to begin with; it's extremely competitive. It took me over 5 years of work; constantly refining my project in order to reach almost 5000 stars on GitHub and 50K downloads per week. Those are pretty big numbers for a framework targeted entirely at developers but in terms of monetization, it's barely enough to cover a single salary and my project is one of the lucky ones because it's back-end and it's somewhat more opinionated.

Front end frameworks and non-opinionated back end tools (e.g. the ones that just implement existing standards/protocols) have near zero monetization opportunities. For front end frameworks, you need a huge success like Vue.js or React in order to monetize at all. It seems that Evan You from Vue.js gets enough money to cover his salary - But Vue has over 100K stars on GitHub and is one of the top 10 repos in the world. If Evan was a musician or an actor, he'd have a multi-million dollar salary for sure.

nostrademons• 8 years ago

I see Red Hat's business model as a product of the times - there was this cheap, underutilized, but difficult-to-use competitor to the OS they really wanted to use, and a lot of corporations that really wanted to spin-up big IT departments and web services quickly. They filled the market need that existed in the late 1990s; it's unclear if this model fills any major market needs that exist now, when turnkey SaaS solutions exist for a lot.

This doesn't mean there won't be other successful open-source companies. It seems like each generation of open-source projects fills a different economic need, all specific to the business landscape at the time:

In the early-mid 2000s we had open-source as a retention bonus for good engineers and PR campaign for the companies involved (Hadoop, Nutch, Protobufs, Snappy, and other Google open-source projects, Cassandra, HBase, Hive, HipHop, and other Facebook open-source projects).

We also saw open-source as a means of securing cooperation between different corporate interests (WebKit) or securing critical partnerships & goodwill by assuring the source code would be available (Chromium, Android).

Starting in the early 2010s, we saw companies releasing an open-source product and then making money off a hosted version of it (Elastic, Datastax).

Nowadays in the crypto world we see organizations releasing an open-source software system, reserving some of the currency for the development team, and profiting off the capital appreciation of the currency. (Ethereum etc.)

yuhong• 8 years ago

A good example of some of the problems of the debt-based economy is Sun/Oracle.

Why There Will Never Be Another Red Hat: The Economics of Open Source (2014)

Related Stories