Hacker News Reader

Canonical arrived and proved that yes, there is room for another profitable Linux vendor.

Cananical only had $65.7 million in revenue in 2013, and made losses of $21.3 million[1].

I think that proves his point rather well.

So Redhat doesn't have the resources for development that Microsoft has and therefore will produce an inferior product? Have you seen Windows 8? Vista? Money does not guarantee success.

Microsoft's product development problems are well documented. Their money making capacity is equally as well known. Their last quarter broke revenue records for them[2], helped by growth in both enterprise and cloud.

Even without the same resources, the main Linux kernel has an enormous number of top-talent developers working for free. This is a serious economic advantage over traditional closed source companies.

There are a number of unpaid contributors to the kernel, but the vast majority are paid by companies. 83% of kernel 3.10 was funded by a range of companies.[3]

Why on earth is the article comparing RedHat to Amazon? One is a Linux distribution vendor, the other is a global internet shopping mall.

Are you trolling? The author posted a graph showing how AWS (Amazon web services) has more than double the revenue of RedHat.

That's why other operating system vendors are (eg, Microsoft) are pushing into the cloud space, and why RedHat's lack of a coherent story on cloud is a big problem for them.

[1] http://arstechnica.com/business/2014/01/ubuntu-maker-boosted...

[2] http://www.pcworld.com/article/2055455/record-first-quarter-...

[3] http://arstechnica.com/information-technology/2013/09/google...

A lot of Microsofts growth in recent years has been in enterprise tools, not operating systems.

Amazon do not break out their AWS revenue from their main revenue, hence one of the reasons it's a dodgy claim. Not to mention the fact that the article included the total Amazon revenue in the charts for no good reason. Likewise for Microsoft, who have multiple large revenue streams in the form of Office, etc...

Not to mention Oracle, who are a database company that dabble in Linux (I've never actually met anyone who uses "Unbreakable" linux - yes, yes, anecdotal I know).

When you remove the specious data points from the graph, you're left with RedHat on it's own. Even if you include the AWS estimate, you still have a much more reasonable graph that indicates RedHat competing in the same ballpark as AWS.

A lot of Microsofts growth in recent years has been in enterprise tools, not operating systems

That's exactly the point (?!)

I think you are looking at Redhat, and seeing them as a successful operating system vendor.

Levine is looking at them as an enterprise software company who haven't grown revenues and/or profits at anything like the rate that other enterprise software companies have.

He then sees that there are no other large software companies which have a business model similar to Redhat.

He uses those two pieces of evidence to argue that there are better business models for Open Source than what Redhat does.

You appear to be picking at things like if Redhar competes with AWS. That isn't his point at all - he's showing that AWS is an alternate business model which appears to have grown revenues much more successfully than Redhat.

Fair enough. But the entire point of the article is that "There will never be another RedHat".

It's the title, it's presumably the main point of the article - and it's ridiculously over-generalising.

RedHat as a company, works.

The market share is still largely with Microsoft, thus still available to be stolen away. Therefore, there is still room for a huge amount of competition in the OS vendor/support space. I can't see how anyone can argue against that.

I'll grant you that perhaps RedHat could be making more money by spreading out into different sectors like SaaS and PaaS, but that doesn't mean that there can't be another RedHat in the market.

I guess my main complaint is with how this point is argued. The charts are deeply misleading for the point that is trying to be made.

>2b) Even without the same resources, the main Linux kernel has an enormous number of top-talent developers working for free. This is a serious economic advantage over traditional closed source companies.

I am not familiar with the actual org structure of Linux kernel contributors besides Linus being the leader but isn't it the case that most of the major contributors are in fact employed by companies (including Red Hat)?

In other words not devoting their time working for free...or am I off?

Many of the kernel developers are paid for their efforts by somebody, yes. Red Hat is the top single corporate employer, but their contributions are roughly 12%:

http://www.linuxfoundation.org/sites/main/files/publications...

From Red Hat's point of view, even though someone else is typically paying for the effort, code that's not from their developers is "free." (Well, Red Hat is also one of the top reviewers, so they're paying people to review the contributions of others, too. So it's a bit more complicated than that.)

Now, some of the companies on that list are in the same business as Red Hat (and some companies are in the same business but don't contribute significantly to the kernel, like Canonical). But many of them are in very different lines of business and view the Linux kernel as either a complement (that's why Intel is involved, for instance -- encouraging free operating systems to sell processors) or as an infrastructure piece for their products (yo, Google). So Red Hat is able to leverage the efforts of those companies to deliver a product for far, far fewer man hours than it would cost if they did all the work themselves.

RedHat contributes ~10% of total development:

https://lwn.net/Articles/451243/

so from the point of view of RedHat a large part of kernel development is actually free. Not actually 90% though since they don't care about all changes.

Getting unrelated and often competing companies to cooperate in creating the infrastructure layer of software is a huge success case of open source.

For RedHat it's free.

http://arstechnica.com/information-technology/2013/09/google...

If 10% are paid by RedHat, 90% aren't - therefore count as free to RedHat. Even then, the single largest group are non-aligned developers.

I agree Red Hat and pretty much every other contributor's result is a net gain. The point I was making is that the overall contribution isn't free. ie the folks working on it are paid by various companies. I believe all the major tech companies have a few devs that they basically loan out to the kernel development effort.

Yes, there's potential for a Tragedy of the Commons style problem to happen with kernel development, where each corporate developer suddenly decides to save money and not contribute, whilst letting everyone else pay. Eventually everyone would follow suit and nothing would get done.

Not exactly the same as the Tragedy of the Commons, but similar in the sense that it works best when everyone works together.

NB: It's also important to note that some kernel work is actually on drivers for specific products. eg. Intel's contributions are largely making sure that all their hardware will work in a Linux server/desktop, not necessarily the kind of general Linux ecosystem development that Canonical might do.

I don't see the tragedy of the commons concern here. Presumably, each company is contributing because they need/want something with the kernel that is not already being provided. Instead of keeping the development in house, which would involve maintaining a fork and almost certainty breaking copyright, they decide to contribute it back to the mainline kernel. At the point where no one wants to contribute, I would expect that the expected value for some companies of having a single employee work on the kernel is greater than the That this happens to benifit everyone else is just a side effect.

Most contributions are from volunteers, but I do not know about contributors. I assume volunteering contributors are even more, because they are sort of the fly by crew.

Not sure of that...I may be wrong but I believe there is a karma type process in place for instance regarding who can contribute patches...I can't imagine there is much fly-by-night contribution...except perhaps bug reporting.

> Why on earth is the article comparing RedHat to Amazon?

I don't understand that and not because:

> One is a Linux distribution vendor, the other is a global internet shopping mall.

I don't understand that because Red Hat already does what Amazon does:

http://www.redhat.com/solutions/open-hybrid-cloud/

I don't know if Peter Levine knows that.

Yes, but the article is comparing the total Amazon revenue in the $70bn+ range to the RedHat ~$1.3bn. A completely misleading comparison, considering that Amazon don't actually break out their AWS figures in their financial reports.

Those charts are highly misleading and borderline duplicitous.

He shows estimated AWS revenue on [1], where he appears to be using the $3.8 billion revenue estimate. I regard that estimate as defensible, and a reasonable indicator of the size of the market.

[1] http://tctechcrunch2011.files.wordpress.com/2014/02/peterlev...

[2] http://www.zdnet.com/amazons-aws-3-8-billion-revenue-in-2013...

You have to remember how AWS started - Amazon selling their excess datacentre capacity in their offpeak downtime.

It really isn't possible to distinguish how much of AWS is a money maker for Amazon, because you can't separate the costs of running it from the costs of the Amazon business as a whole. For all we know, they could be selling the excess capacity at a massive loss, which is still justifiable for them because at least it recoups some of their everyday running costs.

Yes, because "providing service and support for software running on someone else's hardware" is the same as "maintaining large datacenters, including hardware, power and cooling costs," and therefore let's just compare revenues without looking at costs.

If I was an investor, and I saw Redhat growing revenue at the rate AWS seems to be then I'd be buying as much Redhat stock as I could - even if their profits were depressed because of the investment they were making.

I'll quote the investment guidance from the link I posted above[1]:

Using our estimate of $3.8bn for 2013 AWS revenues, and applying a ~5x multiple based on the comps noted above, we arrive at a valuation of ~$19bn for the business on an EV/Sales basis (equating to ~$41/share of AMZN stock). Importantly, we believe this to be a conservative valuation multiple, as AWS revenues are growing much faster than any of the comps incorporated above. At an 8x valuation multiple, we estimate the AWS business could be worth $30bn as a stand-alone company, or ~$66/share.

Note that they value AWS as a $30bn company right now, almost triple Redhat's $11bn market cap.

From [2], we see that Amazon has has invested $4.6 billion in its business over the past year, including $1.4 billion to purchase the building that houses its Seattle headquarters.. Taking out that $1.4 billion, we get $3.2 billion, which almost matches estimates for AWS revenue.

So we see AWS is growing like mad, pretty much owns the whole cloud hosting category[3] and MIGHT NOT EVEN BE LOSING AMAZON MONEY, even while they are investing so much that no one else can compete with them on scale of their offerings already.

Take my money now.

[1] http://www.zdnet.com/amazons-aws-3-8-billion-revenue-in-2013...

[2] http://www.informationweek.com/cloud/infrastructure-as-a-ser...?

[3] http://readwrite.com/2013/08/21/gartner-aws-now-5-times-the-...

Red Hat already does what Amazon does

No, they don't. You can't pay RedHat money and run software on their infrastructure.

Instead, Redhat is building a set of open source platforms, which they will sell support to you for.

Their IAAS platform is based on OpenStack, and their PAAS is OpenShift[1].

The problem isn't the software, it's that they are relying on support revenue for these too. As Levine points out, that isn't as profitable source of revenue as running your own cloud service, or selling a SaaS app.

I don't know if Peter Levine knows that

Levine was the the CEO of XenSource (who wrote the Xen hypervisor). Pretty sure he knows the cloud space fairly well.

[1] https://www.openshift.com/

"You can't pay RedHat money and run software on their infrastructure."

You definitely can.

https://www.openshift.com/products/pricing

Canonical is not profitable.

The parts that do what Red Hat do are; the overall lack of profitability is because Canonical is losing money on other market segments. (In other words, they're generating money to be invested in developing new products, exactly what the article says open source companies can't do.)

My bad. Perhaps I should have said sustainable. Canonical is not going to disappear. They may have unprofitable ventures, but ultimately they can cut it back to the core business and carry on going. Got to spend money to make money!