It takes time, leadership, and a fair amount of luck to successfully build an open-source community. It also takes money. Lots of it, if IBM's $1 billion commitment to Linux is any indication.
Unfortunately, the return on such open-source community investments may be permanently scuppered by the European Commission's misguided defense of MySQL from Oracle's intended acquisition. If the EC is going to punish successful open-source endeavors like MySQL, will investors still clamor to finance the rise of open source?
In many ways, MySQL is the quintessential commercial open-source success story. On the financial side, MySQL managed to build a vibrant business, doing north of $90 million at the time of its acquisition by Sun Microsystems in February 2008.
Equally compelling, however, is the exceptional user and developer community that formed around the open-source database project, registering tens of millions of downloads and a massive developer community.
This community augmented MySQL's financial fortunes, of course, but it also protected MySQL database users from the whims of the company, as former MySQL CEO Marten Mickos wrote to European Competition Commissioner Neelie Kroes:
Even if Oracle for whatever reason would have malicious or ignorant intent regarding MySQL (not that I think so), the positive and massive influence MySQL has on the DBMS market cannot be controlled by a single entity - not even by the owner of the MySQL assets. The users of MySQL exert a more powerful influence in the market than the owner does.
Unfortunately, the EC seems intent on punishing MySQL--both community and company--for its success. Already the MySQL database project has started to fracture into competing forks, while business rivals like EnterpriseDB and IBM collect confused customers.
More worryingly, the EC's actions may end up diminishing potential returns to investors in other open-source projects, particularly those that take the added time and cost to build global communities.
Technology mergers and acquisitions activity is at a 20-month high. Open-source companies, however, may miss out on this resurgence, particularly those, like Acquia and EnterpriseDB, that build on successful open-source communities (Drupal and Postgres, respectively).
Indeed, based on the EC's actions, perhaps the worst thing these companies could do is foster successful open-source communities. Maybe they should just take the cash and run?
Consider: the EC didn't challenge Yahoo's acquisition of Zimbra, VMware's acquisition of SpringSource, Citrix's acquisition of XenSource, etc. What do they have in common? Rising revenue but, except in the case of SpringSource, much more limited communities than MySQL. (Even the Spring community pales in comparison to MySQL, impressive though it is.)
Granted, the major difference with Oracle/MySQL is that the two are ostensibly competitors, as CNET points out. In the letter referenced above, however, Mickos dismisses such competition. The reality is that MySQL and Oracle compete in two different database markets.
Regardless, as well as MySQL was doing, $90-plus million is spare change in the global database market. The EC, in other words, isn't trying to protect MySQL's business. It's trying to protect MySQL's community.
Such mollycoddling of an open-source community is destructive to all future investments in similar endeavors. Why should commercial entities bother fostering community--the very community that makes them less susceptible to hostile takeover and anticompetitive forces--if doing so simply ends up ruining financial returns?
The EC means well, but it is not doing the right thing for MySQL, its community, or other open-source commercial efforts. Quite the opposite. Just as the commercial open-source community has been pondering a move back to community-controlled open source, the EC threatens to hobble the shift.
The EC may well end up with less competition, not more, by blocking Oracle's proposed acquisition of Sun and its crown jewel, MySQL.
Novell has been positioning itself as the Avis of Linux, a distant but gaining Red Hat competitor that "tries harder." Like Oracle, Novell argues that it can give customers Red Hat value at a lower price.
What, me worry?
It's true that adoption of unpaid Linux like CentOS is booming, and that this no-cost alternative to more expensive solutions like Red Hat is a real threat to Red Hat. This is no doubt why Red Hat has made "free-to-paid" a core element of its ongoing strategy, as related in its recent earnings call.
But it's a much bigger threat to Novell and Oracle, both of whom are trying to position themselves as cheaper alternatives to Red Hat Enterprise Linux.
If a customer really wants Red Hat at a lower price, they're not going to move to an incompatible distribution that may or may not run their applications properly. They're going to jump to CentOS, which is basically a carbon copy of RHEL, minus the trademarks (and price tag).
Oracle, for its part, is clearly not in the Linux market. It's in the market to eradicate Red Hat, so as to claim top-to-bottom control of its software stack. But even as Oracle tries to squeeze Red Hat into oblivion, CentOS provides an excellent hedge against commercial competition from Oracle (and Novell), making its pitch ring hollow.
CentOS: Red Hat's biggest annoyance and greatest friend?
It's not dissimilar to the role that piracy plays for protecting Microsoft's Windows dominance against Linux, especially in emerging markets. Quite possibly the worst thing that Microsoft could do, as IDC has also suggested, is to succeed in its anti-piracy efforts.
Were Microsoft to raise its pricing above $0.00 in such markets, suddenly Linux would look like a much better alternative.
Back to Novell and Oracle. It's not enough to try harder. Red Hat has created a dominant global brand that CIOs trust. It's not worth a few dollars here and there to disrupt that to shift to SUSE or Oracle Enterprise Linux.
Not when those CIOs can shave 100 percent of their RHEL subscription costs by moving to CentOS.
I know some CIOs who have, but they tend to be enterprises with lots of developers that are comfortable supporting themselves. Fortunately for Red Hat, few CIOs care to take that risk. Unfortunately for Novell and Oracle, those who do want to save all of their Linux subscription fees, not just some of them.
Even as the recession continues to cool CIO appetites for software purchases, Linux is bucking the trend, according to a new IDC report.
IDC is projecting Linux revenue to expand at a compound annual growth rate of 16.9 percent from 2008 to 2013, topping $1.2 billion in 2013.
As IDC notes, this growth will comprise just 4 percent of total software market revenue by 2013, up from 2.2 percent in 2008. However, for the second time, IDC has also examined nonpaid deployments of Linux, revealing some troubling data.
I've always assumed Red Hat's primary Linux competitor is Novell. And based on IDC's numbers, it does appear that Novell is increasingly a real threat to Red Hat.
But it is the nonpaid usage of Red Hat's software that may well pose a bigger risk.
Novell has 27.9 percent market share of paid deployments and 20.1 percent of the total paid and nonpaid market. This doesn't look so great at first glance; after all, more people use Red Hat (including Fedora) for free than pay for Suse Linux Enterprise Server.
However, in growth, Suse stands out. On paid shipments, Red Hat's 2007 to 2008 growth was 1.9 percent, while Novell's Suse was nearly double that at 3.5 percent.
On revenue, Novell comes in at 29.8 percent market share. That represents 50.3 percent growth in market share, versus Red Hat's 14.8 percent growth. Granted, Red Hat has a much larger base of revenue from which it's growing ($319.5 million compared with Novell's $112.6 million in 2007), but Novell's Linux revenue growth has outpaced Red Hat's since 2007.
I don't particularly like Novell's partnership with Microsoft to promote Linux, but it does appear to be paying off for Novell.
If Red Hat could elect to eliminate one competitor tomorrow, though, I'm wiling to bet that it would not choose Novell's Suse. It would choose unpaid Red Hat Enterprise Linux (RHEL), which accounts for a big chunk of the overall Linux market.
This may seem trivial, given that Red Hat earned a 62.2 percent share in the overall market for new license paid shipments/subscriptions, measured by deployments, or 64.7 percent, measured by revenue.
Sounds great, right?
Maybe. Intriguingly, Red Hat also claims 28.6 percent of the nonpaid market...for RHEL, its Linux distribution that should only be available to paid subscribers, but which many companies dishonestly use without paying (e.g., they may violate their contract by running more RHEL servers than they actually pay for).
Add Red Hat's paid and nonpaid deployments together, and Red Hat accounts for 47.6 percent of the global Linux market, whether users are legitimate customers or pirates.
It gets better (or worse, depending on your view). If one adds in the RHEL clone CentOS and Red Hat's own community distribution Fedora Core, Red Hat and its offspring dominate the global Linux deployments market with 57.1 percent market share.
This might not be so bad, if the trend were toward more paid Linux adoption, but it's not. While paid Linux server deployments will grow at an impressive rate, nonpaid deployments will grow even faster, nearly reaching parity with paid deployments in 2013.
Why this growth in nonpaid Linux?
Undoubtedly some of it stems from enterprises wanting to get something for nothing. Rather than pay for value, they attempt to cheat the system, leaving less money to help develop Linux.
But it may also be that the longer the world uses Linux, the less it feels the need to pay for it. Noted technology CTO Jon Williams once posed a dilemma to me at the Open Source Business Conference. He indicated that the longer his team works with an open-source project, the less need it has for support and maintenance from a vendor.
In other words, the minute the customer becomes profitable to the vendor is the same minute the customer no longer needs that vendor.
We could be seeing this in Linux. Still, the fact that they seem to be stealing RHEL rather than adopting Ubuntu or another "community-led" Linux distribution suggests that we're seeing enterprise IT attempt to cheat vendors rather than do without them.
All of which may mean that the world increasingly recognizes that Linux is a superior server operating system...and doesn't want to pay for it.
How comforting...and alarming. It's not as if Linux development costs nothing. Red Hat pays over $100 million each year to develop Linux, and it's not the only company making such hefty investments.
UPDATE @ 11:03 PT: For the reading-impaired: When I talk about "stealing RHEL" I'm in no way referring to CentOS or Fedora, as my post clearly states. I'm talking about using RHEL without paying for it. Not CentOS. Not Fedora. RHEL. Red Hat has a fair number of companies that actively underpay on RHEL: that is, companies use more RHEL than they are legally allowed to use as per their contract with Red Hat.
So, please read the post, and don't get worked up by the word "steal" and "CentOS" in the same post. I'm not referring to CentOS or other legitimate uses of Linux. I'm talking about theft of RHEL (which is what IDC is talking about, too. Maybe you should buy the report and read it before commenting so that we can have an informed discussion.
Follow me on Twitter @mjasay.
What's the takeaway when one free operating system displaces another when running your application?
That's the question I'd be asking myself at Groundwork in the wake of news from the Works with U blog that 29 percent of its new users are using Ubuntu to power its IT management application instead of CentOS, the Red Hat Enterprise Linux clone. When your users are falling all over themselves to avoid paying for an operating system, can the application be far behind?
Within my own company, we weight leads heavier if they are running our application with a paid version of an operating system, application server, or database. As the reasoning goes, if a prospect is willing to pay for one of these, it will be more willing to pay for our software, as well.
I'm guessing that the operating-system data cited by Works with U relates to raw deployments, and not to Groundwork's actual customers, which would have a much higher population of paid OSes running the application. This would comport with the data that we see at Alfresco: strong adoption of Ubuntu in evaluation but production usage sticks with Red Hat Enterprise Linux, for the most part, or Windows.
Ironically, an open-source vendor's best determinant of financial success may well be how much its prospects already pay for proprietary software. Such prospects may look to open source to shave costs, but not to evade them completely. I'll feel much better about this Groundwork data when the Ubuntu adoption is paid adoption, not simply free-riding.
Follow me on Twitter at mjasay.
Infoworld does an annual review of the best enterprise open-source applications, called the BOSSies, and just announced the 2008 winners. An Infoworld editorial team makes the selections, so this isn't a matter of open-source projects rallying the troops to vote for their projects (which sometimes has odd results, though often gets things right)
- Alfresco, Content Management
- Compiere, Enterprise Resource Planning
- dotProject, Project Management
- Hyperic HQ, Application Monitoring
- Intalio BPMS, Business Process Management
- JasperReports, Reporting
- Liferay Portal, Enterprise Portal
- Magento eCommerce, E-Commerce
- Pentaho Open BI Suite, Business Intelligence
- SugarCRM, Customer Relationship Management
Other categories include the best open-source productivity applications, best open-source middleware, and other categories. You can see the full details here or a snapshot view of the winners over on OStatic.
One nit? I don't like that CentOS was listed as the top open-source operating system. True, it wasn't listed in the enterprise category, but CentOS (which is just Red Hat Enterprise Linux without Red Hat's trademarks) is, in my mind, the worst sort of open source: It sucks money out of the system while giving nothing back. CentOS contributes no innovation to the Linux kernel and instead makes it harder for companies like Red Hat and Novell to invest in research and development.
- prev
- 1
- next
