The day open source became big business is the day that open-source development exploded. Yes, open source predates the moneyed interests hankering to use it to competitive advantage, but it really wasn't until IBM dropped $1 billion on Linux that companies began paying employees to write free software that the movement saw broad adoption.
That's when open source became more than an efficient way to develop software, and also became a great way to build a business.
However, adding open source to one's business is not magical pixie dust that guarantees its viability. As IBM's Bob Sutor explains:
The basic principles around revenue, profit, loss, taxes, payroll, overhead, accounting, sales, incorporation, health care, and human resources all apply. You can be a starving open source software entrepreneur as easily as a starving proprietary software entrepreneur. No one will excuse basic business failures and screw-ups just because you use open source. Make sure that you will produce a product that people want and in some way will pay for, no matter how indirectly.
Sutor's counsel applies to any company or individual that wants to build a business around open-source software, but arguably some of the industry's best projects are not the product of any one company, but rather of several. Linux, Mozilla, Apache Software Foundation, Eclipse, and other collaborative communities represent an interesting way to use open source to competitive advantage.
In many ways, open source has become a critical component of the software industry because the market has largely moved from vertical businesses (i.e., companies controlling all aspects of production, distribution, etc.) to horizontal markets (i.e., companies focusing on their core competencies and depending on others for complementary functions).
Linux: Peace, love, & squeezing Microsoft
But this business control system has a inherent risk. Should an organization monopolize a specific segment of a value chain system they can extract a higher percentage of its total proceeds. If the product, or service, in question is price elastic than those additional proceeds will come from other participants in the value chain system.
Case in point? Windows. By owning the operating system, Microsoft threw a wrench into the collective cogs of horizontally oriented software firms like Intel, IBM, and others.
The industry's response--Linux--is a classic example of the open-source approach to mitigating individual choke holds within an industry, as Prentice goes on to write:
What then does a CEO do when facing a squeeze on their profits because a direct, or downstream, supplier is dominating a segment of the value chain system? Besides negotiating a better deal - if they can - they've been left with little choice but to get directly into that segment of the value chain system themselves. But by doing so their organization is distracted from focusing on its own core competency.
The risk of such an undertaking can be mitigated if there is a collective response by similarly affected members of the value chain system. After all, it is usually a shared problem. But collective responses have always had an inherent, and often fatal, flaw. Who owns the resulting assets? Either organizations enter into complex joint venture agreements to sort this out or run the risk of shifting the distortion in the value chain system to another organization.
Again, Linux offers the perfect example. IBM, Intel, Red Hat, and others aren't investing in Linux because they're all chums at the country club together, but rather because they're looking for ways to reduce Microsoft's hold on their own businesses through its control of personal computer and server operating systems.
As an added benefit, it's a great way for companies to collaborate without running afoul of antitrust laws. It's collusion without the collusion.
Intriguingly, even Microsoft is getting into this game. Microsoft's partnership with open-source ad serving company OpenX indicates that Microsoft, too, is figuring out how to use open-source complements to loosen strangleholds competitors like Google may be hoping to throw in its way.
This is why open source is growing so much faster than the rest of the industry, as IDC finds. It's not because we love each other more. Quite the opposite. It's because proprietary vendors have figured out that open-sourcing key complements to their core businesses can be strategically decisive in hurting competitors while helping themselves.
For years, open-source advocates have been praying for someone to free us from Microsoft's proprietary grasp. We've prayed in vain as Linux, OpenOffice, and other open-source software programs have failed to dent Microsoft's dominance.
Until now.
Google, not Red Hat or Sun, appears to be the long-awaited redeemer of both personal computers and servers, and has even staked a credible claim in the mobile world, as well. Google achieves this, in part, by writing copious lines of open-source code, but pays for this "generosity" with insanely profitable proprietary services, services that have long appealed to consumers but increasingly appeal to enterprises, too.
Google, in other words, is arguably not the open-source savior we were expecting, but it's probably the one we deserve.
(Credit:
Matt Asay)
Despite more than a decade of trying to make "pure" open-source software businesses work, it's telling that only one company--Red Hat--has managed to pull together more than $100 million per year in revenues for its troubles. For its part, Red Hat is quick to downplay the relevance of its revenue model for just about any other business.
Hardly a rallying cry to the still-growing open-source ecosystem.
Yes, MySQL got to $94 million before Sun gobbled it up, and yes, other start-ups (my own, included) are getting closer to the mark, but none, including MySQL, is wholly dependent on selling open-source software subscriptions to achieve this goal.
We also include proprietary add-on value. Like Google.
So we're left with Google, which is, perhaps, the world's largest open-source company, contributing more open-source software and resources than any other, in my estimation. (Sun likely wins on sheer volume of code, but being an "open-source company" involves more than simply code.)
How does Google do it? Well, for one thing, it learned long ago that monetizing open-source software directly is tough. So it simply uses open source to shepherd prospective customers to its other services, like Search or Google Apps.
Indeed, it is the success of these proprietary products that enables it to be such a generous open-source benefactor, much like IBM, Intel, or, for that matter, Sun (which sells a lot of proprietary hardware). Take away these companies proprietary product lines, and overnight we'd see dramatic decreases in their investments in Linux, Apache Software projects, etc.
And we'd all be the poorer for that.
In an ideal world, open-source software companies would thrive by simply giving away lots of code, and having enterprises and government organizations serve their long-term interests by paying for support.
We don't live in that world. Some organizations do buy support for open-source software, of course, though many others do not, and some only pay long enough to become self-sufficient whereupon they dump their support contracts, as former CTO of NBC iVillage Jon Williams once declared.
Until we cross the border into Utopia, we're going to continue to see the biggest investments in open-source innovation come from Google and its peers: companies with wallets fat with proprietary profits.
As I said, this may not be the open-source world for which we've hoped, but it's the one we deserve, because it's reflective of what we value and, hence, what we pay for.
See also Mark Hinkle's response to this post.
Gartner has had a rocky relationship with open source in the past, but recent research suggests that its views on open source have evolved. It's therefore time for the open-source world's views on Gartner to evolve, too.
Gartner hasn't historically been much of a friend to open source. While Forrester, Redmonk, the 451 Group, IDC, and other analyst firms long ago began recording the rise of open source within enterprise computing, Gartner seemed to side with the proprietary vendors in steadfastly arguing that open source's impact was negligible.
This resulted in some suggesting that Gartner's research was simply a reflection of which companies paid it the most money (and recently netted the analyst firm a lawsuit).
I made similar accusations myself.
Gartner responded to such attacks, defending the integrity of its research. Yet its blind spot to open source seemed to persist.
Not anymore. Whatever the reason for the erstwhile overlooking of open source, Gartner analysts' current views on open source have changed, in some cases dramatically.
It used to be that open-source companies and projects never made it into Gartner's Magic Quadrant (MQ), which have tremendous power for, if somewhat limited utility to, enterprise buyers.
Now, you'll find that Gartner lists Drupal ("Drupal is in the Visionaries quadrant because of its use of the open source model to drive adoption and popularity, while providing enterprise services via organizations such as Acquia"), Liferay, and MindTouch in its newest "Social Software in the Workplace" MQ, while Alfresco, MySQL, JasperSoft, Pentaho, and others are listed in a variety of other MQs. (Disclosure: I work for Alfresco and am an adviser to MindTouch and JasperSoft.)
Gartner also recognizes the broad adoption of open source in the enterprise and how open source is affecting even proprietary software vendors.
This isn't to suggest that Gartner finally "gets it" because it's writing favorably about open source. In fact, some of Gartner's best, most interesting analysis is available for free on the blogs section of its Web site, not all of which is positive about open source.
It is, however, balanced and often quite insightful, particularly the work of Gartner analyst Brian Prentice.
Given all this, it's time for open sorcerers to stop using Gartner as a straw man for poor analysis on open source. This isn't helpful and, increasingly, it's not remotely accurate.
(Credit:
Amazon.com)
Open-source software may have a lot in common with the global soccer (football) business: while it generates a tremendous amount of value for users, very little of that value can be converted into cash. At least, not directly.
That's the thought that struck me while reading the exceptional "Soccernomics: Why England Loses, Why Germany and Brazil Win, and Why the U.S., Japan, Australia, Turkey--and Even Iraq--Are Destined to Become the Kings of the World's Most Popular Sport." Among other things, the book tackles the economics of soccer, and yields some counterintuitive insights:
...[I]f Deloitte ranked [soccer] clubs by their profits, the results would be embarrassing. Not only do most clubs make losses and fail to pay any dividends to their shareholders, but many of the "bigger" clubs [like Real Madrid and Manchester United] would rank near the bottom of the list....[Note: my Arsenal is one of the few profitable football clubs on the planet.]
Whichever way you measure it, no soccer club is big business....This feels like a contradiction. We all know that soccer is huge. Some of the most famous people on earth are soccer players, and the most watched television program in history is generally the most recent World Cup final.
Nonetheless, soccer clubs are puny businesses. This is partly a problem of what economists call appropriability: soccer clubs can't make money out of (can't appropriate) more than a tiny share of our love of soccer....[T]he world earns more from soccer than the soccer industry itself does.
It's the world's biggest game...with some of the world's worst financial returns. We buy the replica shirts. (Um, I buy many.) We pay to attend games. (Er, I pay to watch many.) We try to give the game our money. But it generates very little top-line revenue, and almost never any bottom-line profits for soccer clubs.
Like football, there's no question that open source is exceedingly popular these days. Virtually every company--indeed, every person--on the planet uses it in some way, whether it's the free Firefox browser with which someone reads this blog post or the Linux operating system serving up a wireless carrier's phone system, open source is everywhere and highly useful.
It's just not big business.
Yes, Red Hat is nearing $1 billion in annual sales, but it's the exception. And that's OK.
Open source, like soccer, doesn't have to directly generate mountains of cash to be immensely valuable to the companies in its ecosystem. For every Real Madrid squeezing annual revenues of $475 million out of soccer there are scores of broadcasters, sports apparel companies, etc. making billions on the back of the sport.
In similar manner, Google, IBM, and others like them make billions with the help of open-source software, but they make very little directly from open-source software.
Like the soccer economy, the open-source software economy is best measured by the total value it creates, which will have very little to do with the direct sales the Red Hats of the world report. Open source saves enterprises billions of dollars in license fees, and arguably has the potential to collectively add trillions of dollars in productivity gains.
That's big value, even if it's not big money. Not in the pockets of software entrepreneurs like myself, anyway.
Technology companies are generally quick to publicly announce and highlight their customer wins. But in what strikes me as a first, Novell has publicized a customer loss, announcing to the world that the City of Los Angeles dropped it for Google Apps, including Gmail, Google Calendar, and Google Docs.
Google recently announced the City of Los Angeles as its latest high-profile customer win for Google Apps, one that reflects growing momentum for the cloud-computing giant's enterprise business. Until Novell's announcement, I had no idea that Los Angeles had dumped Novell GroupWise in favor of Google Apps, and suspect few others did, either.
Novell, however, attacks Los Angeles' decision, arguing "The City of Los Angeles should have opted for this proven product [GroupWise] to ensure the security of its data and to save taxpayer money. They have taken a risk with no reward."
Translation? "We think Los Angeles is run by a bunch of fools who aren't smart enough to know what's good for them."
This is no way to treat customers.
It's not made any better with this throwaway line: "However, as a valued customer, Novell will continue to offer our world-class support to the City of Los Angeles during the transition."
Well, that's comforting, but given that it comes at the tail end of Novell publicly excoriating its "valued customer," it's doubtful that LA will linger long with such a "valued vendor."
This isn't the Novell that I know. I used to work for Novell, and have never seen the company publicly criticize a customer, not even for defection, of which Novell has seen plenty over the last decade.
It's unclear who Novell is hoping to persuade with the announcement, or what benefit it hopes to derive from it. Is it trying to stem a tide of customers dropping GroupWise for Google Mail? If so, why has it not done the same for all the companies (and there have been plenty) leaving GroupWise for Microsoft Exchange or IBM Notes/Domino?
In fact, the only companies that benefit from this kind of customer abuse are IBM and Microsoft, because Novell slams Google Mail's alleged security and cost deficiencies, without them having to sully their hands with the negative marketing. It certainly won't endear Novell to the City of Los Angeles or to its other customers.
Mozilla isn't just about browsing anymore.
While the foundation made its name with the increasingly popular open-source Firefox browser, it is quickly moving beyond its roots, particularly in the area of e-mail. With the launch of Raindrop, its Google Wave-like unified messaging and collaboration system, as well as corporate uptake of Thunderbird, Mozilla may soon extend its reach well beyond its browser base.
Corporate America hasn't done much with Mozilla's Thunderbird, a competitor to Microsoft Outlook. Europe, however, has given it a warm reception. For example, the French tax authority recently selected Thunderbird to power 130,000 of its personal computers, replacing IBM Lotus Notes and Microsoft Outlook.
It's a massive deal for Mozilla, though in the grand scheme of things, it's still tiny. Even so, it's an indication that Mozilla's e-mail story is credible, and could lead to greater adoption of Thunderbird and, eventually, Raindrop.
Much of Firefox's early traction was in Europe. The same could hold true for Thunderbird and Raindrop.
The question for me, however, is how it gets funded. Google has essentially funded Mozilla's browser development for years. It's unclear who the "Google" is for Mozilla's messaging ambitions, or whether the foundation intends to sell subscriptions to use the software through its for-profit corporation.
Regardless, Mozilla's presence in the messaging market is welcome. While we already have an exceptional open-source competitor to Microsoft Exchange and IBM Lotus Domino in Zimbra, given the importance of messaging and collaboration to enterprise computing, it's useful to have an open-source foundation involved, too. About the only organizations that won't like this increased competition are the proprietary incumbents.
You can look at Google's growing market share in Android, its dominance in search, and elsewhere as signs that it's winning in its markets. But for me, the best indicator that Google is winning is the increasingly vitriolic attacks piled on it.
You want a piece of me?
You can always spot a winner by the bull's-eye painted on it. No one bothers to diss a loser.
Or sue them. Red Bend Software has launched what appears to be a specious patent claim against Google, alleging that Google's Chrome browser violates its patent (6,546,552) by including the Courgette algorithm, which enables Google to push compressed software updates to the browser.
As Microsoft learned years ago, success breeds patent lawsuits. Microsoft rarely sues over intellectual property infringement, but has endured hundreds of patent lawsuits, nearly all of them ultimately found worthless.
But it's not just patent trolls that are on the scent. Symbian, the one-time leader in mobile phone operating systems, has gone on the offensive, claiming Google is "evil" and fear-mongering about what Google will do with consumer data gathered with its Android software.
Is this an indication that Symbian can't compete in the market and must instead resort to FUD?
Google may ultimately be able to get out of the Red Bend lawsuit cheaply, and it's unlikely that Symbian's noise will unduly distract it. It may not end up dominating mobile, for a variety of reasons, but it's going to be a significant competitor, just as it is in search and increasingly in enterprise computing.
After all, Google is innovating in Android, as CNET reports, and generally pushing the envelope on what's possible in computing: mobile, "desktop," and cloud/server. Importantly, open source is a central strategy in each of these areas, which may be one of the things that most riles the incumbent competitors in its markets.
For Google, the increasing vehemence of the attacks on it should signal that it's doing something right. In fact, many "somethings" right.
Ars Technica's Ryan Paul wants to know, "Can a [truly open smartphone] be done?" But the real question is, "Should we care?"
Hello? Can I get some freedom around here?
Hence, Bradley Kuhn of the Software Freedom Law Center expresses anxiety about the future of freedom in mobile...
We are in a very precarious time with regard to the freedom of mobile devices. We currently have no truly Free Software operating system that does the job.
...when he really should be concerned with choice in mobile. Right now, we're spoiled for choice in mobile, what with Apple's iPhone, Google Android, Symbian, LiMo, Moblin, etc., which suggests that users are free to move between devices.
In this case, it's not the license that makes users free. It's the market.
Open-source software plays an important role in ensuring user choice, but it's not the sum total of the freedom/choice equation. It's just one factor. As Tim O'Reilly reminds us, it's not even necessarily the most important factor, either.
Kuhn and other free-software advocates worry that the nuts and bolts making up the software on mobile phones be free, but this is surprising given the increasing irrelevance of single-node freedom when it's tied into a network. This is what I've described as "the Hotel California of tech," and it suggests we should be far more concerned with freedom between nodes than freedom of the nodes themselves.
In other words, the real concern should be over open data, not open phones. No matter how open my phone's software may be, it's meaningless if I can't move my data between devices or wireless providers.
Even here, there's cause for hope. For example, Funambol's open-source mobile cloud synchronization and push e-mail software is in use by 10 of the leading mobile service providers, as identified in a new report, which arguably should be more relevant to the Freedom fighters than whether Bluetooth is open source.
Glyn Moody, a journalist with strong free-software leanings, understands this. That's why he makes the case for an open cloud, and not simply "open node in the cloud":
Ideally, what we need is a completely open source cloud computing infrastructure on which applications providing people with things like (doubly) free email and word processing services could be offered....The trick here is not to fight the battle on the opponents' terms, but to come up with something completely different.
For example, how about creating an open source, *distributed* cloud? By downloading and running some free code on your computer, you could contribute processing power and disc space that collectively creates a global, distributed cloud computing system. You would benefit by being able to use services that run on it, and at the same time you would help to sustain the entire open source cloud ecosystem in a scalable fashion.
One can quibble with the feasibility of this approach, but at least Moody is thinking at the right scale. Those who are still stuck in the Open Source 1.0 of isolated, client-side software are not.
I suppose someone has to fixate on upper-case Freedom above all other priorities. Like usability. Or ubiquity. Or...well, anything.
But most of us don't think this way, because the world is a lot more complicated than Freedom on one hand, and Slavery on the other. Also, the focus of freedom has evolved in our networked world, though some free-software advocates seem mired in Freedom 1.0.
It's time to upgrade. Freedom is more than a license. It derives from a competitive market, one that is assisted by open source but not exclusively or even primarily defined by it.
One App Store to rule them all?
(Credit: Apple)Apple has an app store, of course. So does Microsoft. Google has two, one for Android and now one for Wave. In fact, it's hard to find anyone who doesn't have an app store these days.
We're swimming in app stores. Or drowning.
I'm serious. At the Symbian conference in London on Tuesday, I attended a panel that was overrun with app stores. Nokia, Symbian, GetJar, Sony Ericsson, Handmark, and Handango were all promoting their respective app stores, each talking about how great theirs is.
They're probably right. They probably are all great. But how am I, as a lay consumer, going to figure out which one to use?
More particularly, how will developers decide which platforms to target?
After all, everyone wants to be a platform these days. Does that mean that no one is?
Developers may be spoiled for choice, but "choice" in this case may not be what they want. Developers need to feed their families and will follow the money. Money is more easily made when choice is manageable (which is a euphemism for "limited").
This means we'll see plenty of application developers remain with Apple (though it's debatable whether the iPhone is the land of milk and honey for anyone but Apple), but we'll also continue to see a stampede to Google Android.
At present, every other mobile platform is playing for third place, but this could change: Symbian, as a foundation, is in a good position to launch an effective challenge to both Apple and Google if it can get its marketing and execution right.
Outside of mobile, it's unclear what role app stores will play. It's nice that Google Wave is getting an app store, but it's just one more "forge" among many. Every vendor (my employer, included) seems to feel an irresistible urge to create a forge/app store where third-party developers can "add value" to their "platforms."
Do we really need these? Or do we need more general repositories like Google Code and SourceForge?
I wish I had a definitive answer. I'm just not sure that these competing app stores do anything more than appeal to vendor vanity, and they could end up causing customer confusion.
As a consumer, I don't want to have to think about sorting among competing app stores. I just want applications.
Presumably, if I use a Sony Ericsson phone, I'll automatically find myself within its app store (unless my wireless provider doesn't slot me into its app store first, that is). But if that's the case, what's the point of making a big deal over a glorified catalog of applications that work with my given device/software/etc.?
It strikes me that app stores, like the cloud, are simply a way to dress up old ideas. If they help to organize potential buyers and sellers of software, great. But I still think I'd prefer meta-repositories of applications, similar to SourceForge, than individual application repositories for every single device or piece of software that I happen to buy.
How about you?
The U.S. military is no laggard when it comes to open-source software adoption, but apparently thinks it can do better. The U.S. Department of Defense on Tuesday issued new guidelines designed to remove roadblocks to open-source adoption, arguing that open source can help the Defense Department "anticipate new threats and respond to continuously changing requirements."
And to think open-source software like Linux used to be considered a threat to secure Defense Department systems.
While Department of Defense CIO David Wennergren's revised guidance (PDF) is not intended to create new policy, it does provide clarity that suggests open source is very welcome at the Defense Department.
Apparently, the Defense Department's guidance on open source, issued in 2003, wasn't resulting in as much uptake as the CIO desired.
Hence, the new guidance specifies that open-source software meets internal purchasing requirements for "commercial computer software," and as such gets statutory preference in purchasing decisions, just like software from Oracle, Microsoft, or others.
But the guidance goes beyond neutrality to suggest reasons that open-source software might be better than such alternatives, including:
- The continuous and broad peer-review enabled by publicly available source code supports software reliability and security efforts through the identification and elimination of defects that might otherwise go unrecognized by a more limited core development team.
- The unrestricted ability to modify software source code enables the Department to respond more rapidly to changing situations, missions, and future threats.
- Reliance on a particular software developer or vendor due to proprietary restrictions may be reduced by the use of OSS, which can be operated and maintained by multiple vendors, thus reducing barriers to entry and exit....
- Since OSS typically does not have a per-seat licensing cost, it can provide a cost advantage in situations where many copies of the software may be required, and can mitigate risk of cost growth due to licensing in situations where the total number of users may not be known in advance...
- OSS is particularly suitable for rapid prototyping and experimentation, where the ability to "test drive" the software with minimal costs and administrative delays can be important.
Ultimately, the Defense Department CIO leaves it to individuals to determine which software best meets Defense Department requirements in a given scenario, but the memo hardly reads like neutral guidance. This is consistent with a wise policy of preferences, not mandates, for open source.
It's also an indication of much more Defense Department open-source adoption to come.
(As an aside, special thanks to John Scott for alerting me to this news, and for his work with the Defense Department to help this happen.)
