Open Source and Commercialization

I recently spent some time speaking with a popular Yankee Group analyst who covers the enterprise sector in the US, focusing in on open source and where the movement may go in the next few years.

Just to be clear, I differentiate, as most industry watchers do, between Linux and open source. While Linux is open source, the primary Linux distributors have caught on to how they need to position themselves for success and are starting to run their businesses just as any proprietary software company does.

Red Hat and SUSE make prime examples, realizing the path to long term success and revenue streams resided in proving themselves enterprise worthy to larger businesses and institutions, have shifted business models or been acquired by organizations with roots in the enterprise.

Her views, while not always popular in the open source community. are right on point if open source seeks widespread adoption and a permanent seat at the table for longer term financial success.

There are a few obstacles open source proponents need to accept and move forward on:

1) It will be more costly for a company to migrate away from Windows to Linux, even in light of slightly reduced ongoing maintenance and improved security and uptime.

While I have not always agreed that the costs are higher, having migrated corporate systems to Linux in the past, their research showed it to be true in many cases — especially when migrating beyond standard web hosting and email systems. The costs are higher when factoring in re-certifying drivers, application integrity and training.

2) To truly become entrenched as a viable financially-rewarding option (meaning open source companies make money and create jobs), a shift toward commercial software models is necessary. This does not mean forgoing open source, however, what it does mean is developing a structure for development, distribution, patching and support that passes muster with corporate IT managers who could be investing substantial amounts of money in open source.

What it boils down to is that while open source has definitely revolutionized software, and it is found internationally in companies large and small, businesses still pick software because it provides a solution not just because it is open source.

The fact that it is cheaper or free simply means the user will save money, but this does not win the favor of those buyers who could be injecting millions into open source projects rather than proprietary software makers.

I would use Firebird as a model. In an interview with Helen Borrie, forthcoming in my July column on SitePoint, she noted that since many Fortune 500 companies are using an open source database like Firebird speaks volumes to the maturing of their project and open source at large.

The reason as I see it, is due to the treatment of Firebird like an enterprise scale proprietary software project. They have a well managed developer community and active support lists, commercial offerings for support through partnerships with several companies, and commercial development projects for corporate clients.

If more open source projects looked at Borrie’s team model and discipline in development and support, we just might see more penetration that attracts longer and more profitable contracts and work for those like us in the SitePoint community.

Win an Annual Membership to Learnable,

SitePoint's Learning Platform

  • http://www.phppatterns.com HarryF

    It will be more costly for a company to migrate away from Windows to Linux, even in light of slightly reduced ongoing maintenance and improved security and uptime.

    You mean relative to staying with Windows? Does this include recurring costs of Windows licensing / upgrades?

    The costs are higher when factoring in re-certifying drivers, application integrity and training.

    On the drivers front, that assumes (if we’re saying Linux cf. Windows) that systems need upgrades as frequently. There’s generally less need to keep upgrading Linux, when used as a server.

    Re application integrity, think thats very hard to research accurately – kind of a wooly comment that needs qualification.

    On the training side, it’s an interesting area where it’s kind of like comparing Apples with Pears.

    Windows generally hides administrators from much of what’s really happening, so it’s probably easier to train someone to the point where they’re feeling confident but given serious problems, who do you turn to?

    *Nix effectively exposes administrators to everything so more time is required to reach the point where sysadmins are confident. Once they reach that point though, they’re typically capable of handling anything. The result is stable systems. I’d also argue that a single *Nix sysadmin is capable of maintaining a greater number of systems (scripts / automation etc.) although no figures to back that.

    Firebird is an interesting example. The flip side of Firebirds way of doing things seems to be the Open Source “community” is largely unaware of it (compared to, say, MySQL).

  • http://www.practicalapplications.net bwarrene

    Yes – on costs – Linux was actually found to be more expensive in numerous cases compared to staying with Windows. This is unfortunate as I am a proponent of finding migration paths from Windows to Linux for stability and administration automation. However, the research did show the total cost of ownership eventually balances out, it simply is much more expensive at the outset than staying on a Windows upgrade path.

    This survey (partially on site with staff and others via questionnaire) – 1000 companies with 5000 or more employees – found that they did have to certify drivers at the initial migration, certify all new disk images, provide training or certification to adhere to corporate policy, buy indemnification insurance, perform migrations, test, establish support contracts and finally, pay about a 15 percent premium when bringing in certified L:inux staff.

    The benefit if the company decided to take the financial hit: over an extended period they experienced the benefits of Linux – uptime, experienced admins and flexibility of the platform.

    Application integrity was ambiguous in the study – however – managers cited it constantly when trying to retire commercial Unix and move apps to Linux, needing certification that an entire applications runs exactly as before.

    Perhaps it is time for the open source community to begin establishing central organizational points that act as clearinghouses – like Open Source Development labs does for Linux – to certify open source applications on a major scale.

  • andrecruz

    I beg to differ on Harry’s view about Firebird. Firebird is not as popular as MySQL because 1) it’s a newer project (project, not software) and 2) MySQL support comes built into PHP; no need for additional software. Firebird requires either recompilation or loading this DLL into the extension space.

  • jmcginty

    It was nice to read about your chat with L… DiD… (why are we keeping her name secret?).

    Second, I don’t understand your distinction between Linux and Open Source. Maybe I’m slow or something, but what it seems to boil down to is:

    “Open Source = unprofessional
    Proprietary = professional (unstated)
    Linux = open source, but starting to become professional despite itself by acting like proprietary.”

    Well I’ll grant you there are a lot of unprofesssional Free Software projects out there; but the same is true of proprietary. Bad proprietary programs are slightly less likely to see the light of day, but there’s still a bevy of them out there.

    Now, on the assertion that Linux companies are succeeding by acting like proprietary companies: there’s truth and non-truth to it. On the one hand, Red Hat and SuSE have no doubt learned a lot about management, marketing, and good business practices from established companies. On the other hand, an effective open source player does not act the same as an effective proprietary player: there are all kinds of issues with dealing with the developer community that are not an issue in the proprietary world: they bring plusses and minuses, but have to be dealt with rather than ignored.

    And I will note that Red Hat, the most successful Linux distributor, is a pure-play Open Source vendor: they do not ship proprietary code. In fact, they devote a lot of developer time to a community distribution that they make no direct money on (but do get free testing from). Likewise, one of the first things Novell did after its so-far successful acquisition of SuSE was to GPL SuSE’s proprietary installer. This suggests that while good management is indispensible in anythin, Open Source ventures should not be running off and trying to ape proprietary vendors blindly.

    Finally, there’s a big difference between the way mass-market shrinkwrapped proprietary software and the way big-iron stuff is. With big-iron stuff you often have consultants in the field, lots of direct customer feedback, maybe even code sharing under NDA with the client: in short, it works a lot like an Open Source project. And that’s where Open Source has shined: *nix boxes, web servers, network infrastructure, compilers, developer tools, and increasingly RDMSes. With mass shrinkwrap you have to do much more seeking out of customer needs on your own and also be prepared to tell customers to shove it and wait for the next release. On stuff like this (desktop guis and apps) Open Source has been less successful.

    At least one high-profile OSS desktop project (Mozilla) was a legendary quagmire for a long time and is only beginning to claw its way back. Many of the mistakes came from not being open to community input (“dammit, we don’t need a whole platform, just a good browser”) as any good project of any kind should be. Thing is, no one has a clear idea of how to be usefully open to community input on a mass-market OSS project yet: the twin dangers of adding every requested feature or my-way-or-the-highway-ism have been so far hard to avoid.

    Personally, I think the question of the Open Source desktop is given too much importance. Windows server shipments still account for 60% of the market, so it’s not like that area is all sewn up. A company that wants to avoid vendor lock-in would do best to migrate its server infrastructure first – that’s gonna be least painful and probably highest long-term benefit. Then maybe desktop apps, the maybe desktop operating system.

    On MySQL vs. Firebird: yes, MySQL is more widespread, but they’re used for entirely different things.

  • http://dag.wieers.com/ Dag Wieers

    I’m a bit confused to why you want to differentiate between Linux (eg. Red Hat) and Open Source.

    Red Hat releases source packages and contributes largely to Open Source projects, both in resources as in code. Improvements by Red Hat are included in SuSE and vice versa. Everybody wins.

    This ensures that Red Hat will have to be the best on its own merits. Competition will always be lurking around the corner to take over. Despite that, Red Hat is doing a good job.

    You cannot compare this to proprietary vendors were your money goes into the big company bucket being used for the next version that you have to pay for again.

    If I can choose I’d rather pay for services, if it guarantees that the money is used for Open Source development. If my Open Source vendor goes belly-up, its work is still available for anyone to use.

    Paying for Open Source just guarantees you that you have freedom and are never tight to any vendor. Red Hat is just one example to show that the money is used for the good of the public.

    And if you don’t have deep pockets, there’s still Fedora, CentOS, TaoLinux or Whitebox. Plenty of competition in the same vendor segment. Hard to beat IMO.

  • wevervain

    “I recently spent some time speaking with a popular Yankee Group analyst”

    Why should an analyst’s social standing be mentioned in a technical piece? It has no bearing on the matter in hand if they are popular or not.

    Also why have you not named the analyst in question? I know of several analysts used by the Yankee group who have shown a complete lack of understanding of the realities of technological business matters.
    Without knowing the name of your source the public can not judge how much credence to give to your views.
    Perhaps that was the idea?

  • Ron Johnson

    One thing I notice that is never mentioned when talking about Windows vs. Linux TCO is virus & worm costs. Both the cost of AV s/w and clean-up after an infection sneaks into the corporate LAN. That *huge* expense will never be borne by a Linux shop.

  • Matthew C. Tedder

    Yes.. The commercial software industry has been migrating toward subscriptions, on-demand, etc. and especially professional services.

    These are actual ideal for Open Source and Free Software–which provide the advantages of self-marketing, the argument of multi-vendor support, full supportability, and full customizability.

    Self-marketing comes from (1) employees bringing the products in under the radar and avoid standard procurement processes (seeking support contracts, etc. after deployment), and (2) the ability to fully test the software before making an irreversible commitment.

    The argument of multi-vendor support is a helpful sales pitch. Of course, the reality of it is more useful to the customer. Marketing against commercial products though, this is a huge statement.

    Full supportability exists because you have full access to the source code–and, in most cases the developers.

    Full customizability exists because you have full access to the source code and can make any modifications. This is where professional services companies make the bulk of their income. IBM Global Services, for example, is the biggest breadmaker of the organization.

    Note: I think it’s very important to restate the saying that was popular in the earlier days of GNU/Linux… “Free as in free speech, not free beer”. No open source license (and definitely not Free Software) has even guaranteed free of charge copies to anyone. This largely occurs, but it is not a point of Open Source or Free software. This saying was created to emphasize the fact, because people weren’t getting it. Their not getting it even worse, today.

    Matthew

  • http://www.practicalapplications.net bwarrene

    No dark secret in not revealing the name of the analyst. I was having an extended discussion beyond an interview for another publication where the analyst was cited.

    Analysts are often called into question for their capabilities, however, we all need to remember, they base their reports on working directly with real world corporate IT. The statistics they generate are based on this research.

    Finally, hopefully I have reflected my advocacy for open source in this blog. That said, just as the ‘closed’ world gets roasted for its problems, we need to be willing to look at Linux and the open source community, good and bad, and fix issues when they arrive.

    We could spend the next five years shouting at how it is superior to the proprietary world. If it is not convincing to those holding the checkbooks and budget money, we have failed. Microsoft is successful for one reason (regardless of how it achieved it) – companies and individuals invest money in MS products.

    Linux and open source need to work toward the same – dollars flowing into the community to generate new research and development, central organizations to promote and support it, and most importantly, paying jobs.