Can open source stop navel gazing and get real?
Enterprises and other users deploy open-source software because it works. For those of us in the open-source vendor community, however, too often we waste time talking about issues that have relatively little resonance for the vast majority of users.
We miss the mark on open-source marketing. In fact, it's often the case that the very standards we seek to set for the software world--interoperability, transparency, etc.--are better observed and delivered by open standards than by open source.
As a case in point, Red Hat and other open-source companies (including Alfresco, my employer) routinely advertise "no lock-in" as a key reason to buy open source. There are two problems with this marketing pitch: one, it's only technically true, and two, customers don't care, as Redmonk's Stephen O'Grady recently noted.
On the first, it's true that open source can reduce vendor lock-in by ensuring that a customer can get support and ongoing code development from someone other than the original developer of the software. But this is only trivially true.
Once a customer invests in a particular vendor (be it Red Hat or Canonical or Novell or MySQL or...), there will always be a cost associated with leaving that vendor, a cost that arguably isn't much different whether that vendor's code is open source or proprietary.
Cost aside (which is easier said than done, as cost is the primary consideration for the buyer), the support options for Vendor X's code from Vendor Y or Z are unlikely to be on par with what Vendor X can deliver. Just ask Red Hat about CentOS or Oracle Enterprise Linux support. ("Compatibility with Red Hat Enterprise Linux can only be verified by Red Hat's internal test suite.")
Apparently there's no lock-in...so long as you stay with the original open-source developer. :-)
The reality is that open-source vendors should be pitching real value to real customers. As Josef Assad presented at the Open Source Days 2008 conference, open source should strive to "lose the TCO (total cost of ownership) war with proprietary vendors." Open-source value is about performance and flexibility at a great price--and not necessarily about absolute freedom from lock-in.
Red Hat gets this. That's why most of the time it sells the value of its subscriptions, and not the hocus-pocus "no lock-in" story. Red Hat doesn't have 75 percent of the paid Linux market (or, probably more accurately, 62 percent, according to IDC) because of its lock-in story.
Would-be customers don't care about that. Really. They just want Red Hat's performance and price, especially compared with Unix.
In fact, to the extent that customers really do want interoperability and reduced vendor lock-in, it's open standards that they should be asking for, not open source.
IBM's Savio Rodrigues points this out in his analysis of the different permutations on the open-source WebKit project. Serdar Yegulalp adds to the analysis:
Source code is a building block, not a standard. It's something you turn into other things. A standard is something that stands above and apart from all of those things, a guideline for what that finished product ought to be like....
The problem with using code as a standard is simple: it's too fluid. The minute you implement it in something, it's not the same code anymore. It almost always has to be changed to fit its container, as water changes to fit.
Open source is an indispensable complement to open standards, but it's not a substitute for them.
This isn't the only area where open-source vendors misread customer tea leaves. For years open-source insiders have debated definitions for "open-source vendor," even as customers shrugged their shoulders and continued using open source--from different vendors with very different business strategies--without worrying about the various shades of ideology and pragmatism that fuel open-source development. I'm as guilty of leading this foolish march as anyone.
Real customers simply don't care.
This is why I think Sun open-source guru Simon Phipps' proposed expansion of the Open Source Initiative's charter is misguided, though very well-intentioned. (The 451 Group's Matt Aslett also weighs in on the proposal.)
Phipps wants the OSI to establish a "holistic vision of software freedom against which businesses can be benchmarked" because too many companies, apparently, are calling themselves "open source" without a consistent definition for what this means.
I don't think it matters. The reality is that businesses don't seem to have any trouble adopting open source regardless of such "truth-in-labeling" initiatives. Gartner suggests that 85 percent of businesses are already using open source. Forrester tells us that a big majority of enterprises are adopting open source because it's delivering real cost and quality benefits to them.
And so the problem is...?
Well, the problem is that open-source advocates are often out-of-sync with open-source adopters. We probably need a new breed of open-source advocate, as ZDNet's Jason Perlow suggests, the kind that reflect customer interests in pragmatic adoption and not advocates' interest in controlling and fine-tuning that adoption.
We don't need paternalistic oversight of open-source adoption, and we don't need to fuel it through vague and inaccurate marketing. Open source is a fantastic way to develop and distribute software. Customers recognize this and don't need to be cajoled or confused into buying.
Matt Asay brings a decade of in-the-trenches open-source business and legal experience to The Open Road, with an emphasis on emerging open-source business strategies and opportunities. Matt is vice president of business development at Alfresco, a company that develops open-source software for content management. He is a member of the CNET Blog Network and is not an employee of CNET. Disclosure. You can follow Matt on Twitter @mjasay. 
Also saying 85% of companies is using Open Source . . . well what are they using it for?
If the business is big enough (say, 50 employees or more), one would hope you had a competent admin on staff, or at least had one on retainer.
Otherwise, RedHat and Novell (and Oracle) have support packages that solve all of the questions you've stated.
We don't win with 'no lock-in' and we won't win with 'we cost less'...we will win with a combination of superior technology, lower cost and more flexibility.
Err, okay...
* you're asserting that data is locked-in to one vendor (say, RedHat) even if the customer wants to go to SuSE, Oracle OpenLinux, or etc? Doesn't make any sense there, Matt. Even the apps can be transported, so long as they're written to static libs, as opposed to distro-specific ones (yes, locations for those libs might change between distros, but that's what "ln -s" is for ;) ).
* In general, customers may or may not care. OTOH, if they're locked into one vendor (as opposed to multiple vendors for the same type of OS), they do tend to care as the bills pile up.
At some point you might consider that there's a link between the unlockinable nature of F/OSS and the lower cost you keep going on about.
And that's the big problem, as a free software developer I have the needs of users and fellow developers as my priority.
Customers, not so much..
>(or, probably more accurately, 62 percent,
And that link sends you straight to the Novell site...
In a 'he said, she said,' debate about how much Red Hat is kicking Novell's ass in the market you picked Novell's numbers, so at least own up to it.
Not saying either is right, just interesting how youd skip that part.
>according to IDC) because of its lock-in story.
See, you can own up to it.
Of course, IDC gets every single FLOSS numbers from some dark nether region so no one really buys their spiel.
"Once a customer invests in a particular vendor (be it Red Hat or Canonical or Novell or MySQL or...), there will always be a cost associated with leaving that vendor, a cost that arguably isn't much different whether that vendor's code is open source or proprietary."
That's an extremely dubious assertion. I haven't seen a really comprehensive study on it, but I would suggest it's extremely likely to be cheaper for an enterprise to move from Red Hat Linux to Novell Linux than it is to move from either one to Windows, or vice versa. This is partly a function of technical similarity, true, but it's also partly a function of open source rejecting lock-in. There's no way Red Hat or Novell can use proprietary formats or database structures or whatever to lock your data to their applications. As open source _ensures_ those formats must be open, it will always be possible to convert the data relatively trivially (or at least with the minimum difficulty necessitated by their technical differences, rather than with a level of difficulty artificially inflated by lock-in techniques which have nothing to do with intrinsic differences between the formats).
"Just ask Red Hat about CentOS or Oracle Enterprise Linux support. ("Compatibility with Red Hat Enterprise Linux can only be verified by Red Hat's internal test suite.")"
Well, of *course* that's what we'd say. We want your money, after all. You will notice, however, that CentOS and OEL both exist and are actually used, in the real world, on real servers, by real people. Oracledows and Centdows are both noticeable by their absence. Our (Red Hat's) party line about OEL and CentOS is only sustainable as long as it's actually based in fact - as long as we actually provide better support for our own product than CentOS or Oracle do. If that were to cease to be the case, our publicity just wouldn't work any more. There is genuine competition in support for RHEL-based operating systems, even if we're currently winning that competition in the perceptions of most. If we suddenly jacked up our costs 500%, put everyone on hold for ten hours when they called, dropped the ball on hardware certification and security fixes, our customers _would_ be able to move to CentOS or OEL - or keep running RHEL bits and source support and updates from CentOS or Oracle! - with relative ease. This is simply not the case for Microsoft and Windows. The 'pain bar' - the point at which it's less hassle to switch operating systems or support providers (well, you can't do the second without doing the first, when it comes to Windows) than it is to keep dealing with the level of support you're getting - is massively higher for a Windows shop than a RHEL one (or a SUSE one).
"Would-be customers don't care about that. Really. They just want Red Hat's performance and price, especially compared with Unix."
This is ultimately true, but the lack of lock-in is an important factor in ensuring RH continues to provide good performance - and support and security - at a low price. The fact that open source makes it practically impossible to create vendor lock-in ultimately keeps RH honest. We _have_ to provide good support at a competitive price, because the threat of competition in our own back yard, as it were, if we don't is significant.
Blanket statements are almost always wrong. Surely you know that, so why make them?
What point are you trying to make, anyway? That software marketing largely sucks, but that customers buy the stuff despite vendors' lame positioning statements because they have to get some work done? Give a concrete example of what you would rather see vendors say. And explain why more information in the form of a broader open source certification is a bad thing. Do you think it will make customers make bad decisions? If so, why?
- by pentest October 15, 2009 1:15 PM PDT
- Customers do care about lock-in. Intelligent ones anyway. They know that when they buy MS SQL Server, that not only are they going to pay big today, they know they are also going to pay big down the road, when they upgrade hardware and are forced to upgrade that expensive MS crap.
- Like this Reply to this comment
-
(16 Comments)Those that don't know it or care, will feel the pain when the day arrives, and they will learn the hard way.