Open Source Could Suffer from Exaggerated Expectations

There is an article/advertorial floating around the Internet by Roger Burkhardt, CEO of Ingres, that asks the question, “Can Open Source Help the Economy?”.

I am afraid Burkhardt is suggesting possibilities that open source software terms, conditions and marketing techniques cannot deliver. As such, it sets up expectations which—when not met—will spread more misinformation and doubt about the open source culture and movement. Open source cannot relieve “pressure on global IT budgets” in any kind of significant way for two reasons, neither of which can be controlled by the open source culture and its fans in the supplier world:

1. First, annual “new-new” licensing of software is a small percentage of the overall IT spend in any given year (about .5%--$20 billion of $5 trillion--depending on whose market-size spending numbers are used). That means the opportunities to save vis a vis an individual IT budget are also small, even if Burkhardt is saying open source could replace the "new-new" license spend totally (very unrealistic). IT budgets consist mostly of personnel costs, overhead, systems and storage hardware acquisition and maintenance, implementation and other types of consulting, IT-based services such as software as a service (SaaS), maintenance on existing installed software and “new-old” licensing, where new but required versions of installed software are not included in the maintenance fee. Of course, Burkhardt's company naturally treats the latter two types of spending as sales opportunities but very few such situations turn into migrations simply because of human nature and the odd happenstance that a large number of IT users are happy with their software.
2. Second, there is a significant personnel cost (training, implementation time, and so forth) associated with bringing in open source software and no one in the open source movement would tell you otherwise. In addition, typically—but not always—a maintenance support cost equal to the maintenance fees associated with any other kind of software is involved. As with point 1, anyone looking at total costs of ownership is not going to find a large nut to trim. The open source movement is always careful to position its deliverables as free as in speech (or some such similar analogy where speech isn't free), not "free as in beer."

Burkhardt cites 25% savings found by Forrester. But I would guess they are realized against the half of a percent mentioned in point 1 above. So now the "global IT budget savings" that could be realized with open source software are down to around a tenth of a percent. Every little bit helps, as they say, but don’t promise “much lower costs.” (The 25% savings on hardware costs he mentions could be realized by using any software on commodity servers.)

Even if “proprietary vendors” were foolish enough to raise license fees by 45% (unlikely in a down economy), everyone solely competing in the software market--whatever the terms and conditions of their licenses--is competing for this sliver of the "global IT budget."

And what is a "proprietary vendor" anyways? Ingres' top competitor Oracle (ORCL) just published how it's involved in the open source culture and its activity probably dwarfs Ingres', simply because of scale.

So the answer to his question is "No, open source cannot help the economy very much, and should not be expected to." For Linux in particular, publicly available IDC research illustrates that Linux is replacing older Unix versions as would be expected because of the way IBM, HP (HPQ) and other systems suppliers promote Red Hat (RHT) and Novell (NOVL) as a replacement for their own heritage Unix software (Q3 IDC data is available but I have not yet posted on it; the trend remains the same). I do not see any suppliers making a similar recommendation for the “higher level software applications” that Burkhardt mentions. There is one systems supplier making a recommendation like that vis a vis Oracle RDBMSs, but I don't see how Sun (JAVA) pushing MySQL helps either open source or Ingres.

Leave open source to what it can deliver--more open development processes that save suppliers (trickling down to users) research and development expense. As it has delivered for the last 50 years.

Comments

[Tony Barreca:]

Thanks, Dennis. Too bad that the Sun senior management team didn't do the business analysis that you do in this article. Had they done so, they might have realized that they needed a plan rather than just a religion based on open source.

2008 Dec 17 02:05 PM

[bigmoney:]

I've been a UNIX / Linux enthusiast for more than 15 years. However, I do recognize it involves a significant learning curve, that even talented Windows techs are often not willing to bridge.

It's not just learning curve though. Windows O/S, and Windows apps, such as MS Office, Exchange, IIS, SQL Server, anti-virus, etc. are often provisioned in a "gray" market, and are as "free" to the user as anything open source. Windows techs typically run many PCs and laptops at home, each with virtual machines, and if they actually purchased licenses for each, they would go broke. So they get the best of both words. Minimal learning curve. Minimal software expense.

2008 Dec 18 12:45 AM

[Seeking Tru...:]

Agree this argument is not a strong one. But it is a red herring for both open source opponents and proponents to use for their own gain. This is a distraction from the real value of open source.

The economic benefits of open source come over the long term rather than the short term, primarily because OSS frees IT shops from vendor/data lock in, and the astronomical ELA costs that come with that lock in. Since you can leave one OSS product for another, there is more choice and competition, always good for the economy and for budgets. And, of course, there's the basic cost benefit that it usually costs less to acquire (download), although training and support is not free. Even then, support and training for OSS will be priced lower than proprietary vendors, since the code is available for anyone to view and support - not a vendor who owns the source code.

Standards-based, vendor-backed open source software has a multitude of benefits over proprietary software that go beyond economics (most notably: security) and are much more important than whether or not open source software can "save the economy."

And let's be honest - if Oracle were to truly get involved in the open source community (much less commit to an open source model) their shareholders would run for the hills and the company would go bankrupt. It conflicts with their core business model. They may contribute code and some time to community projects, but do they truly provide enterprise class open source products, much less support them?

2008 Dec 18 12:13 PM

[FUD Redux:]

If you believe that the cost of switching software is drastically reduced just because the software is open source, you must be smoking dope. Switching from one piece of software to another (whether OSS or not) usually comes with a huge amount of perceived (and possibly real) risk; applications may stop working; staff skills go out the window and have to be replaced. Not to mention the user revolt problems that may occur if the software you are replacing has a large end-user population. The fact that the code is open source has little implications for most IT shops. There are a few companies (like Google, Yahoo, etc.) that will actually hack the source code, but most standard IT shops don't have the resources and would rather have name-brand vendors support their mission critical apps. And you have to pay for good support no matter what. So that leaves us with the question whether open source software would have cheaper support costs merely by virtue of being open source (as opposed to the specifics of the level of competition in a particular market). One could argue that the fact that the source code is available would allow anyone to start a company providing support for a specific piece of open source software thus increasing competition. Well, Oracle tried that undercutting Red Hat on their own Linux. Maybe it has helped to keep Red Hat's fees in check, but as far as I can tell, Oracle's cheap support for Red Hat has not been a huge success. And if a big name vendor like Oracle can't siphon support revenue from Red Hat, who else can?


On Dec 18 12:13 PM FUD Killer wrote:

> Agree this argument is not a strong one. But it is a red herring
> for both open source opponents and proponents to use for their own
> gain. This is a distraction from the real value of open source.<br/>
>
> The economic benefits of open source come over the long term rather
> than the short term, primarily because OSS frees IT shops from vendor/data
> lock in, and the astronomical ELA costs that come with that lock
> in. Since you can leave one OSS product for another, there is more
> choice and competition, always good for the economy and for budgets.
> And, of course, there's the basic cost benefit that it usually costs
> less to acquire (download), although training and support is not
> free. Even then, support and training for OSS will be priced lower
> than proprietary vendors, since the code is available for anyone
> to view and support - not a vendor who owns the source code.
>
> Standards-based, vendor-backed open source software has a multitude
> of benefits over proprietary software that go beyond economics (most
> notably: security) and are much more important than whether or not
> open source software can "save the economy."
>
> And let's be honest - if Oracle were to truly get involved in the
> open source community (much less commit to an open source model)
> their shareholders would run for the hills and the company would
> go bankrupt. It conflicts with their core business model. They may
> contribute code and some time to community projects, but do they
> truly provide enterprise class open source products, much less support
> them?

2008 Dec 19 04:24 AM

[Seeking Tru...:]

Response to above, "If you believe that the cost of switching software is drastically reduced just because the software is open source, you must be smoking dope."

I never said the cost of switching was drastically reduced. This just goes to Byron's ongoing rant that people read what they want to read, versus what is actually written.

I said that the economic benefits of OSS come over the long term, not the short term. When you can get away mandatory, costly OS/productivity suite "upgrades" every few years (that eventually render your old document formats useless - what's the cost of losing your data?), and when you can pay for the support you need versus costly, mandatory database ELA renegotiations every 3 years, you will save money and have much more flexibility for the long term.

And it sounds like we agree on how OSS should be supported (by a vendor with enterprise/mission critical support), and how the open source model fosters competition to keep support prices fair. Anyone who thinks open source support should be free is, as you say, smoking dope - if you are running mission critical apps, open or not, you pay for support (from a vendor or via salaries for in-house employees). You are right about Oracle - compared to their overall business model the Linux support business is beans. But it certainly benefits end users. I am willing to bet that 10 years from now that will be the more viable business model, as more and more enterprises wake up to this opportunity (and vendors come up with ways to move 30 years worth of locked-in data to open standards/formats).

Never did I say that switching costs are reduced, drastically or otherwise. It does cost money upfront to switch, train, and engage in different support models (and like Byron said, open doesn't mean free) - but that upfront cost should be dwarfed by the money you will save within the first 3 years alone (after you skip the next forced upgrade or ELA renegotiation), and get even better over the long term.

Change always costs money but if you are executing the right strategy and thinking beyond the next purchase order it means more savings, profits and competitiveness for the long run. Think like a senior executive, not an first-year accountant.

2008 Dec 19 08:41 AM

[Tony Barreca:]

"In the long run, we are all dead."

--J.M. Keynes

2008 Dec 19 12:04 PM