My presentation at the "Paving for eFuture" conference in Reykjavik 13.09.07, in a debate with a Microsoft representative
Citation preview
1. Open source / free software vs proprietary software what is
best for business?
Kaido Kikkas Tallinn University *** Estonian IT Society Paving
for eFuture Reykjavik, September 13, 2007
2. For those unfamiliar with the free world...
... these terms and concepts are worth studying:
free software
open source
GNU General Public License
copyleft
hacker
hacker ethic
Linus' Law on work motivation NB! Due to the presentation's
small timeframe, additional arguments, data and links are provided
by the complementary webpage (including these slides) at
http://www.kakupesa.net/kakk/docs/reykjavik2007/
3. Looking at the title...
I'd like to ask about a small detail: Best for WHOSE
business...?
I try to keep the user's (as opposed to vendor's)
perspective
Plus, in the next slides I try to look at a business
considering a shift from proprietary to free modeland give some
arguments to support the decision
4. What would a business expect from its IT?
doing the necessary thing
reasonable acquiring costs (esp. SME)
reasonable running costs
reasonable (re)training costs
reliability
interoperability
security
support
flexibility, extensibility and scalability
choice of services and providers (no lock-in)
5. 1. doing the necessary thing
All software regardless of licensing model can be appropriate.
Or not.
A caveat when moving to software with different licensing
model, do not assume that all your previous knowledge remains
valid. Or to put it simpler Linux is NOT Windows
Yet the added benefit of open source code allows for better
modifications. Also, the market is open one can opt for in-house
improvements or choose the best partner instead of paying extorting
prices to a market dominator
6. 2. acquiring costs
The smaller the enterprise, the more important (typically)
issue it is
FOSS results in large savings in this stage (probably not
denied even by proprietary vendors)
But even in the free world, going blindly for the seemingly
cheapest option may not be wise
In the business world (somewhat opposed to the NGO, education
and private spheres) using commercially-backed solutions (which may
cost quite a lot) may be justified. But not always an important
factor is the in-house IT capacity
7. 3. running costs
Similar to the former, but has more variables in it
Leaving your homework undone may sometimes hit quite hard
Due to the increasingly unreasonable 'intellectual property'
system, may run into various artificial obstacles (patents etc)
when not careful. In Europe, the problem is much smaller than in
the US
Earlier, finding qualified staff was somewhat an issue (not
much anymore, but depends on the location)
8. 4. (re)training costs
Can be substantial when moving large numbers of employees to a
new platform
In essence, do not depend on licensing model
Often cited as a prohibitive factor in moving to free models
yet the same applies to proprietary systems
E.g. for a typical desktop user, moving from MS Office 2000 to
the new 2007 is arguably more difficult than moving to
OpenOffice.org
Free model can result in more flexible training again, you do
not need Authorized Trainers
9. 5. reliability
Depends also on the maintenance skills of the tech staff
incompetent technicians can work wonders (in a negative sense)
Free systems (e.g. BSD or Linux) have excellent reliability
marks worldwide
10. 6. interoperability
Proprietary systems tend to be interoperable as long as you use
the products of the same company
Sometimes interoperability is considered directly
counterproductive to the company's goals (the earlier case of MS
Office documents, or also the current OOXML debate). Conflict of
interests?
Free systems have more potential here, although it should not
be taken for granted in some cases the initial author does not have
need for it and thus will not stress it
Open standards are the key but more than often, there is a
strong correlation with software freedom
11. 7. security
A long-time plague in MS software: Trojan horses and viruses
are 99% Windows-specific (in fact, a Linux virus is like the Yeti
some people claim it exists. Never seen one yet)
Being locked into a single platform also contributes towards
weaker security an attack will only need a single vector
A side remark: regardless of platform, the biggest security
risk is always located between the keyboard and the chair => a
training issue
12. 8. support
At the first glance, this one is a clear win for proprietary
systems. Linux has no support...
Actually, surprisingly large number of free systems have
commercial support available. Moreover, the market is open (again)
and thus it is much harder to charge excessive sums for support
services
Support can be obtained both in a traditional way (by
purchasing the software; e.g. Red Hat) or from third parties
13. 9. flexibility, extensibility and scalability
Clearly better in free systems. Examples:
Most of the Top 500 supercomputers run Linux
Free NetBSD operating system supports more than 50 hardware
platforms
Flexibility is an important factor in open source, so is
extensibility. Both stem from the lack of either technical (lack of
source code) or legal (prohibitive licensing) obstacles
14. 10. Choice
Monoculture is dangerous both in biology and in technology
(some call it inbreeding)
Proprietary vendors often strive to create large, unified
solutions on a single (their own) platform, leaving it more
vulnerable to threats
Also, having achieved a lock-in on a customer, the vendor is
able to charge remarkably higher prices than in the case of open
market
15. Where proprietary approach may make sense
In highly professional, specialised fields with turnkey
solutions handed out (e.g. composers)
the client can afford to pay for support
the client's time is expensive losing access to his/her tools
would cost much more than calling for a specialist
But even here I'd consider a free approach for greater
flexibility and playing room for support
The more common the application, the more obvious should using
the free model be
16. Personal opinion: if I had a business
I'd run my IT sector roughly as follows
MS Windows only where specific applications demand it;
preferrably also locked into a separate network cluster; prefer XP
over Vista as long as possible; using free applications on Windows
where possible (app compatibility)
MacOS X is an option for presentation/sales
The rest would run on free systems (exact methods support etc -
depend on circumstances)
And I would be far from the first one doing that
17. Conclusion
Free models have been discussed from a variety of viewpoints in
this presentation we left aside ethical and social issues and
focused on professional ones only (my personal reasons to avoid
proprietary software are 50/50 a business decision and an ethical
statement). But even these are sufficient