Open Source Commentary from Navica's CEO,
Bernard Golden
March 2007
In This Issue
-
The New Economics of IT
-
Navica News
The New Economics of IT
If you look at most IT organizations, they are trapped in
a labyrinth of legacy systems – a mix of applications
built on obsolete architectures, specialized home-grown applications
created by people who have long since left the organization,
and packaged applications from vendors who offer a seductive
vision of a completely integrated software stack that, once
entered, quickly turns into a House of Mirrors offering no
safe exit.
All of these components of the enterprise architecture have
one thing in common: they were built for the old world of
IT of underpowered hardware and expensive software. Because
yesterday’s hardware had so little power, it was critical
to create software architectures that used as few processing
cycles as possible. Every potential software optimization
had to be taken in order to save hardware
resources. Because it was vital to conserve those resources,
highly integrated software architectures, usually delivered
by an infrastructure software vendor, were desirable. And,
of course, endless hours went into hand-optimizing the overall
IT infrastructure, just to ensure minimally acceptable system
performance.
Infrastructure software vendors employed the best talent
available, so that the software could be optimized by software
geniuses to be shoehorned into the very expensive hardware
needed for minimally acceptable performance. And, since the
software companies had specialized capabilities due to their
talent pool, they could charge enormous premiums for the capability
they delivered.
Trained specialists were employed by end user organizations
to tweak the hardware components of the infrastructure to
ensure every precious cycle was rationed out to gain all possible
processing.
Essentially, end users paid gigantic prices to access the
work of elite talent. The result is obvious today: an infrastructure
designed to incorporate the complexity necessary to achieve
acceptable performance. Underlying the infrastructure design
is a whole set of assumptions about the skill sets of the
people working on the infrastructure, the processes to be
followed in managing the infrastructure, and the appropriate
costs to be invested in running the infrastructure.
Based on those assumptions, today’s IT organizations
stumble through their self-created infrastructure labyrinth,
shuddering each time it must be modified, terrified that they’ll
touch something and bring it all crashing down around their
ears.
There’s only one thing wrong with this: it doesn’t
make sense anymore. All the antediluvian economic assumptions
about the way IT should be managed are no longer applicable,
superseded by the new economics of IT: free software, cheap
hardware, and people too expensive to devote to endless, unnecessary
infrastructure tweaking. We need to move from a thoroughbred
perspective to a draft horse perspective, because today’s
draft horse IT infrastructure is plenty good enough.
Free Software and Virtualization: A Marriage
Made in Heaven
First off, free software. I have argued, at some length,
that the critical innovation of open source is how much less
expensive it is compared to its proprietary counterpart. The
other aspects of it about which people pour an overwhelming
stream of logorrhea out into the Internet – source code
availability, redistribution, expansive licenses – are
all important. In fact, they ensure that open source remains
useful by precluding the possibility of proprietary handcuffs
shackling open source products. These aspects ensure open
source stays open.
However, if open source had all these aspects and yet had
cost parity to proprietary alternatives, it would be a clear
loser. Most of these aspects impose some additional responsibility
and work – after all, improving a product through modifying
its source code by definition means you have to do some work
on it! If the underlying cost of open source was equivalent
to using a proprietary product, there would be little advantage
for most organizations in taking on the additional marginal
cost of these aspects of open source.
But when the software is free – well, that’s
a different matter. The marginal cost imposed by open source
pales in comparison with the cost of proprietary software.
Most people are willing to spend a dime to save a dollar,
and that’s the economic reality of open source.
Turning to hardware, the impact of Moore’s Law is hardly
news. Hardware gets cheaper every year. But we’re just
about to see an amping up of Moore’s Law. For most organizations,
Moore’s Law has run out of usefulness. The machines
they bought three years ago are still adequate to run their
software, so buying the four-times-faster machines available
today through the magic of Moore’s Law is hardly compelling.
But the recent trend of virtualization will change all that.
Instead of the “one application, one server” mindset
appropriate to a world of expensive hardware and fragile software,
virtualization can put software payload after software payload
onto a single box and still keep them isolated to ensure they
don’t interfere with one another. In a virtualized world,
you want to take advantage of Moore’s Law as much as
possible: the more powerful the hardware, the more you can
take advantage of it. Virtualization puts IT organizations
back on the efficient frontier of computing power.
While many in IT scoff at virtualization, preferring to cite
the complications it imposes (meaning, I have to learn some
new skills to take advantage of this, so I’d rather
stick with my vestigial practices so I can continue sleepwalking
through my job) rather than examine how it can transform the
cost structure of the hardware portion of the IT infrastructure.
Make no mistake about it: virtualization is no flash-in-the-pan
technology. I believe so strongly in it that I am writing
a new book on it: Virtualization for Dummies. Look for it
later this year.
Open source and virtualization are perfect complements. Combined,
they make the building blocks of IT infrastructure one hundredth
the price of- just five years ago. And when the price of a
good changes that dramatically, it forces change in the organizational
processes that are in effect.
The organizational processes in place at most IT organizations
are totally inappropriate for the new economics of IT. IT
organizations devote most of their time to tweaking software
to achieve peak performance. What time they don’t spend
on tweaking their infrastructure is devoted to fretting about
vendor management: how does Oracle’s Fusion plan affect
my five year strategy? Or, how can I stretch my budget to
cover the mandatory upgrade that provides little new functionality
but is forced on me by a vendor?
In a world of free software and cheap hardware, having people
devote time to obsessing about saving cycles and engaging
in shoving matches with suppliers is mindless. People are
too expensive to have them focus on efforts with miniscule
payoff. IT organizations should be re-evaluating the opportunity
cost of their employee’s effort. Instead of wasting
time on patching their existing infrastructure, they should
be planning how to tear down entire sections of it and replace
them with infrastructure better suited for the new economics
of IT.
Next month: How IT organizations can take advantage of the
new economics of IT.
Navica News
You can hear me speak at these upcoming events:
March 20, 1:00 p.m.: "Building an Open Source Business",
SDWest, Santa Clara, CA, joint presentation with Bill Weinberg
March 20, 5:30 p.m.:"War
and Peace in the Open Source Era", Association
of Strategic Alliance Professionals, Sunnyvale, CA
May 10, 10:00 a.m.: "Open Source Virtualization: Creating
an Action Plan", Red Hat Summit, San Diego, CA
If you are interested in having me speak at your
organization:
Contact me directly via email.
You might be interested in reading my blog posts
at CIO Magazine:
Open Source Odds
and Sods: Sightings from the Ecosystem
Oracle's Unbreakable
Linux Unleashed on an Uncaring World
|
