Open Source Commentary from Navica's CEO, Bernard Golden
November 2006
In This Issue
-
The Open Source Underground Economy
-
Virtualization Seminar: Virtualization
Smackdown Debrief
-
Navica News
The Open Source "Underground
Economy"
Whenever I go down to my local Home Depot, there are dozens
of Latino men hanging out in the parking lot. They were never
there when I first moved to my community 15 years ago; their
presence reflects the social changes in the surrounding cities;
Silicon Valley has seen enormous immigration from Mexico and
other Latin American countries over the past few years.
These men are standing around waiting to be hired for casual
day labor. If you need digging, carrying, lifting, or the
like, striking a bargain to get it done takes no more than
a minute. And it's not just homeowners with a small job that
employ these guys. Contractors doing commercial jobs rely
on them; indeed, they may make up the bulk of the employment
deals struck in the parking lot.
It goes without saying that this work is done under-the-table.
The work is paid for on a per-day basis in cash. No taxes
are paid by either party. For the worker, it's better money
than they could get in their native country; for the employer
its better quality work than they could get from a native
worker. At the end of the job, each party goes its separate
way, satisfied with the arrangement.
Of course, this type of work is by no means confined to non-legal
immigrant workers. Anytime a plumber does a job “on
the side” or a hair stylist stops by your house and
gives you a haircut for cash, it bypasses the official tax
and legal system. This kind of thing isn't even limited to
services. Many times I've been offered a better deal on a
product if I would pay cash – with both parties understanding
that the transaction would never go onto the company's books.
The formal name for the aggregation of this type of work
is “the underground economy.” Depending upon your
perspective on the employment of people working in this country
outside the legal requirements, this segment of the underground
economy is a boon for workers and employers or is a unacceptable
avoidance of legal requirements regarding taxes and benefits.
Whatever your perspective, this is a significant portion
of the overall economy. It's critical to understand that,
despite its absence from the “real” economy, the
underground economy represents real work done by real people
that generates an economic benefit. While the absence of formal
wages and taxes means that the underground economy does not
generate official statistics that are incorporated into overall
economic numbers, there are estimates as to its overall size.
To form these estimates, various techniques are used, including
looking at the amount of cash in circulation and comparing
it to the overall economic activity of the nation, although,
as Wikipedia
points out, that measure does not capture the more genteel
underground economy activities of the offshore tax and banking
havens. Overall, estimates of the size of the US underground
economy range to nearly one billion dollars. (See here
for a good recent piece on the topic).
All well and good, but you're probably asking “What
does this have to do with open source software?” Just
this: Similar to how actual work gets done by the casual laborers
hired outside Home Depot, but nothing official ever ends up
in the national accounts, open source software gets deployed
and performs a useful function even though you'll never see
it counted in the official statistics of the software industry.
This curious situation has caused some real problems for
software industry watchers. For example, for several years
IDC ran surveys to estimate how much Linux was being used
by IT organizations. However, it formed the estimates based
on servers shipped by major vendors with Linux pre-installed
on the boxes. The very significant presence of Linux on repurposed
machines (or indeed, on self-built or whitebox machines) made
no impact on the estimates. There's no question that IDC's
estimate significantly underestimated the use of Linux in
data centers, all because their survey methodology, very appropriate
for the traditional software market, was ill-suited for the
new mode of software distribution and use created by open
source.
IDC was at it again recently when they released a survey
of the virtualization software market. Their study indicated
that VMware was far in the lead with its $310 million in 2005
revenues representing 55% of the virtualization market. Microsoft
achieved $49 million in revenues and around 8% of the market.
Xen was lumped in with “others,” who only make
up 3% of the market with $12.4 million in revenues. Again,
IDC's estimating methodology undoubtedly missed significant
use of Xen, although it must be acknowledged that this case
is not as egregious as the Linux example cited above, since
Xen is a relatively new technology just beginning to grow
its installed base.
IDC is not the only party to judge a market solely by revenues
and thereby interpret it in a disadvantageous way to open
source products.
My friend Bill Snyder of TheStreet.Com wrote a story
(although, admittedly, it was based on a report by another
analyst firm, Gartner, illustrating that analyst myopia is
not limited to one firm) discussing the database market. He
noted that open source databases achieved $100 million of
the total database market sales of $13.8 billion; in other
words, a miniscule .007% of the market.
I'll give Bill a pass, since he's a financial reporter and
it's understandable that he would examine the database market
through the lens of its numbers. But I am much less forgiving
of the analysts. Since they should know better, I can only
conclude that they are willfully ignoring the real world evidence
in front of their eyes regarding the actual usage of software,
choosing instead to focus on aspects of software markets that
bear an ever-decreasing relationship to the actual presence
of particular products in working software infrastructures.
And really, this is no longer even an open source-only story.
Measuring installed share of virtualization software by revenues
overlooks the fact that VMware itself has now begun to distribute
a very capable virtualization product, VMware Server, for
free. I have been doing a lot of speaking on virtualization
recently, and can attest that many of the audiences I present
to have people whose organizations are running VMware Server
in production.
And VMware Server is by no means unusual. Free (as in free
beer) software is increasingly a central part of commercial
software strategies, following the razor-and-blade model of
giving something away in expectation of forming a lasting
economic relationship with a customer. However, according
to the estimation methodologies cited above, these free software
products don't exist.
However, just as the underground economy's true size can
be seen in its impact on other elements of the economy, similar
to how a dark star's existence can be confirmed by the gravitational
perturbations of other bodies, we can see the evidence of
open source's importance in other behaviors in the software
market place.
In fact, Bill Snyder's story contains a perfect example of
how open source's importance can be estimated. In his story,
he notes that Gartner states:
“Open-source database sales have lagged, while sales
of the Linux operating system offered by vendors like Red
Hat and various open-source middleware have grown strongly,
in part, because few third-party developers were willing
to provide the software tools needed to support the database.”
He then notes that the report goes on to say that this is
changing, due to vendors like Embarcadero Technology making
their tools available for open source databases.
But really, isn't this backward? Embarcadero Technology and
its brethren are unlikely to modify their tools to work with
unimportant players in the market. I mean, why would they
devote any work to an also-ran product? Isn't is much more
likely that the causal chain is the other way around? That
Embarcadero Technology is making its tools available for open
source databases because their real-world experience is that
the open source products make up a significant part of the
potential user base for their product, and customers are demanding
that they make their tools work with MySQL and other open
source databases?
I think the answer to that question is obvious. I've worked
in a software company deciding whether or not to support a
new partner product, and the decision is never taken lightly.
The partner product has to have significant presence before
a software company will even consider integrating with it,
and the company will evaluate the likely long-term presence
before finally pulling the trigger on a go-forward decision.
The entire software industry is being turned upside-down,
and trying to evaluate it using the tools appropriate to another
age is to miss the forward-looking trends that will distinctively
shape tomorrow's IT industry. I hope the pundits don't continue
to overlook the new software “underground economy”
and continue to mislead the industry about the new marketplace
realities.
Virtualization Seminar Debrief:
Virtualization Smackdown for Real
If you didn't get a chance to attend the virtualization seminar
I featured in the past two newsletters, you missed a treat.
We had a first-class panel with representatives of Microsoft.
VMware, Cassatt, and XenSource discussing their perspectives
on the topic.
I was really impressed with the quality of the panelists.
The companies sent along senior people and some real news
got broken there. Jeff Woolsey, program manager for Microsoft's
virtualization initiative, shared that the main focus of Microsoft's
virtualization work going forward would be Viridian, their
hypervisor-based product, with Virtual Server being put out
to pasture.
And, the event really did turn into a smackdown. The guys
from VMware and XenSource took some pretty good shots at one
another, trading barbs regarding the shortcomings of their
competitor's products. You don't often get to see the gloves
come off at one of these panel events (in fact, the usual
experience is one of tedious boredom), but you should have
been there. It was great. You can read my take on the event
here.
Navica News
You can hear me speak at these upcoming events:
November 7/8, 8:00 a.m.: Day 1 Chair and Speaker "Evaluating
the Business Models for Open Source Software in Mobile",
Open Source in Mobile Conference, Amsterdam -- Register at
the Informa
Telecoms and Media website.
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:
I'm From IT and
I'm Here to Help -- Part II
Open Source Special
Purpose Appliances
Too Soon for Red
Hat to Run Up the White Flag
|
