J.D.'s BorCon diary: Wednesday, July 12
An ongoing, on-the spot report from the 11th Annual Borland Conference.
It's the last day of BorCon.
In previous years, I've always been an exhibitor. So I've spent the last day
of BorCon working in the Windows Tech Journal booth and missing all the
keynotes and other presentations. Then it's booth tear-down time. That's an
amazing thing...you wouldn't believe how fast the exhibit hall can come to
represent a loading zone. Imagine your bedroom while you're packing for a
two-week trip, but expanded by a couple orders of magnitude. The signs come
down, the booth walls collapse, burly guys come and roll up the carpet...before
you know it you're standing in a warehouse full of shipping cartons. It's really
something to see.
CLASSROOM REPORT
I made my second conference presentation yesterday. I called it "The
Linux Markets," and it was all about assessing the potential of Linux in
different segments of the computing market.
Linux is proven as a Web server platform. Apache has emerged as Linux's first
killer app, and the Linux-Apache combination accounts for nearly half the
servers on the Internet. Fast growth continues...this is a place where Linux
indisputably makes sense. (That's great because Borland has recently announced
that Kylix will have special library support and tools for creating Apache-based
applications._
There's a secondary sort of Internet server market that I find interesting.
When I think of Web servers, I think of apps like Apache that serve up Web
pages. But the Internet also connects application servers, audio-video servers,
chat servers, fax servers, list servers, mail servers, news servers, proxy
servers, and more. In each of these niches, there are Linux solutions.
The second major market for Linux is the IT market -- general-purpose servers
that run applications and store data. Linux is slowly making inroads in this
market. It's following the well-worn path of the original PC. Linux is brought
into the company for one specific task, usually running Apache. (Just as the PC
was brought in to run VisiCalc or 1-2-3.) Then, once the platform is there,
early adopters launch pilot projects or stealth initiatives to see if the new
technology can fit into the existing infrastructure. In the Linux world, this is
happening first in smaller companies where the operating system's low price is a
compelling factor.
How much progress has Linux made in displacing NT, Solaris, AIX, and other
operating systems on IT servers? As near as I can tell, no one knows. There are
a lot of reasonable assumptions and good guesses, but we await a market report
that will give us solid numbers.
The third market is the one that excites me the most. It's the embedded
market, where Linux is used in devices that aren't PCs or servers. This is a
huge market. Did you know that only about 10% of the world's microprocessors are
used in PCs and laptops? The rest -- one to two billion processors per
year -- go into data acquisition and control devices, Internet appliances,
telecom and telephony devices, set-top boxes, vending machines, point-of-sale
devices, games, and the like.
This is a huge market. "Everything with a digital heartbeat will be
connected to the Internet," says Scott McNealy, and I think he's right.
Consider your home thermostat. You probably have a smart thermostat that
turns the heat down at night when you're sleeping and warms up the house just
before you get up in the morning -- such devices have become so inexpensive
they're standard in new homes now. Wouldn't you like to be able to communicate
with that thermostat from your desk at work, in case you're coming home early or
working late? How long would it take for improved heating-fuel efficiency to
repay the cost of adding a cheap Internet connection to home thermostats? Better
yet, don't you think your thermostat ought to do a better job of anticipating
weather conditions and outside temperatures? What if The Weather Channel had a
data sideband that your thermostat could tap?
Or what if you had a bricks-and-mortar company that placed vending machines
in businesses and malls all over town. Wouldn't you like to be able to build an
app that queries the inventory level in all those machines? That's not how it
works now. Today, you hire a driver and buy a van, and you send him driving
around 40 hours a week to check inventory levels. Vending machines ought to know
when they're jammed, and they ought to send you an e-mail. How long would it
take to justify that simple application of Internet technology?
Everything's going to get wired. The bandwidth is coming, the hardware is
already laughably cheap, and Linux has exactly the right stuff for many, many of
these applications. If you want to bet on a Linux boom, this is the place for
it.
The fourth and final market I explored was the replacement-for-Windows
market, the desktop. In this market, Linux has some real barriers. Its
number-one competitor has a monopoly position. Linux is not especially
well-designed for use as a desktop operating system for the masses. And there's
little incentive for non-geeks to give up the operating systems they already
have.
Nonetheless, despite virtually no marketing and demonstrably primitive
distributions, Linux gained four percent of the desktop market in 1999,
according to research firm IDC. It's easy to scoff at that number, and in my
presentation I did. But Dan York of the Linux
Professional Institute stayed after class to chat about that. He noted that
Apple has spent 16 years pushing the Macintosh as hard as it could, and has
managed to achieve market share of five percent. Linux hit four percent without
a single full year of marketing.
The four percent number is also tough to verify. Linux is free, so it's
reasonable to guess that some percentage of the people who acquired Linux CDs
retired them to shelfware, maybe without ever installing them.
On the other hand, the IDC numbers do not include copies of Linux that are
downloaded over the Internet instead of distributed on CDs. And it doesn't have
any way of tracking the numbers when Linux is installed on multiple computers
from one CD.
IDC's numbers are the best we have, but there are still some wide holes in
our knowledge of Linux's market share.
WINDING DOWN
After I post this article I'll be done with my BorCon responsibilities except
for the Advisory Board wrap-up meeting this afternoon. Then, as conference-goers
pack up and head for home, I'll check my bags and leave on vacation. I'm going
to Russia for pleasure, can you believe it? I'll tell you all about it when I
get back.
Keep hacking.
JDH
Connect with Us