Open Source Commentary from Navica's CEO,
Bernard Golden
July 2005 Newsletter
-
-
-
See Bernard Speak at LinuxWorld
Hell Freezes Over Redux
The May
2005 Navica Open Source Newsletter discussed two events
illustrating the increasing importance and adoption of open
source. The first was the review of OpenOffice Associated
Press published in a number of newspapers around the US. The
publication of this review in mainstream press indicated that
open source is breaking free of its early adopter base. That
such an establishment institution as the Associated Press
was willing to lend its credibility to OpenOffice is extremely
telling about the progress of open source.
The second event that indicated the dropping temperature
in the nether regions was IBM's purchase of Gluecode. Gluecode
is an company focused on the open source J2EE application
server Geronimo. Many people read the purchase as a response
to the growing adoption of JBoss, another open source J2EE
server.
While the purchase is certainly a shot across JBoss's bow,
more importantly it also indicates IBM depth of belief in
open source. It is so committed to open source as the direction
of software that it is willing to compete with its own J2EE
server, Websphere. I know from experience how difficult it
is to introduce technologies that compete with a company's
existing offerings, so the purchase of Gluecode definitely
qualifies as a "hell freezes over" event.
A couple of weeks ago another event occurred that underlined
the meaning of IBM's Gluecode purchase. Sun announced that
it would join IBM and JBoss and release its J2EE application
server under an open source license.
Let's be clear: Sun's application server is a minor part
of the J2EE market and an insignificant part of Sun's revenue,
so this move doesn't carry financial implications for Sun,
unlike IBM's GlueCode purchase.
However, it represents a tipping point in the J2EE space.
With Sun's move, there are only two vendors pursuing a commercial
J2EE play: Oracle and BEA. Of the two, only BEA truly builds
its business on J2EE (Oracle bundles their server with their
products, so more or less gets market presence by riding along
with the database and apps).
With Sun's move, open source now represents the majority
of major offerings in this market. Proprietary J2EE products
are the minority. From this point forward, open source will
rule the J2EE market. New J2EE implementations and ISV J2EE
bundling will all be open source. Oracle and BEA will be left
with harvesting past design wins and installed base maintenance
streams.
Frankly, this is surprising. There is still relatively rapid
evolution in the J2EE space -- the underlying specification
is still being improved, which would seem to provide room
for commercial vendors to innovate and realize revenues. It
seems that a fairly static product category -- something like
SQL databases -- would have been the first one to migrate
to open source dominance. Nevertheless, hell has truly frozen
over. Open source has invaded and conquered a commercial software
sector.
Open Source Skills: Solving the Wrong Problem
A lot of discussions about open source in general, and open
source TCO in particular, focus on the challenge of finding
open source skills. I recently got a piece of email from someone
complaining that it was much harder to find Linux skills than
just to post a job on Monster.com for Windows admins and wait
for resumes to pour in. (In fact, it seemed like the mailer
held me personally responsible for the dearth of Linux skills
and expected me to solve his problem!).
This is a funny perspective. It is nothing to do with open
source at alll. It reflects the organizational challenge that
acquiring new technical skills always presents. The emailer's
bleat about how easy it is to find Windows admins is true
-- today. There was a time when Windows NT was a new product
(and it wasn't that long ago), and finding NT skills
was a challenge, because everyone knew Unix, or VMS, or whatever.
IT organizations have a long history of addressing skill shortages,
and have well-established mechanisms to meet the challenge.
Technical skills are human capital. They require education,
formal or informal, and experience. The capital increases
gradually as people learn new skills as part of their job
or on their own through home study. It's not surprising that
it's more challenging to find less-available skills. Naturally
it requires more creative ways to solve the skill shortage
problem -- just like you have to be more creative to solve
any resource-constrained situation. And really, how hard is
it to find Linux skills? Go to any Linux User's Group. Post
to a Linux forum. Call a recruiter.
As I noted, IT organizations are very experienced in upgrading
their skill sets. However, open source presents them with
a far greater challenge than skill issues, and one they are
not nearly as experienced in solving. I refer to the cultural
and process challenge that open source will force upon IT
organizations. This will be a significant issue in the next
five years.
The reason is very simple. Open source doesn't really present
any insurmountable skill/techical challenges (in fact, some
of those Unix skills from the past can help today!). It does,
however, present an enormous challenge in terms of the playing
field upon which those skills are used. Simply put, open source
changes the rules of the game. The fundamental assumptions
about how IT organizations work will have to be rethought
because of open source. The business model of open source
will play havoc with IT processes.
Today, most IT projects have, at their heart, vendor-centric
assumptions. Projects are all about selecting the right vendor
or upgrading a vendor's products (often forced by a vendor's
mandatory upgrade dictate). IT organizations have the luxury
of expecting lots of help from the vendor in return for the
large licensing or maintenance fees they pay. Another way
to say this is that IT vendors do the work of selling, which
includes the direct selling effort but also all of the marketing,
partner education, and so on.
With open source, vendors are passive. Open source software
is bought, not sold. Users are expected to take care of their
own selection and assessment. Once they've decided on a product,
the vendor (if there is one associated with the product) will
offer support and maintenace.
Of course, this is not unique to open source. It is the reality
of any market moving from high margin to low margin. As margins
shrink, vendors offer fewer bundled services. Users are expected
to do more for themselves.
This change in IT project assumptions will be earthshaking.
I predict an extended period of shock as IT organizations
come to understand the complete implications of open source.
The realities of open source business models will force everyone
to modify their assumptions about how IT is done. Get ready
for it.
See Bernard Speak at LinuxWorld
LinuxWorld next month is going to be really interesting.
I expect many announcements about open source companies getting
VC funding. Attendance should be higher, with a larger percentage
of attendees coming from mainstream companies.
I will be speaking three times at LinuxWorld, twice in the
Addison-Wesley/Prentice Hall booth, and once as part of a
panel on open source trends. Here are the details:
August 10, 12:00: Press Panel
August 10, 1:30: Addison-Wesley Booth: "Enterprise Open
Source Adoption: Trends, Opportunities, and Challenges"
August 11, 10:30: Addison-Wesley Booth: "Open Source
ROI: The Real Story"
Please stop by and see me. I look forward to meeting many
of you in person.
|
