Research analysts John Rymer and Noel Yuhanna at GigaWorld 2004. This
session was attended by fewer than 30 of the approximately 800 GigaWorld
attendees, but raised some interesting points, especially about how to
competing commercial and open source offerings.Rymer and Yuhanna came across not so much as open source partisans as people who
accept that fact that open source is now a viable part of the enterprise
software ecosystem and is destined to become a larger part of it over time.
Rymer said that when it comes to corporate use of open source, "For the
vast majority, the question should no longer be 'whether' but 'where and
when.' We often find ourselves pushing open source. Yes, it's new, but you should be
Yuhanna pointed to results of a Forrester survey of 140 companies (some
large, some small) taken in March. Some highlights:
- 16% are using Linux now; 45% plan to use it in the near future.
- 8% currently use Apache; 38% plan to use it in the near future.
- 6% use OpenOffice.org; 9% plan to use it in the near future.
- 4% use MySQL now; 27% plan to use it in the near future.
But the most telling figure, in Rymer's and Yuhanna's eyes, was that 52%
surveyed companies were neither using nor planning to use any open source
"Advocates of open source tend to be very biased. They focus on those
using it but don't notice that over half aren't in the game yet," Rymer said.
He was not saying this to knock open source advocates, but to point out
that there is still a large, untapped market for products and services
based on open source, hence plenty of room for growth.
Yuhanna pointed out that mixed "conventional" (as they kept calling closed
source) and open source solutions are becoming more and more common. As an
example, he cited the fact that 55% of MySQL downloads are for Windows, and
that a majority of MySQL-based solutions are apparently
workstations running Windows but end up being deployed on Linux
The ingredients of a successful open source project
Rymer and Yuhanna displayed this equation:
+ Mature standards
+ Large skills base
+ High conventional software prices
+ Right target
+ Quality code
MySQL was their prime example. They pointed out that the SQL standard was
well-known and widespread, and that many developers and users were
familiar with it before
MySQL came along. Commercial (or their word, "conventional") database
were charging stiff license fees. People who needed a fast, simple
primarily "read-only" applications like Web site backends were a perfect
target for an open source database. And MySQL's code was good enough to
this kind of application even in its early releases. Since all these
ingredients were in place, MySQL quickly accumulated a community of
developers (often the same people) who helped the project develop and grow
at a rapid rate.
Rymer noted that JBoss and Tomcat have the same "ingredient mix," with
J2EE as the "conventional" software product against which they both
Comparing commercial and open source competitors
While it may not always be possible to make a direct comparison between
competing "conventional" and open source software products, said Rymer and
Yuhanna, there are ways you can highlight their comparative strengths and
weaknesses if you look at them from the proper angles:
- An open source project's "to do" list should be considered as
same thing as a "conventional" software company's "product road map."
- The "commercial ecosystem" (i.e. third-party support providers)
an open source project can be compared to the "partner ecosystem"
a "conventional" software product.
- When looking at an open source project, see if there
are commercial support alternatives. When looking at
"conventional" software, where support from the product's vendor is
assumed, evaluate the quality of that support.
- When evaluating long-run product viability, compare an open source
community with a "conventional" software product's customer base.
Predictions about open source adoption and "conventional" software
The easiest word to use when describing what Rymer and Yuhanna expect to
happen with corporate open source use betwen now and 2007 is "more."
They expect to see more Apache, more Linux, more open source development tools, more
OpenOffice.org, and, they say, large increases in several areas where open
source solutions are not widely used today, most notably integration
A strong shift they're seeing on the vendor side of the open source
equation is toward selling and supporting complete open source-based "stack"
solutions. They point out that HP and
Novell are moving rapidly in this direction, while IBM seems to be lagging
because much of IBM's current "open source" effort seems to be oriented
IBM's "conventional" software and middleware solutions (i.e. Lotus Notes and
commercial Websphere components) on top of Linux.
(While apparently beneath the notice of Forrester analysts, many smaller
service companies already make good livings selling complete open
hardware + software + service "stack" solutions; this is a proven
business model for free and open source software vendors that is only
adopted by some of the field's bigger players.)
And with more and better (and better-integrated) open source solutions
to market, say Rymer and Yuhanna, "conventional" software vendors are being
forced to lower their prices considerably -- or at least to offer larger and
more frequent discounts if they are reluctant to reduce their "list" prices.
Within 18 to 24 months they expect to see significantly lower prices for
"conventional" database and Java products, with this effect spreading to
parts of the enterprise software market as the 52% of companies that aren't
currently using or considering open source gradually change their minds.