NYCPHP Meetup

[Jayesh Sheth]

Greetings fellow LAMPers and LAMPesses,

I came across an interesting article (or rather case study) about how 
MySQL was able enter the formerly impenetrable database market.
It is called (somewhat obtusely) 'MySQL Open Source Database in 2004'. 
It explores how MySQL, a once tiny company gained a foothold in the 
database market by providing an open soure 'commodity' database. Since 
this article is only available in 'protected electronic' or paper 
format, I thought I would share this summary of it with this list, since 
many people on this list use PHP and MySQL together.

There is a lot of focus on how MySQL has 'less features' than other 
database providers (such as Oracle and IBM), yet how it managed to gain 
a following among individuals and small businesses like itself. There 
are the inevitable quotes from Oracle top brass; a typical sampling: " 
Ken Jacobs, Oracle's VP of product strategy, saw MySQL as a complement 
to Oracle's DBMS and believed it would take many years before MySQL 
would be a true competitor. He formulated Oracle's official position: ' 
Technically, in terms of capabilities, MySQL is a decade or more behind 
Oracle. For example, MySQL lacks the most basic relational database 
features ... such as stored procedures, triggers and views. [...] MySQL 
[has not been useful for] the more transaction processing and business 
intelligence applications that constitute the majority of enterprise 
database deployments."

This case study was published on June 1st, 2004. One can however find 
some or all of these 'missing' features in MySQL 5 today. An excerpt 
from the mysql.com site ( 
http://dev.mysql.com/doc/mysql/en/mysql-5-0-nutshell.html ): "New 
development for MySQL is focused on the 5.0 release, featuring stored 
procedures, views (including updatable views), rudimentary triggers, and 
other new features."

This article also presents the other side of the story, with Zach 
Urlocker of MySQL AB explaining that they "compete against 
nonconsumption". In other words, people who would not otherwise have a 
database installation use MySQL. In my experience, I often find MySQL's 
(versions 3.23 and 4.0) features more than sufficient. The only thing 
that I have really missed is enforced foreign key constraints and 
transactional / rollback support, but those features are available if 
one uses the InnoDB engine. The problem is that today, most hosting 
companies only install MySQL with MyISAM engine support, and MyISAM 
lacks these features. Then again, with careful error checking and coding 
on the PHP end, I have been able to work around this issue. Granted, if 
one is building a 500,000 user online banking system, then perhaps MySQL 
4.0 may not be the best database to choose. For for more small business 
Intranet use, or for small business e-commerce and informational 
websites, it is more than sufficient, especially when combined with PHP 
to retrieve and display its data.

The case study also mentions MySQL's alliance with (and investment from) 
SAP in 2003. SAP had this aging, yet full-featured database called SAP 
DB that was used by approximately 5% of its ERP customers. They 
partnered with MySQL so that MySQL would continue maintenance and 
support of this product, now renamed Max DB. MySQL would also continue 
development of its own product line, with a view towards making it 
'enterprise ready' / 'SAP ready' in the next three to four years. This 
alliance seems to have been mutually beneficial: "Both MySQL revenues 
and SAP DB's percentage share in SAP installations increased sharply 
since the start of the alliance in June 2003".

Regarding MySQL's business model (a.ka. "how they make money"), the 
article mentions their dual licensing scheme, where companies that want 
to incorporate MySQL into their products can purchase a commercial 
license and bypass the open source GPL license. The article (Exhibit 2) 
quotes Josh Berkus of the PostgreSQL project as saying that MySQL is 
really "shareware" and not open source. The article mentions that all 
MySQL code has been written by MySQL employees, thus enabling them to 
offer dual licensing. It then draws the conclusion that MySQL is 
"proprietary" and not open source because it owns all the code, unlike 
the case of Linux, where Red Hat, for example, does not own all the code 
itself. The main stated business benefit of MySQL open availability of 
source code is rapid market penetration and widespread testing by the 
thousands of "GPL edition" MySQL users.

There also also mentions of how MySQL is run as a "virtual organization" 
with 120 employees scattered throughout the world, many of whom work 
from their home offices. This, the article says, enables MySQL to keep 
its costs down while recruiting the best and most dedicated developers 
wherever in the world they may be - developers who have already proven 
themselves in the MySQL open source community.

The article ends on a mostly upbeat note, with a discussion of the 
possibilities for MySQL to grow from into a $100 million a year company. 
It says that while large scale success may eventually come to MySQL, 
MySQL must also be wary of the "seismic forces that could easily 
overwhelm" a small company such as MySQL. I suppose those seismic forces 
are Oracle, Microsoft and IBM.

So, as developers, what do you think? Does MySQL meet your needs 
sufficiently? Do you see it growing along with your needs? If you do not 
use MySQL, which database do you use and why? Any thoughts regarding 
IBM's new alliance with Zend to bring DB2 and Derby (Cloudscape) support 
to PHP and how this might affect MySQL's market share among PHP users?

For those interested, more information is available at: 
http://harvardbusinessonline.hbsp.harvard.edu/b01/en/common/item_detail.jhtml?id=SM124

- Jay