Open to Discussion

Will open source databases ever have a major role in business-critical applications?

What About Support?

The first-tier support for any open source product is the source code itself, which is accurate, free, and always available. The assumption is, however, that the users possess the skills and the time to read and, if necessary, modify the source code. Mailing lists are another version of free first-tier support, but they make no guarantees on timely answers either. For all open source database products, other than MySQL and PostgreSQL, these support options are the only ones available.

MySQL and PostgreSQL, however, do have vendor support in the form of email, Web sites, and call centers. NuSphere Corp. and Great Bridge LLC are two companies that have been recently publicizing their support options for the two databases, including training classes and materials. Both companies sell packaged versions of the product, similar to the Linux distribution companies, and offer to send out product updates on a specified periodicity. Consulting services are also available.

The current release of the Red Hat Inc. database packages PostgreSQL 7.1.2 with Red Hat Linux 7.1, integrated under a single installation mechanism. The base package includes a month of unlimited phone support and monthly software updates for a year. Additional phone and Web-based support of varying levels can, of course, be purchased for fees comparable to those charged by NuSphere and Great Bridge. The great advantage to the Red Hat Database lies in having both operating system and database software supported by the same vendor, which should create a single point of reference that will simplify the lives of both database and systems administrators.

Using Open Source Databases

Today, commercial databases are superior to open source ones. In order for open source databases to be competitive with products such as Oracle 9i, IBM DB2, and SQL Server, open source databases must improve dramatically in reliability and scalability. Features such as automatic crash recovery, reliable site replication, clustering and parallel server options, and enhanced query optimization strategies are needed before the products can really begin to compare. Furthermore, open source databases currently offer little, if any, additional support for decision-support applications; creating a terabyte-sized data warehouse on any open source database would be impossible to accomplish today.

However, for smaller-scale applications, especially applications that are Web-based with a Java or Perl interface, open source databases like MySQL and PostgreSQL are a viable alternative. They can provide a streamlined database server to support small to midsize transactional systems, replacing databases such as Microsoft Access or FoxPro. Although I was able to find reference cases in which MySQL servers were said to be housing more than 50,000 tables containing several million rows, the vast majority of the user community is focused on more compact systems. From scanning the Web-based support lists, most MySQL and PostgreSQL databases seem to be in the 10 to 50GB range, with no more than 10 to 20 concurrent users at peak levels. Furthermore, many in the user community chose the open source alternative because they simply could not justify the cost of a commercial database and had personnel who were willing to take time to learn and administer a new product. Often, the organizations or groups that adopted the open source technology had no prior investment in database software.

The Steps Ahead

Rate This Article Rating: 1 (best) 2 3 4 5 (worst) Comments: Optional e-mail address:

In the long run, commercial software will continue to get more expensive, and open source software will continue to improve. The catch is that improvements in open source are directly proportional to the software usage, so a large and diligent user and test community today will guarantee, or at the very least help the chances of, an improved product tomorrow. As Eric Raymond writes in his classic open source essay The Cathedral and the Bazaar, "Given enough eyeballs, all bugs are shallow." Fundamentally, the more the software is used, the better it's going to get.

Database software, unlike operating system software, is unique in that the majority of the truly high-end usage exists in enterprise computing; corporate DBAs make up a large percentage of the most knowledgeable and expert users of database software. Encouraging the use of open source database software in enterprise architecture, even in a limited context, will pave the way for a time when open source products will be able to compete head-on with commercial database engines. Seven years ago, Linux was still in its infancy; today, it has the potential to replace virtually every other operating system in the enterprise. This is a powerful message, and its implications cannot be ignored. I believe that open source database software will follow the same trend, but that its progress will be more closely intertwined with corporate acceptance and usage.

In the grand scheme of things, is seven years really such a long time to wait?

Protima Banerjee [protima@ieee.org] is a technology architect for the database technologies practice of Idea Integration, a leading systems integration consultancy. She has more than seven years of experience in designing and building decision-support solutions.

RESOURCES

Great Bridge: www.greatbridge.com

MySQL: www.mysql.org

NuSphere: www.nusphere.com

PostgreSQL: www.postgresql.org

Red Hat: www.redhat.com

For more information about open source licensing, see www.gnu.org and www.bsd.org.

Open source database benchmarks are available online at osdb.sourceforge.net.