No choice you have is ideal. MySQL is in transition, the EnterpriseDB company is untrustworthy, and Oracle “Classic” is expensive.

If I were consulting to your project, I’d try to understand more about your application needs, now and in the future, to try to figure out in what areas you need to pay up (if need be) for top-end features and in which areas cheap/good-enough gets the job done.

http://www.dbms2.com/2009/02/25/even-more-final-version-of-my-tdwi-slide-deck/ focused on analytics rather than OLTP, but if I did an OLTP one, it would have a similar general philosophy.

Best,

CAM

Currently we deploy Oracle on 5 yo Solaris machines with active/passive using Veritas clustering. With additional business needs coming up, we are upgrading our machines (going intel and linux) to have something similar to no-single-point-of-failure architecture (a hw loadbalancer in front, 2 machines for app server, 2 dedicated database machines). This new set of machines were thought of keeping Oracle’s RAC in mind. However, the licensing costs are prohibitive for us.

We are looking out for alternates and are stuck between MySQL and PostgreSQL (or even the thought to bear the cost of Oracle if MySQL and Postgre and not up to the task).

regards,
mnsharif

For instance, in this application and for 1 single operation, we have 2 SELECT / 3 INSERTS / 6 UPDATES.

This “single operation” is repeated tons of time each minute, and we basically need to be able to allow more and more of these in the future (that’s why we were thinking about “mysql cluster”)

We currently have a postgresql web application, using a lot of join queries.
Some times we need to process a lot of INSERTs / UPDATEs, we don’t use SELECT massively.

We are now thinking about rewritting our app, and now that mysql matured, we are considering it for our new version. What interests us the most in “MySql Cluster” which seems to allow an easy / fast way to raise our capacity when expecting use pikes.

At this time, April 2009, would you seriously considere both solutions, with maybe a preference for mysql for the reasons described above ? or would you rather stick to postgresql (though at that time, we don’t use procedures / views … )

Thank you for your answers.

You need to say more to get useful advice.

You won’t find that with PostgreSQL. Nor EnterpriseDB Postgres Plus and Postgres Plus Advanced Server.

This week, a number of mentions about the Marten Mickos of MySQL being interviewed in the Wall Street Journal. He talks about his secret sauce being his teams culture. Seems like this culture is having problems getting a release out the door in a timely fashion and as planned. According to an article in ComputerWorld on April 15, 2008 of this year, (from the article) “Marten Mickos, the former CEO of MySQL who is now a senior vice president at Sun….When we released MySQL 5.0, it didn’t really meet our quality standards, he said today at the start of the MySQL Conference & Expo in Santa Clara, Calif. “With 5.1, we are being much more conservative, much harder on ourselves…..Mickos said it will release the production version by the end of June, or about three months later than planned.”

Here it is July and no signs of 5.1 shipping. This week Bruce Momjian of the PostgreSQL community will be speaking in Boston about the open source development process for PostgreSQL at the Massachusetts Technology Leadership Council Open Source Meeting. Maybe MySQL could learn something from Bruce’s talk as PostgreSQL has had a great history of shipping highly reliable and very high quality product releases on time. Its ready when its ready and when it IS ready, it is made available. I heard from a MySQL sales executive a month ago that the field sales team is very frustrated with the delay in their product shipping.