quote:
Originally posted by ThePreacher:
If a website like yahoo, which probably gets millions of hits a day, can use MySQL, then it is more than powerful enough to handle almost everything and microsoft and IBM might actually have something to worry about.
Don't get me wrong, I think MySQL is an excellent database (actually I like PostgreSQL more which is also Free/GPL). I use both MySQL and PostgreSQL heavily. PostgreSQL is actually more like a DB2/MSSQL/Oracle/Sybase than MySQL is but still fall way short.
However, both MySQL and PostgreSQL fall way short of DB2/Oracle/Sybase for many things large companies use databases for, in many areas. I use PostgreSQL/MySQL for as much as I can and where it makes sense and where they don't fall short or fail (they make great web server databases and are excellent for small databases). But for many applications there is no comparison, they are on two completely different planets. I have done much testing in this area.
Now MySQL and PostgreSQL can actually beat the high end databases in specific areas because they don't have the overhead. For instance, in a specialized application since you have the source code for MySQL you can tune and modify MySQL to give ultimate performance for that specific application. Now most people don't have the expertise to understand what it takes to tune and modify database source code to match an application.
With the larger databases you have tuning and profiling applications to analyze and optimize your queries, etc. And the tool sets are extensive. You *can* get in touch with vendor database engineers directly to help you solve a problem you may be having.
I'll try and give you a good example of this without revealing who I worked for or what the system was for. I was responsible for a revenue system where a 20-30GB Sybase database on a 10 processor Sun Enterprise server. The system consisted of a distributed processing where around 30 workstations automatically chugged through the 30GB of historical data nightly and made forcasts and sales predictions for the coming year and adjusted inventory accordingly and automatically.
Around 30 analysts worked in the database manually during the day and made minor adjustments for special circumstances. The work that the automated part did nightly may have required thousands of analysts to do. Now, occasionally I would run into performance or corruption problems on specific versions of Sybase.
There were several occasions where I needed to escalate my problems to the top Sybase developers/engineers who were able to help me figure out my problem and develop a patch or solution very promptly. For this particular database it was critical that it be up and functioning efficiently nearly 100% of the time. There were very few short windows where I could bring the system down and work on it.
Now we paid around $45,000/yr for support on this one database (which wasn't even the highest level of support we could get) but it was a small price to pay compared to the $1M/week in revenue this system generated for the company.
Out of the box, MySQL can't even begin to scratch the surface of what this database does. Now, if you put the entire MySQL development team on the project along with the vendor of the revenue software, maybe in a year you could come up with a specialized version of MySQL capable of doing this specific application, and maybe even better than you could do it with Sybase. But that's going to cost you far more than what you are paying for the high end system that can already do the job. Do you know how much high end database server developers make, let alone a team of them?
