Home/ Questions/Q 7884273
In Process

The Archive Base Latest Questions

Editorial Team
  • 0
Asked: 2026-06-03T04:45:18+00:00

I have an image serving script in PHP that I have been optimizing. I

  • 0

I have an image serving script in PHP that I have been optimizing. I have been working a lot on the server-end of optimization, but I also want to make sure I optimize the mysql side of things. I currently have the following actions being taken with MYSQL on each loading of this image serving script.

  1. Connect to the database
  2. Check to see if the id provided is valid and return the row if it is
  3. Update the stats of the image by increasing the number of hits column by 1
  4. Increment the stats table with the specific controller and action by 1 as well
  5. I don’t close the connection, i was told to let it sit as it takes less time and it will die on it’s own

I ran “mysqladmin status” on my server and found out there were 63.8997 avg being run at the moment. I currently have around 17 req/s going through apache loading this script. I also have eaccelerator configured on the server to keep a compiled version of this script ready to use. My question is are there any ways I can further optimize the number of queries i can process? I am trying to make sure I keep all the #s as low as I can possibly do and moving as fast as I can?

My goal is to have my server handle 4.32 million image serves a day with no issue (that’s 50/req a sec in apache).

Share

Leave an answer

You must login to add an answer.

Need An Account,

1 Answer

  1. Editorial Team

    Everytime you increment the counter and commit it to the DB, you probably will be writing to the disk. Instead, you can try caching the number of hits in memory for a few minutes, and every few minutes dump the cache back into the database. This will reduce the number of writes and reads you have on the database.

    You can also go a step further in caching the entire row in memory. This will only be practical if the data doesn’t change very often, and if the data is requested very often. Also, memory constraints may become an issue if you start caching huge amounts of data.

    I know in some very high performance database environments, the entire database instance will be stored in memory, and writes to the disk will only occur periodically. This would also drastically decrease query times, and reduce the need for caching as everything would already be in memory. This would require some serious hardware though.

    Additionally, you may want to eventually move away from MySQL. MySQL isn’t the best for high performance, namely becausse MySQL and stored procedures don’t go hand in hand very well. Stored procedures are compiled on the database, and don’t require parsing of SQL in order to be executed. If you were to switch to something, like Oracle for example this entire transaction could be done with a compiled stored procedure. MySQL doesn’t easily provide this flexiblity, if at all.

    • 0