Lai Liu-yuan wrote: > > Well, this may be off topic. > > In my case, I store tens of thousands of images, gradually growing. All of them are > quite small, most around 30*30 gray scale. Would it still be faster to store them on > disk? > In most cases yes. We have over 3.5 million images stored and mananaging them from the database side would be almost impossible. There are many benifits to using the file system instead of the database.
1. You can archive the images on cd/dvd/tape as new ones are received/added. 2. Try converting or moving databases with 200GB of blob files. 3. The larger the database, the longer it takes to retreive information. 4. Data from a database is retreived one row at a time so you have to wait for each image. Storing the path in the database allows you to fetch a row, spawn a child process to fetch the image, and continue to fetch rows from the database while the child processes handle getting the images. 5. Mysql will not cache the images. The OS however will cache disk reads. 6. The database has a finite set of resources. You can add many file servers, each specific to what "type" of image you're pulling. We store stuff by date so we know to pull 1998 images from fileserver a and 1999 images from fileserver b, etc... The OS on each fileserver can then cache what is requested most often. Good luck! walt --------------------------------------------------------------------- Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail <[EMAIL PROTECTED]> To unsubscribe, e-mail <[EMAIL PROTECTED]> Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php