Am 04.03.2015 um 17:48 schrieb Jean Weisbuch:
In my opinion, the A option is the way to go, it will be simpler to manage, simpler and easier to backup/restore, easier to scale/load balance and i dont see why it would be more of a security issue.
if the folder where you store them is not reachable from the web and you bypass the images via PHP based on authentication it's fine to only store the path, otherwise *it may be* problematic depending how sensitive the files are
Le 04/03/2015 17:12, Peter Boston a écrit :
I don’t know the following questions are valid here or not. I just hope I get some help from experienced experts here.
I am planning to use MariaDB on 64bit Linux for a new medical application project. I have to store lots of images. The size of one image is about 263KB. There are about 50 images per patient. It’s about 14MB.
I have two options here for these image data:
Option A: store them in file system and have their location linked in database table;
Option B: store them in database in a separated table as BOLB with a unique id;
I know there are lots of implementations go with Option A. My concern about Option A is the security and backup. I have to do them separately. Do you have any good idea how to do them?
For Option B, I am worry about the size of that table and the database. I read somewhere saying the BOLB and TEXT columns in a table are not stored “inline” and will not be counted to the table size. Where I can find the detailed information about this?
Oracle has a special way to handle this situation and is pushing save images into database. Is there a similar mechanism in MariaDB?
How about compress a series of images for a patient in Option B? Is it worthy to scarify perform for gaining some spaces?
Eventually the data size will be very big, what are the options for me to cleanup and archive old data, including the image data?