My server 's currently having 2GB of RAM, our documents are mainly scanned PDF and video files , the average size is ~ 200MB for each download. I'm facing problem with memory overloading, because the real FREE memory to serve is about 1GB (the Tomcat and JVM stack uses about 1GB already).
I've heard of Node.js and its concurrent model to serve request. So i'm thinking of using a sepate Node.js server (different port behind a proxy), so each time a user click on download button in Dspace, Dspace will redirect it to the node.js server with nessary parameter linked to the file path, and node.js will use its model to serve file downloading.
I don't know if this kind of setup will improve performance for my need or not. Do you have any suggestion for this use case ?
Thank you very much.
TRUONG HOANG DUNG
Information and Library Centre
Hai Phong Private University