One thing came to my mind. Nowadays the internet connection speed is often comparable to hard drive speeds. It means I can fully imagine there might be cases when it would be better (for a selfish web client) to retrieve a resource from network than to read it from disk. Especially when the size of the resource is relatively small the time cost of disk operations may exceed the download time.
Do web browsers have such a functionality?
The reason for my question is pure curiosity, not a real problem. BTW, I am not convinced such idea is worth implementing as the more bandwidth is taken for some very small local benefits.