How effective are reverse proxies as a web application security device? Which types of threats do they mitigate against and which don't they?
For example are they more effective in preventing exploits against the proxied application web servers rather than against the applications themselves?