> I've read about a way to secure webservers, which must not be directly > exposed to the Internet, using a reverse proxy, e.g. MS ISA Server or > Squid on a UNIX box. >
I have yet to use ISA Server for this, but using Squid is an *excellent* method for keeping the actual HTTP server isolated from public access. > Now my question would be: Has anyone experience with that? Is it really > more secure (compared to firewalling and port forwarding)? Is the MS ISA > Server Webpublishing rule (which implies reverse caching) doing an > application layer filtering or does it just the mentioned caching? Can a > Squid reverse proxy solution fulfill that too? > Again, not sure on using ISA, but using Squid has been an very nice option. Basically in a layout with a lot of Windows boxes running IIS as the actual content/web servers, with private ip connections to a public Squid server has been an excellent solution. Of course, this should not be your only means of securing the network, but since your web services are probably public, this extra step should help you out tremendously. > If not, what are the steps necessary to accomplish this? > I don't know the exact meaning of this question...maybe worded wrong. Here is a link to a good SANS article that covers the general scope and usage of reverse proxy implementation, and covers a brief overview of how to implement with Apache...really adjusting the process here is not that difficult to apply to IIS as the backend web server. http://www.sans.org/rr/web/reverse_proxy.php If you need more detailed information or advice, I can be contacted directly and we can go over a setup more specific to your environment. Personally I prefer Squid since this is what I have used for this exact type of scenario and the cost comparison to licensing an ISA server...well Squid is free so you can't beat that. Adam McCarthy [EMAIL PROTECTED] > Your input is appreciated.