Riddle me this...

[Ginny Palmieri]

Ok - I confess my understanding of how a machine grabs an IP address is shaky at best, but I did think it was sort of constant per log-in, at any rate. So, how come, when I go to whatsmyip through Mozilla I get a different response than I do when I go through IE two seconds later?

Illumination continues to be my goal....

[Russell W. Sprague]

I have a theory, but in order not to make a fool out of myself in front of the university (and, more importantly, my co-workers), can you post the two IP addresses?

[E. Axel Larsson]

Originally Posted by Ginny Palmieri

Ok - I confess my understanding of how a machine grabs an IP address is shaky at best, but I did think it was sort of constant per log-in, at any rate. So, how come, when I go to whatsmyip through Mozilla I get a different response than I do when I go through IE two seconds later?

Illumination continues to be my goal....

What you are seeing when you use one of those sites is the public “real” internet address that your computer is accessing the site through at that moment.

If you run ipconfig from the command prompt, you'll see your computer's current internal IP address. It used to be the case the Drew had internet routable addresses for all of the computers on campus. However, as the IP address space has become more scarce, and when we switched internet providers, we had to start using NAT (Network Address Translation).

With NAT, PCs on the network are assigned internal IP addresses that are not routable on the Internet. We have a pool of "real" Internet IP addresses that are shared amongst all clients on the network and that are managed by our Internet firewall. When your PC tries to connect to a site on the Internet, the firewall dynamically maps your internal address to some address/port combination in this public pool, and maps the response back to your PCs internal IP address. The mapping is dynamic and can change from request to request. NAT allows us to use a much smaller range of real Internet routable addresses compared to the size of our internal IP address space.

This is basically a grown-up version of what happens at home if you use a home networking router with a cable or DSL provider. The provider only assigns your one real IP address as part of your service. The router uses NAT to enable all of the PCs on your home network to share this address.

[Ginny Palmieri]

Originally Posted by Russell Sprague

I have a theory, but in order not to make a fool out of myself in front of the university (and, more importantly, my co-workers), can you post the two IP addresses?

Through IE: 70.18.81.103

Through Mozilla: 192.107.39.18

And your theory is???

Ginny

[Ginny Palmieri]

Thanks, Axel. I'm going to have to read this a couple of times, but I get most of it.

[E. Axel Larsson]

Originally Posted by Ginny Palmieri

Through IE: 70.18.81.103

Through Mozilla: 192.107.39.18

And your theory is???

Ginny

In this case, a slightly different explanation from the one I gave above...

192.107.39.18 is the address of Drew's proxy server... 70.18.81.103 is just one of our public addresses from the NAT pool... So, this means that you have Mozilla configured to use the Proxy, but IE isn't. (When you have the proxy set up and you are on-campus, it directs all traffic through the proxy.. for caching purposes... When you are off-campus, only the research resources and other sites that need to be proxied get proxied...)

[Jennifer A. Heise]

I notice that a bunch of faculty computers in my department, including mine, appear to have the Automatic Proxy server configured to point to:
http://proxy.drew.edu/default.pac

Is this really necessary on campus? What does it do? I'm familiar with proxy servers set up for off campus access to library databases that are restricted by IP address, but this is new to me.

-- Jenne Heise

[Jennifer A. Heise]

Oh, duh. I found the answer to my question on the proxy page:

"The Drew proxy server is also provided for on-campus users. When used on-campus, your web browsing performance will be enhanced because frequently accessed web content will be cached locally on Drew's proxy server. Using the proxy server on-campus also reduces the burden on Drew's internet connection. "

Dumb question, but does that proxy everything, set up that way, or only specified sites? Is there a way to have it only proxy some things? I'm concerned about cached copies of news pages and other things with rotating content.

Originally Posted by Jennifer A. Heise

I notice that a bunch of faculty computers in my department, including mine, appear to have the Automatic Proxy server configured to point to:
http://proxy.drew.edu/default.pac

Is this really necessary on campus? What does it do? I'm familiar with proxy servers set up for off campus access to library databases that are restricted by IP address, but this is new to me.

-- Jenne Heise

[Ginny Palmieri]

Thanks, Axel. That's a big DUH from me. I should have thought of that and looked to see if that's what I'd done. Your earlier explanation, though, in the larger context of things, is really helpful to my understanding. Thanks.

[E. Axel Larsson]

Originally Posted by Jennifer A. Heise

Oh, duh. I found the answer to my question on the proxy page:

"The Drew proxy server is also provided for on-campus users. When used on-campus, your web browsing performance will be enhanced because frequently accessed web content will be cached locally on Drew's proxy server. Using the proxy server on-campus also reduces the burden on Drew's internet connection. "

Dumb question, but does that proxy everything, set up that way, or only specified sites? Is there a way to have it only proxy some things? I'm concerned about cached copies of news pages and other things with rotating content.

When you use the PAC file on-campus, it configures the browser to send most requests out through the proxy.. Only the on-campus application servers bypass the proxy in that configuration.

News sites and the like shouldn't be an issue because the proxy respects the Expires and Cache-Control headers returned by the web servers. A site displaying dynamic content should be setting "Pragma: no-cache" (for HTTP 1.0 clients) and the "Cache-Control: no-store, no-cache, must-revalidate" (for HTTP 1.1 clients) headers.