I'm running a Ironport WSA in transparent mode.
In advanced proxy config - DNS Options I find the following:

Find web server by:
Specify how the appliance should find the location of the requested web server.
0 = use DNS answers in order
1 = use client supplied address then DNS
2 = use ONLY client supplied address
3 = use client supplied address for next hop connection and Web Reputation (Warning: Destination IP based policies will still use DNS).
Default Value: 1  

Can anyone explain why it makes sense to use 1 as the default?
With 1. a user can edit his hostfile and point www.google.com to The request goes to the proxy, which does a dns lookup for www.google.com, but will still forward the request to
Why is the proxy doing a dns lookup, if it doesn't use the answer?

Can anyone see a good reason not to use 0?
It makes sense to me, since the proxy does a DNS request anyway, and this will make the proxy logs more trustworthy.

  • Did any answer help you? If so, you should accept the answer so that the question doesn't keep popping up forever, looking for an answer. Alternatively, you can post and accept your own answer.
    – Ron Maupin
    Jan 5 at 2:22

This could first of all be necessary if for some reason, the client has access to more or different name resolution data than the proxy. You gave one simple example yourself: an entry in the client's host file that is not present on the proxy. Or the client could be talking to a different DNS server.

It could also be an efficiency thing, since with option 0 all DNS requests will be performed twice, once by the client and then again by the proxy. In option 1, the proxy will only perform a DNS lookup as a fallback if the client supplied address doesn't work.

Also note that when a domain uses round robin DNS, the client and proxy might get a different answer to their queries. This would result in a connection to a different server, not the one the client expected to be connected to. This could cause issues with protocols which somehow depend on being connected to the instance returned by DNS. Such a protocol would work with option 1, but not always with option 0.

  • Thanks for the reply. In my environment, the clients and proxy are using the same DNS infrastructure. Even with option 1 it seems to me that the proxies always do a DNS lookup, which puzzles me. Round robin is a good point, but can you think of any web app that would have issues if the client ended up on a different server than expected?
    – sk0yern
    Jun 21 '13 at 14:40
  • I think I remember Windows Update having trouble with transparent proxies for this reason, but I can't seem to find any proof on microsoft.com. It's also not really a web application I guess.
    – Gerben
    Jun 24 '13 at 7:37

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.