RE: [squid-users] Verisign puts wildcard in .com and .net TLDs

On Tue, 16 Sep 2003, Sam Pointer wrote:

> More on-topic, how will Squid handle this? Say I visit `Verisignsucks.com`
> and get their sitesearch page. Squid caches it. Now I go to
> `zzzxfeuoiwqueoiwqe.com` and get the exact same content. Does Squid cache
> another whole instance or recognise that it already has the correct files,

Squid caches on the URL. The URL is different as the hostname component is
different.

But how much this accounts to in cache use depends on verisign. I have not
looked at their sitesearch page but if they are smart they refer to all
images etc via their official domain name and have the HTML content
uncacheable (their goal is presumeably to collect detailed information
about use of unregistered domains)

> albeit from the wrong domain? I'll wager that it stores another instance.
> Therefore, every time a user does a typo on a URL we'll have another load of
> redundant stuff hitting our caches rather than the normal sane behaviour of
> a failure in DNS resolution. Great.

Indeed.

But Verisign is not the first in this madness.. a friend of mine pointed
out that NuDomains already have a similar construct for the .nu domain.
Not as big problem as very few non-web systems use the .nu domain.

Regards
Henrik
Received on Tue Sep 16 2003 - 11:18:15 MDT