Spammers' Upend DNS
Saint Aardvark writes "eWeek reports on the latest trick of spammers: getting around DNS-based lookups. By registering a domain *after* the spam goes out advertising it, they can get around blacklists. However, that causes all sorts of problems for ISPs and anti-spam services. Paul Judge, CTO at Ciphertrust, says "Even in large enterprises, it's becoming very common to see a large spam load cripple the DNS infrastructure.""
You've misunderstood the problem ...
The domains sending the email exist, but the ones advertised in the email do not. Because SpamCop (et. al) punish not only the sending IP block, but also the advertised host/IP block, spammers are advertising sites that won't exist for a few hours, tricking SpamCop (et al) into reporting on domains that don't exist and therefore cannot be penalized.
2advanced.net - Business Quality Hosting
When a DNS query goes to an ISPs DNS server, and the entry does not exist, does it go to the root servers?
When we make a DNS query, it goes to our name server. If the name server does not have a result for that query cached, it queries a higher-level server for information on which name server is authoritative for that domain. It is possible that any DNS query where no component of the domain name is cached to require a query of the root name servers. This is true for any existant or nonexistant domain name.
Secondly, do invalid domain names get cached (I'm thinking not)?
I don't know about all implementations, but contemporary versions of BIND all perform "negative caching" for some amount of time. The invalidity of DNS records can be cached.
Failed requests (non existent domains) always go to the root servers.
2advanced.net - Business Quality Hosting
Sadly it is quite easy:
1) profit
cost to spam: (tiny fixed amount) + ($0.00 * number_of_spam)
means that there is very little cost to the sender of spam. This means that if 1 in 100000 buy their product they have made money.
(Remember this: There Aint No Such Thing As A Free Lunch. We ate our `free' lunches and now we have to pay.)
The article is just wrong, and there's a feedback post on the same page that explains why very well. (Although, what's with the stupid formatting?)
That's the one that by default, sends spam bounces to forged email addresses?
One line blog. I hear that they're called Twitters now.
OpenBSD's spamd will initially reject all mail from previously unknown sources. It will only permit access to sendmail after an attempt at redelivery. This has brought my spam load down to about zero.
Unless a spammer using the above trick attempted redelivery (which is unlikely), it would not cause a DNS flood.
spamd is only one of a great many reasons to consider OpenBSD on your critical servers.
Not according to the article:
The sending domains *don't* exist.
Honestly, this seems pretty overrated - any mail coming into our domain gets a single lookup - if the domain doesn't exist, it gets a 500. If the domain exists, but the DNS servers time out, it gets a 450.
Why anyone would accept mail from a domain that doesn't exist is beyond me.