Up until Matasano mistakenly let the cat out of the bag about the DNS forgery attack that Dan Kaminsky found, lots of experts were downplaying the problem as old and known. Once the details were released, those same folks agreed, that yes, the problem Kaminsky found was that bad. Since Kaminsky gave his presentation about the DNS vulnerability (along with two blog posts explaining Why So Serious and a Summary), a lot of noise is being made about the impacts.
Can I take a moment to say, "well, duh!" Alright, got that out of my system. The short story is just about everything relies on DNS. Web surfing, e-mail, anti-spam, SSL, content delivery, and so on. It's not just that those services rely on DNS, those services can be made party to the attack. Kaminsky has no less than 17 detailed slides on why attacking DNS is fruitful. Kaminsky then dives into why SSL is not a silver bullet. Every network device, connected to the Internet or not, relies on DNS for name resolution.
DNS is more than simple name resolution. Just look at everything else DNS stores. The IETF DNS extensions working group has been adding new resources records such as storing the location of services in the SRV resource record (RFC 2782), the distribution of phone numbers and services (ENUM) in the NAPTR resource record (RFC 3403), or digital certificates and certificate revocation lists in the CERT resource record (RFC 4398). DNS is a distributed data base. Address records are just one type of record.
Kaminsky's slide deck is great stuff. I bet the presentation was useful (now wish I had gone to Vegas), but his slides give enough detail to pick up the points. For more details and a more accessible explanation, check out Steve Friedl's "An Illustrated Guide to the Kaminsky DNS Vulnerability."
