Details of Major Internet Flaw Posted by Accident

July 22, 2008 – 5:53 AM

The bug has to do with the way DNS clients and servers obtain information from other DNS servers on the Internet. When the DNS software does not know the numerical IP (Internet Protocol) address of a computer, it asks another DNS server for this information. With cache poisoning, the attacker tricks the DNS software into believing that legitimate domains, such as, map to malicious IP addresses.

In Kaminsky’s attack a cache poisoning attempt also includes what is known as “Additional Resource Record” data. By adding this data, the attack becomes much more powerful, security experts say. “The combination of them is pretty bad,” Liu said.

An attacker could launch such an attack against an Internet service provider’s domain name servers and then redirect them to malicious servers. By poisoning the domain name record for, for example, the attackers could redirect the ISP’s users to a malicious phishing server every time they tried to visit the banking site with their Web browser.

Kaminsky declined to confirm that Flake had discovered his issue, but in a posting to his Web site Monday he wrote “13>0,” apparently a comment that the 13 days administrators have had to patch his flaw before its public disclosure is better than nothing.

“Patch. Today. Now. Yes, stay late,” he wrote.


You must be logged in to post a comment.