3 ms·
I think this is a terrible article. Paul Vixie has done nothing to fix this situation. He certainly does not deserve the fame of the article. Other DNS servers
by st3fan 18y ago
I think this is a terrible article. Paul Vixie has done nothing to fix this situation. He certainly does not deserve the fame of the article.
Other DNS servers like DJB-DNS and PowerDNS have implemented proper port randomization as part of their design a LONG time ago. As a result of that those servers are completely unaffected by this DNS exploit.
Vixie and his Bind crew ignored the whole thing for a long time until it blew up in their face. Now it it just an excuse to roll out the monster that is called DNSSEC of course. Great marketing.
- tlrobinson 18y agoWell, I'm no expert but it sounds like source port randomization is a bandaid, while DNSSEC is the better lasting solution.
- tptacek 18y agoDNSSEC is a debacle of epic proportions. It has taken over 13 years for that one standards effort to solidify to the point where it is today, which is a "secure DNS" protocol where there is still no real agreement on how to prevent arbitrary people on the Internet from dumping the contents of your zone files. I predict, without any real evidence to back me up, that DNSSEC is DOA for a simple reason: the total Internetwide deployment of resolver libraries with a "gethostbyname()" interface; none of these libraries can handle transient or "soft" DNS security failures. SSL, a protocol that is far, far easier to deploy and manage than DNSSEC, sees transient errors so often that users are rebelling against the size of the error messages Firefox generates for them. DNSSEC transient failures kill your lookup. I can give lots of other reasons why DNSSEC isn't going to work, but that's one you might not have thought of.