Hi, I think there are three potential risks, none of them particularly major. 1) Information leakage If an attacker issues non-recursive queries against the server, they can see what domains have been looked up, and also infer when from the TTL. In fact even if you disable recursion they may still be able to do this. For BIND I advocate denying all queries at the top of the config file, and then specifically allowing queries for each authorative zone. 2) Unauthorized use of resources People can hijack you as a free resolver. This is unlikely to be much of a problem in reality. Personally I always know a few open resolvers, so I've got backup nameservers if the local ones are down. 3) Potential cache poisoning This is more a theoretical risk than anything. All the easy cache poisoning attacks (ID prediction, birthday, etc.) are vulnerabilities in specific versions of software. Better servers like DJB or BIND 9 randomize both the transaction ID and the UDP source port, giving an attacker about 2^26 space to brute force (and no handy shortcuts). This is scarcely easier than brute forcing a TCP connection. Regards, Paul omifix omnifix wrote: >Hi all > >can anybody explain me what the problem is when my >external DNS server supports recursive DNS queries? > >People are telling me that a DNS server is prone to >cache poisoning when recursive DNS queries are >supported. > > -- Paul Johnston Internet Security Specialist Westpoint Limited Albion Wharf, 19 Albion Street, Manchester, M1 5LN England Tel: +44 (0)161 237 1028 Fax: +44 (0)161 237 1031 email: paul@private web: www.westpoint.ltd.uk _______________________________________________ Full-Disclosure - We believe in it. Charter: http://lists.netsys.com/full-disclosure-charter.html _______________________________________________ Plugins-writers mailing list Plugins-writers@private http://mail.nessus.org/mailman/listinfo/plugins-writers
This archive was generated by hypermail 2b30 : Thu Mar 04 2004 - 10:18:41 PST