1. Introduction
As specified in RFC 1035std13 (section 6.4), the IQUERY operation for DNS queries is used to look up the name(s) which are associated with the given value. The value being sought is provided in the query's answer section and the response fills in the question section with one or more 3-tuples of type, name and class. As noted in [RFC1035], section 6.4.3, inverse query processing can put quite an arduous burden on a server. A server would need to perform either an exhaustive search of its database or maintain a separate database that is keyed by the values of the primary database. Both of these approaches could strain system resource use, particularly for servers that are authoritative for millions of names. Response packets from these megaservers could be exceptionally large, and easily run into megabyte sizes. For example, using IQUERY to find every domain that is delegated to one of the nameservers of a large ISP could return tens of thousands of 3-tuples in the question section. This could easily be used to launch denial of service attacks. Operators of servers that do support IQUERY in some form (such as very old BIND 4 servers) generally opt to disable it. This is largely due to bugs in insufficiently-exercised code, or concerns about exposure of large blocks of names in their zones by probes such as inverse MX queries. IQUERY is also somewhat inherently crippled by being unable to tell a requester where it needs to go to get the information that was requested. The answer is very specific to the single server that was queried. This is sometimes a handy diagnostic tool, but apparently not enough so that server operators like to enable it, or request implementation where it is lacking. No known clients use IQUERY to provide any meaningful service. The only common reverse mapping support on the Internet, mapping address records to names, is provided through the use of pointer (PTR) records in the in-addr.arpa tree and has served the community well for many years. Based on all of these factors, this document recommends that the IQUERY operation for DNS servers be officially obsoleted.
