draft-ietf-dnsop-bad-dns-res-02.txt

bind 9.3结合mysql数据库

   With glue fetching enabled, however, an authoritative server will
   generate queries whenever it needs to look up an unknown address
   record to complete the additional section of a response.

   We have observed situations where a glue-fetching name server can
   send queries that reach other name servers, but apparently is
   prevented from receiving the responses.  For example, perhaps the
   name server is authoritative-only and therefore its administrators
   expect it to receive only queries.  Perhaps unaware of glue fetching
   and presuming that the name server will generate no queries, its
   administrators place the name server behind a network device that
   prevents it from receiving responses.  If this is the case, all
   glue-fetching queries will go answered.



Larson & Barber         Expires August 16, 2004                 [Page 7]

Internet-Draft    Observed DNS Resolution Misbehavior      February 2004


   We have observed name server implementations that retry excessively
   when glue-fetching queries are unanswered.  A single com/net name
   server has received hundreds of queries per second from a single name
   server.  Judging from the specific queries received and based on
   additional analysis, we believe these queries result from overly
   aggressive glue fetching.

2.4.1 Recommendation

   Implementers whose name servers support glue fetching should take
   care to avoid sending queries at excessive rates.  Implementations
   should support throttling logic to detect when queries are sent but
   no responses are received.

2.5 Aggressive retransmission behind firewalls

   A common occurrence and one of the largest sources of repeated
   queries at the com/net and root name servers appears to result from
   resolvers behind misconfigured firewalls.  In this situation, a
   recursive name server is apparently allowed to send queries through a
   firewall to other name servers, but not receive the responses.  The
   result is more queries than necessary because of retransmission, all
   of which are useless because the responses are never received.  Just
   as with the glue-fetching scenario described in Section 2.4, the
   queries are sometimes sent at excessive rates.  To make matters
   worse, sometimes the responses, sent in reply to legitimate queries,
   trigger an alarm on the originator's intrusion detection system.  We
   are frequently contacted by administrators responding to such alarms
   who believe our name servers are attacking their systems.

   Not only do some resolvers in this situation retransmit queries at an
   excessive rate, but they continue to do so for days or even weeks.
   This scenario could result from an organization with multiple
   recursive name servers, only a subset of whose traffic is improperly
   filtered in this manner.  Stub resolvers in the organization could be
   configured to query multiple name servers.  Consider the case where a
   stub resolver queries a filtered name server first.  This name server
   sends one or more queries whose replies are filtered, so it can't
   respond to the stub resolver, which times out.  The resolver
   retransmits to a name server that is able to provide an answer.
   Since resolution ultimately succeeds the underlying problem might not
   be recognized or corrected.  A popular stub resolver has a very
   aggressive retransmission schedule, including simultaneous queries to
   multiple name servers, which could explain how such a situation could
   persist without being detected.

2.5.1 Recommendation




Larson & Barber         Expires August 16, 2004                 [Page 8]

Internet-Draft    Observed DNS Resolution Misbehavior      February 2004


   The most obvious recommendation is that administrators should take
   care not to place recursive name servers behind a firewall that
   prohibits queries to pass through but not the resulting replies.

   Name servers should take care to avoid sending queries at excessive
   rates.  Implementations should support throttling logic to detect
   when queries are sent but no responses are received.

2.6 Misconfigured NS records

   Sometimes a zone administrator forgets to add the trailing dot on the
   domain names in the RDATA of a zone's NS records.  Consider this
   fragment of the zone file for "example.com":

     $ORIGIN example.com.
     example.com.      3600   IN   NS   ns1.example.com  ; Note missing
     example.com.      3600   IN   NS   ns2.example.com  ; trailing dots

   The zone's authoritative servers will parse the NS RDATA as
   "ns1.example.com.example.com" and "ns2.example.com.example.com" and
   return NS records with this incorrect RDATA in responses, including
   typically the authority section of every response containing records
   from the "example.com" zone.

   Now consider a typical sequence of queries.  A recursive name server
   attempting to resolve A records for "www.example.com" with no cached
   information for this zone will query a "com" authoritative server.
   The "com" server responds with a referral to the "example.com" zone,
   consisting of NS records with valid RDATA and associated glue
   records. (This example assumes that the "example.com" zone
   information is correct in the "com" zone.)  The recursive name server
   caches the NS RRset from the "com" server and follows the referral by
   querying one of the "example.com" authoritative servers.  This server
   responds with the "www.example.com" A record in the answer section
   and, typically, the "example.com" NS records in the authority section
   and, if space in the message remains, glue A records in the
   additional section. According to Section 5.4 of RFC 2181 [4], NS
   records in the authority section of an authoritative answer are more
   trustworthy than NS records from the authority section of a
   non-authoritative answer.  Thus the "example.com" NS RRset just
   received from the "example.com" authoritative server displaces the
   "example.com" NS RRset received moments ago from the "com"
   authoritative server.

   But the "example.com" zone contains the erroneous NS RRset as shown
   in the example above.  Subsequent queries for names in "example.com"
   will cause the server to attempt to use the incorrect NS records and
   so the server will try to resolve the nonexistent names



Larson & Barber         Expires August 16, 2004                 [Page 9]

Internet-Draft    Observed DNS Resolution Misbehavior      February 2004


   "ns1.example.com.example.com" and "ns2.example.com.example.com".  In
   this example, since all of the zone's name servers are named in the
   zone itself (i.e., "ns1.example.com.example.com" and
   "ns2.example.com.example.com" both end in "example.com") and all are
   bogus, the recursive server cannot reach any "example.com" name
   servers.  Therefore attempts to resolve these names result in A
   record queries to the "com' authoritative servers.  Queries for such
   obviously bogus glue A records occur frequently at the com/net name
   servers.

2.6.1 Recommendation

   An authoritative server can detect this situation.  A trailing dot
   missing from an NS record's RDATA always results by definition in a
   name server name that is in the zone.  But any in-zone name server
   should have a corresponding glue A record also in the zone.  An
   authoritative name server should report an error when a zone's NS
   record references an in-zone name server without a corresponding glue
   A record.

2.7 Name server records with zero TTL

   Sometimes a popular com/net subdomain's zone is configured with a TTL
   of zero on the zone's NS records, which prohibits these records from
   being cached and will result in a higher query volume to the zone's
   authoritative servers.  The zone's administrator should understand
   the consequences of such a configuration and provision resources
   accordingly.  A zero TTL on the zone's NS RRset, however, carries
   additional consequences beyond the zone itself: if a recursive name
   server cannot cache a zone's NS records because of a zero TTL, it
   will be forced to query that zone's parent's name servers each time
   it resolves a name in the zone.  The com/net authoritative servers do
   see an increased query load when a popular com/net subdomain's zone
   is configured with a TTL of zero on the zone's NS records.

   A zero TTL on an RRset expected to change frequently is extreme but
   permissible.  A zone's NS RRset is a special case, however, because
   changes to it must be coordinated with the zone's parent.  In most
   zone parent/child relationships we are aware of, there is typically
   some delay involved in effecting changes.  Further, changes to the
   set of a zone's authoritative name servers (and therefore to the
   zone's NS RRset) are typically relatively rare: providing reliable
   authoritative service requires a reasonably stable set of servers.
   Therefore an extremely low or zero TTL on a zone's NS RRset rarely
   makes sense, except in anticipation of an upcoming change.  In this
   case, when the zone's administrator has planned a change and does not
   want recursive name servers throughout the Internet to cache the NS
   RRset for a long period of time, a low TTL is reasonable.



Larson & Barber         Expires August 16, 2004                [Page 10]

Internet-Draft    Observed DNS Resolution Misbehavior      February 2004


2.7.1 Recommendation

   Because of the additional load placed on a zone's parent's
   authoritative servers imposed by a zero TTL on a zone's NS RRset,
   under such circumstances authoritative name servers should issue a
   warning when loading a zone or refuse to load the zone altogether.

2.8 Unnecessary dynamic update messages

   The UPDATE message specified in RFC 2136 [6] allows an authorized
   agent to update a zone's data on an authoritative name server using a
   DNS message sent over the network.  Consider the case of an agent
   desiring to add a particular resource record. Because of zone cuts,
   the agent does not necessarily know the proper zone to which the
   record should be added.  The dynamic update process requires that the
   agent determine the appropriate zone so the UPDATE message can be
   sent to one of the zone's authoritative servers (typically the
   primary master as specified in the zone's SOA MNAME field).

   The appropriate zone to update is the closest enclosing zone, which
   is the lowest zone in the name space.  The closest enclosing zone
   cannot be determined only by inspecting the domain name of the record
   to be updated, since zone cuts can occur anywhere.  One way to
   determine the closest enclosing zone involves working up the name
   space tree and sending repeated UPDATE messages until success.  For
   example, consider an agent attempting to add an A record with the
   name "foo.bar.example.com".  The agent could first attempt to update
   the "foo.bar.example.com" zone.  If the attempt failed, the update
   could be directed to the "bar.example.com" zone, then the
   "example.com" zone, then the "com" zone, and finally the root zone.

   A popular dynamic agent follows this algorithm.  The result is many
   UPDATE messages received by the root name servers, the com/net
   authoritative servers, and presumably other TLD authoritative
   servers. A reasonable question is why the algorithm proceeds with
   sending updates all the way to TLD and root name servers.  In
   enterprise DNS architectures with an "internal root" design, there
   could conceivably be private, non-public TLD or root zones that would
   be the appropriate target for a dynamic update.  However, we question
   if designing an algorithm to accommodate these limited cases is worth
   the load it places on the public DNS in the form of unnecessary
   UPDATE messages.

2.8.1 Recommendation

   Dynamic update agents should not attempt to send UPDATE messages to
   authoritative servers for TLD zones or the root zone by default.  If
   this functionality is supported, it should be require specific action



Larson & Barber         Expires August 16, 2004                [Page 11]

Internet-Draft    Observed DNS Resolution Misbehavior      February 2004


   by a user to be enabled.

2.9 Queries for domain names resembling IP addresses

   The root name servers receive a significant number of A record
   queries where the qname is an IP address.  The source of these
   queries is unknown.  It could be attributed to situations where a
   user believes an application will accept either a domain name or an
   IP address in a given configuration option.  The user enters an IP
   address, but the application assumes any input is a domain name and
   attempts to resolve it, resulting in an A record lookup.  There could
   also be applications that produce such queries in a misguided attempt
   to reverse map IP addresses.

   These queries result in Name Error (RCODE=3) responses.  A recursive
   name server can negatively cache such responses, but each response
   requires a separate cache entry, i.e., a negative cache entry for the
   domain name "192.0.2.1" does not prevent a subsequent query for the
   domain name "192.0.2.2".

2.9.1 Recommendation

   It would be desirable for the root name servers not to have to answer
   these queries: they unnecessarily consume CPU resources and network
   bandwidth.  One possibility is for recursive name server
   implementations to produce the Name Error response directly.  We
   suggest that implementors consider the option of synthesizing Name
   Error responses at the recursive name server.  The server could claim
   authority for synthesized TLD zones corresponding to the first octet
   of every possible IP address, e.g. 1., 2., through 255.  This
   behavior could be configurable in the (probably unlikely) event that
   numeric TLDs are ever put into use.

   Another option is to delegate these numeric TLDs from the root zone
   to a separate set of servers to absorb the traffic.  The "blackhole
   servers" used by the the AS 112 Project [8], which are currently
   delegated the in-addr.arpa zones corresponding to RFC 1918 [7]
   private use address space, would be a possible choice to receive
   these delegations.

2.10 Misdirected recursive queries

   The root name servers receive a significant number of recursive
   queries (i.e., queries with the RD bit set in the header).  Since
   none of the root servers offer recursion, the servers' response in
   such a situation ignores the request for recursion and the response
   probably does not contain the data the querier anticipated.  Some of
   these queries result from users configuring stub resolvers to query a



Larson & Barber         Expires August 16, 2004                [Page 12]

Internet-Draft    Observed DNS Resolution Misbehavior      February 2004


   root server.  (This situation is not hypothetical: we have received
   complaints from users when this configuration does not work as
   hoped.) Of course, users should not direct stub resolvers to use name
   servers that do not offer recursion, but we are not aware of any stub
   resolver implementation that offers any feedback to the user when so
   configured, aside from simply "not working".

2.10.1 Recommendation

   When the IP address of a (supposedly) recursive name server is
   configured in a stub resolver using an interactive user interface,
   the resolver could send a test query to verify that the server
   supports recursion (i.e., the response has the RA bit set in the
   header).  The user could be immediately notified if the server is
   non-recursive.

   The stub resolver could also report an error, either through a user
   interface or in a log file, if the queried server does not support
   recursion.  Error reporting should be throttled to avoid a
   notification or log message for every response from a non-recursive
   server.

2.11 Suboptimal name server selection algorithm

   An entire document could be devoted to the topic of problems with
   different implementations of the recursive resolution algorithm.  The
   entire process of recursion is woefully underspecified, requiring
   each implementor to design an algorithm.  Sometimes implementors make
   poor design choices that could be avoided if a suggested algorithm
   and best practices were documented, but that is a topic for another
   document.

   Some deficiencies cause significant operational impact and are
   therefore worth mentioning here.  One of these is name server
   selection by a recursive name server.  When a recursive name server
   wants to contact one of a zone's authoritative name servers, how does
   it choose from the NS records listed in the zone's NS RRset?  If the
   selection mechanism is suboptimal, queries are not spread evenly
   among a zone's authoritative servers.  The details of the selection
   mechanism are up to the implementor, but we offer some suggestions.

2.11.1 Recommendation

   This list is not conclusive, but reflects the changes that would
   produce the most impact in terms of reducing disproportionate query
   load among a zone's authoritative servers.  I.e., these changes would
   help spread the query load evenly.




Larson & Barber         Expires August 16, 2004                [Page 13]

Internet-Draft    Observed DNS Resolution Misbehavior      February 2004


   o  Do not make assumptions based on NS RRset order: all NS RRs should
      be treated equally.  (In the case of the "com" zone, for example,
      most of the root servers return the NS record for
      "a.gtld-servers.net" first in the authority section of referrals.
      As a result, this server receives disproportionately more traffic
      than the other 12 authoritative servers for "com".)

   o  Use all NS records in an RRset.  (For example, we are aware of
      implementations that hard-coded information for a subset of the
      root servers.)

   o  Maintain state and favor the best-performing of a zone's
      authoritative servers.  A good definition of performance is
      response time. Non-responsive servers can be penalized with an
      extremely high response time.

   o  Do not lock onto the best-performing of a zone's name servers.  A