draft-ietf-dnsop-respsize-07.txt   draft-ietf-dnsop-respsize-08.txt 
DNSOP Working Group Paul Vixie, ISC DNSOP Working Group Paul Vixie, ISC
INTERNET-DRAFT Akira Kato, WIDE INTERNET-DRAFT Akira Kato, WIDE
<draft-ietf-dnsop-respsize-07.txt> February 2007 <draft-ietf-dnsop-respsize-08.txt> November 19, 2007
DNS Referral Response Size Issues DNS Referral Response Size Issues
Status of this Memo Status of this Memo
By submitting this Internet-Draft, each author represents that any By submitting this Internet-Draft, each author represents that any
applicable patent or other IPR claims of which he or she is aware applicable patent or other IPR claims of which he or she is aware
have been or will be disclosed, and any of which he or she becomes have been or will be disclosed, and any of which he or she becomes
aware will be disclosed, in accordance with Section 6 of BCP 79. aware will be disclosed, in accordance with Section 6 of BCP 79.
Internet-Drafts are working documents of the Internet Engineering Internet-Drafts are working documents of the Internet Engineering
skipping to change at page 2, line 5 skipping to change at page 2, line 5
Abstract Abstract
With a mandated default minimum maximum UDP message size of 512 With a mandated default minimum maximum UDP message size of 512
octets, the DNS protocol presents some special problems for zones octets, the DNS protocol presents some special problems for zones
wishing to expose a moderate or high number of authority servers (NS wishing to expose a moderate or high number of authority servers (NS
RRs). This document explains the operational issues caused by, or RRs). This document explains the operational issues caused by, or
related to this response size limit, and suggests ways to optimize related to this response size limit, and suggests ways to optimize
the use of this limited space. Guidance is offered to DNS server the use of this limited space. Guidance is offered to DNS server
implementors and to DNS zone operators. implementors and to DNS zone operators.
INTERNET-DRAFT February 2007 RESPSIZE INTERNET-DRAFT November 19, 2007 RESPSIZE
1 - Introduction and Overview 1 - Introduction and Overview
1.1. The DNS standard (see [RFC1035 4.2.1]) limits message size to 512 1.1. The DNS standard (see [RFC1035 4.2.1]) limits message size to 512
octets. Even though this limitation was due to the required minimum IP octets. Even though this limitation was due to the required minimum IP
reassembly limit for IPv4, it became a hard DNS protocol limit and is reassembly limit for IPv4, it became a hard DNS protocol limit and is
not implicitly relaxed by changes in transport, for example to IPv6. not implicitly relaxed by changes in a network layer protocol, for
example to IPv6.
1.2. The EDNS protocol extension starting with version 0 permits larger 1.2. The EDNS protocol extension starting with version 0 permits larger
responses by mutual agreement of the requester and responder (see responses by mutual agreement of the requester and responder (see
[RFC2671 2.3, 4.5]). The 512 octet message size limit will remain in [RFC2671 2.3, 4.5]), and it is recommended to support EDNS. The 512
practical effect until there is widespread deployment of EDNS in DNS octets message size limit will remain in practical effect until
resolvers on the Internet. virtually all DNS servers and resolvers support EDNS.
1.3. Since DNS responses include a copy of the request, the space 1.3. Since DNS responses include a copy of the request, the space
available for response data is somewhat less than the full 512 octets. available for response data is somewhat less than the full 512 octets.
Negative responses are quite small, but for positive and delegation Negative responses are quite small, but for positive and referral
responses, every octet must be carefully and sparingly allocated. This responses, every octet must be carefully and sparingly allocated. While
document specifically addresses delegation response sizes. the response size of positive responses is also a concern (see
[RFC3226]), this document specifically addresses referral response size.
2 - Delegation Details 2 - Delegation Details
2.1. RELEVANT PROTOCOL ELEMENTS 2.1. RELEVANT PROTOCOL ELEMENTS
2.1.1. A delegation response will include the following elements: 2.1.1. A delegation response will include the following elements:
Header Section: fixed length (12 octets) Header Section: fixed length (12 octets)
Question Section: original query (name, class, type) Question Section: original query (name, class, type)
Answer Section: empty, or a CNAME/DNAME chain Answer Section: empty, or a CNAME/DNAME chain
Authority Section: NS RRset (nameserver names) Authority Section: NS RRset (nameserver names)
Additional Section: A and AAAA RRsets (nameserver addresses) Additional Section: A and AAAA RRsets (nameserver addresses)
2.1.2. If the total response size exceeds 512 octets, and if the data 2.1.2. If the total response size exceeds 512 octets or advertised size
that does not fit was "required", then the TC bit will be set in EDNS, and if the data that does not fit was "required", then the TC
(indicating truncation). This will usually cause the requester to retry bit will be set (indicating truncation). This will usually cause the
using TCP, depending on what information was desired and what requester to retry using TCP, depending on what information was desired
information was omitted. For example, truncation in the authority and what information was omitted. For example, truncation in the
section is of no interest to a stub resolver who only plans to consume authority section is of no interest to a stub resolver who only plans to
the answer section. If a retry using TCP is needed, the total cost of consume the answer section. If a retry using TCP is needed, the total
the transaction is much higher. See [RFC1123 6.1.3.2] for details on cost of the transaction is much higher. See [RFC1123 6.1.3.2] for
the requirement that UDP be attempted before falling back to TCP. details on the requirement that UDP be attempted before falling back to
TCP.
INTERNET-DRAFT November 19, 2007 RESPSIZE
2.1.3. RRsets are never sent partially unless the TC bit is set to 2.1.3. RRsets are never sent partially unless the TC bit is set to
indicate truncation. When TC bit is set, the final apparent RRset in indicate truncation. When TC bit is set, the final apparent RRset in
the final non-empty section must be considered "possibly damaged" (see the final non-empty section must be considered "possibly damaged" (see
[RFC1035 6.2], [RFC2181 9]). [RFC1035 6.2], [RFC2181 9]).
INTERNET-DRAFT February 2007 RESPSIZE
2.1.4. With or without truncation, the glue present in the additional 2.1.4. With or without truncation, the glue present in the additional
data section should be considered "possibly incomplete", and requesters data section should be considered "possibly incomplete", and requesters
should be prepared to re-query for any damaged or missing RRsets. Note should be prepared to re-query for any damaged or missing RRsets. Note
that truncation of the additional data section might not be signalled that truncation of the additional data section might not be signaled via
via the TC bit since additional data is often optional (see discussion the TC bit since additional data is often optional (see discussion in
in [RFC4472 B]). [RFC4472 B]).
2.1.5. DNS label compression allows a domain name to be instantiated 2.1.5. DNS label compression allows the component labels of a domain
only once per DNS message, and then referenced with a two-octet name to be instantiated exactly once per DNS message, and then
"pointer" from other locations in that same DNS message (see [RFC1035 referenced with a two-octet "pointer" from other locations in that same
4.1.4]). If all nameserver names in a message share a common parent DNS message (see [RFC1035 4.1.4]). If all nameserver names in a message
(for example, all ending in ".ROOT-SERVERS.NET"), then more space will share a common parent (for example, all ending in ".ROOT-SERVERS.NET"),
be available for incompressable data (such as nameserver addresses). then more space will be available for incompressible data (such as
nameserver addresses).
2.1.6. The query name can be as long as 255 octets of network data. In 2.1.6. The query name can be as long as 255 octets of network data. In
this worst case scenario, the question section will be 259 octets in this worst case scenario, the question section will be 259 octets in
size, which would leave only 240 octets for the authority and additional size, which would leave only 240 octets for the authority and additional
sections (after deducting 12 octets for the fixed length header.) sections (after deducting 12 octets for the fixed length header) in a
referral.
2.2. ADVICE TO ZONE OWNERS 2.2. ADVICE TO ZONE OWNERS
2.2.1. Average and maximum question section sizes can be predicted by 2.2.1. Average and maximum question section sizes can be predicted by
the zone owner, since they will know what names actually exist, and can the zone owner, since they will know what names actually exist, and can
measure which ones are queried for most often. Note that if the zone measure which ones are queried for most often. Note that if the zone
contains any wildcards, it is possible for maximum length queries to contains any wildcards, it is possible for maximum length queries to
require positive responses, but that it is reasonable to expect require positive responses, but that it is reasonable to expect
truncation and TCP retry in that case. For cost and performance truncation and TCP retry in that case. For cost and performance
reasons, the majority of requests should be satisfied without truncation reasons, the majority of requests should be satisfied without truncation
or TCP retry. or TCP retry.
2.2.2. Some queries to non-existing names can be large, but this is not 2.2.2. Some queries to non-existing names can be large, but this is not
a problem because negative responses need not contain any answer, a problem because negative responses need not contain any answer,
authority or additional records. See [RFC2308 2.1] for more information authority or additional records. See [RFC2308 2.1] for more information
about the format of negative responses. about the format of negative responses.
2.2.3. The minimum useful number of name servers is two, for redundancy 2.2.3. The minimum useful number of name servers is two, for redundancy
(see [RFC1034 4.1]). A zone's name servers should be reachable by all (see [RFC1034 4.1]). A zone's name servers should be reachable by all
IP transport protocols (e.g., IPv4 and IPv6) in common use. IP protocols (e.g., IPv4 and IPv6) in common use. As long as the
INTERNET-DRAFT November 19, 2007 RESPSIZE
servers are well managed, the server serving IPv6 might be different
from the server serving IPv4 sharing the same server name. It is
important to ensure that a zone should have servers reachable by every
IP protocol in common use (e.g., IPv4 and IPv6).
2.2.4. The best case is no truncation at all. This is because many 2.2.4. The best case is no truncation at all. This is because many
requesters will retry using TCP immediately, or will automatically re- requesters will retry using TCP immediately, or will automatically re-
query for RRsets that are possibly truncated, without considering query for RRsets that are possibly truncated, without considering
whether the omitted data was actually necessary. whether the omitted data was actually necessary.
INTERNET-DRAFT February 2007 RESPSIZE 2.2.5. Anycasting (see [RFC3258]) is a useful tool for performance and
reliability without increasing the size of referral response.
2.3. ADVICE TO SERVER IMPLEMENTORS 2.2.6. While it is irrelevant to the response size issue, all zones have
to be served in IPv4 as well to avoid name space fragmentation (see
[RFC3901]).
2.3.1. In case of multi-homed name servers, it is advantageous to 2.3. ADVICE TO SERVER IMPLEMENTORS
include an address record from each of several name servers before
including several address records for any one name server. If address
records for more than one transport (for example, A and AAAA) are
available, then it is advantageous to include records of both types
early on, before the message is full.
2.3.2. Each added NS RR for a zone will add 12 fixed octets (name, type, 2.3.1. Each added NS RR for a zone will add 12 fixed octets (name, type,
class, ttl, and rdlen) plus 2 to 255 variable octets (for the NSDNAME). class, ttl, and rdlen) plus 2 to 255 variable octets (for the NSDNAME).
Each A RR will require 16 octets, and each AAAA RR will require 28 Each A RR will require 16 octets, and each AAAA RR will require 28
octets. octets.
2.3.3. While DNS distinguishes between necessary and optional resource 2.3.2. While DNS distinguishes between necessary and optional resource
records, this distinction is according to protocol elements necessary to records, this distinction is according to protocol elements necessary to
signify facts, and takes no official notice of protocol content signify facts, and takes no official notice of protocol content
necessary to ensure correct operation. For example, a nameserver name necessary to ensure correct operation. For example, a nameserver name
that is in or below the zone cut being described by a delegation is that is in or below the zone cut being described by a delegation is
"necessary content," since there is no way to reach that zone unless the "necessary content," since there is no way to reach that zone unless the
parent zone's delegation includes "glue records" describing that name parent zone's delegation includes "glue records" describing that name
server's addresses. server's addresses.
2.3.4. It is also necessary to distinguish between "explicit truncation" 2.3.3. Recall that the TC bit is only set when the required RRset can
where a message could not contain enough records to convey its intended not be included in its entirety (see [RFC2181 9]). Even when some of the
meaning, and so the TC bit has been set, and "silent truncation", where RRsets to be included in the additional section are not fit in the
the message was not large enough to contain some records which were "not response size, TC bit isn't set. These RRsets may be important for a
required", and so the TC bit was not set. referral. Some DNS implementation tries to resolve these missing glue
records separately which will introduce extra queries and extra time to
resolve a given name.
2.3.5. A delegation response should prioritize glue records as follows. 2.3.4. A delegation response should prioritize glue records as follows.
first first
All glue RRsets for one name server whose name is in or below the All glue RRsets for one name server whose name is in or below the
INTERNET-DRAFT November 19, 2007 RESPSIZE
zone being delegated, or which has multiple address RRsets (currently zone being delegated, or which has multiple address RRsets (currently
A and AAAA), or preferably both; A and AAAA), or preferably both;
second second
Alternate between adding all glue RRsets for any name servers whose Alternate between adding all glue RRsets for any name servers whose
names are in or below the zone being delegated, and all glue RRsets names are in or below the zone being delegated, and all glue RRsets
for any name servers who have multiple address RRsets (currently A for any name servers who have multiple address RRsets (currently A
and AAAA); and AAAA);
thence thence
All other glue RRsets, in any order. All other glue RRsets, in any order.
INTERNET-DRAFT February 2007 RESPSIZE
Whenever there are multiple candidates for a position in this priority Whenever there are multiple candidates for a position in this priority
scheme, one should be chosen on a round-robin or fully random basis. scheme, one should be chosen on a round-robin or fully random basis.
The goal of this priority scheme is to offer "necessary" glue first, The goal of this priority scheme is to offer "necessary" glue first,
avoiding silent truncation for this glue if possible. avoiding silent truncation for this glue if possible.
2.3.6. If any "necessary content" is silently truncated, then it is 2.3.5. If any "necessary content" is silently truncated, then it is
advisable that the TC bit be set in order to force a TCP retry, rather advisable that the TC bit be set in order to force a TCP retry, rather
than have the zone be unreachable. Note that a parent server's proper than have the zone be unreachable. Note that a parent server's proper
response to a query for in-child glue or below-child glue is a referral response to a query for in-child glue or below-child glue is a referral
rather than an answer, and that this referral must be able to contain rather than an answer, and that this referral must be able to contain
the in-child or below-child glue, and that in outlying cases, only EDNS the in-child or below-child glue, and that in outlying cases, only EDNS
or TCP will be large enough to contain that data. or TCP will be large enough to contain that data.
3 - Analysis 3 - Analysis
3.1. An instrumented protocol trace of a best case delegation response 3.1. An instrumented protocol trace of a best case delegation response
follows. Note that 13 servers are named, and 13 addresses are given. follows. Note that 13 servers are named, and 13 addresses are given.
This query was artificially designed to exactly reach the 512 octet This query was artificially designed to exactly reach the 512 octets
limit. limit.
INTERNET-DRAFT November 19, 2007 RESPSIZE
;; flags: qr rd; QUERY: 1, ANS: 0, AUTH: 13, ADDIT: 13 ;; flags: qr rd; QUERY: 1, ANS: 0, AUTH: 13, ADDIT: 13
;; QUERY SECTION: ;; QUERY SECTION:
;; [23456789.123456789.123456789.\ ;; [23456789.123456789.123456789.\
123456789.123456789.123456789.com A IN] ;; @80 123456789.123456789.123456789.com A IN] ;; @80
;; AUTHORITY SECTION: ;; AUTHORITY SECTION:
com. 86400 NS E.GTLD-SERVERS.NET. ;; @112 com. 86400 NS E.GTLD-SERVERS.NET. ;; @112
com. 86400 NS F.GTLD-SERVERS.NET. ;; @128 com. 86400 NS F.GTLD-SERVERS.NET. ;; @128
com. 86400 NS G.GTLD-SERVERS.NET. ;; @144 com. 86400 NS G.GTLD-SERVERS.NET. ;; @144
com. 86400 NS H.GTLD-SERVERS.NET. ;; @160 com. 86400 NS H.GTLD-SERVERS.NET. ;; @160
com. 86400 NS I.GTLD-SERVERS.NET. ;; @176 com. 86400 NS I.GTLD-SERVERS.NET. ;; @176
com. 86400 NS J.GTLD-SERVERS.NET. ;; @192 com. 86400 NS J.GTLD-SERVERS.NET. ;; @192
com. 86400 NS K.GTLD-SERVERS.NET. ;; @208 com. 86400 NS K.GTLD-SERVERS.NET. ;; @208
com. 86400 NS L.GTLD-SERVERS.NET. ;; @224 com. 86400 NS L.GTLD-SERVERS.NET. ;; @224
com. 86400 NS M.GTLD-SERVERS.NET. ;; @240 com. 86400 NS M.GTLD-SERVERS.NET. ;; @240
com. 86400 NS A.GTLD-SERVERS.NET. ;; @256 com. 86400 NS A.GTLD-SERVERS.NET. ;; @256
com. 86400 NS B.GTLD-SERVERS.NET. ;; @272 com. 86400 NS B.GTLD-SERVERS.NET. ;; @272
com. 86400 NS C.GTLD-SERVERS.NET. ;; @288 com. 86400 NS C.GTLD-SERVERS.NET. ;; @288
com. 86400 NS D.GTLD-SERVERS.NET. ;; @304 com. 86400 NS D.GTLD-SERVERS.NET. ;; @304
INTERNET-DRAFT February 2007 RESPSIZE
;; ADDITIONAL SECTION: ;; ADDITIONAL SECTION:
A.GTLD-SERVERS.NET. 86400 A 192.5.6.30 ;; @320 A.GTLD-SERVERS.NET. 86400 A 192.5.6.30 ;; @320
B.GTLD-SERVERS.NET. 86400 A 192.33.14.30 ;; @336 B.GTLD-SERVERS.NET. 86400 A 192.33.14.30 ;; @336
C.GTLD-SERVERS.NET. 86400 A 192.26.92.30 ;; @352 C.GTLD-SERVERS.NET. 86400 A 192.26.92.30 ;; @352
D.GTLD-SERVERS.NET. 86400 A 192.31.80.30 ;; @368 D.GTLD-SERVERS.NET. 86400 A 192.31.80.30 ;; @368
E.GTLD-SERVERS.NET. 86400 A 192.12.94.30 ;; @384 E.GTLD-SERVERS.NET. 86400 A 192.12.94.30 ;; @384
F.GTLD-SERVERS.NET. 86400 A 192.35.51.30 ;; @400 F.GTLD-SERVERS.NET. 86400 A 192.35.51.30 ;; @400
G.GTLD-SERVERS.NET. 86400 A 192.42.93.30 ;; @416 G.GTLD-SERVERS.NET. 86400 A 192.42.93.30 ;; @416
H.GTLD-SERVERS.NET. 86400 A 192.54.112.30 ;; @432 H.GTLD-SERVERS.NET. 86400 A 192.54.112.30 ;; @432
skipping to change at page 6, line 27 skipping to change at page 6, line 48
K.GTLD-SERVERS.NET. 86400 A 192.52.178.30 ;; @480 K.GTLD-SERVERS.NET. 86400 A 192.52.178.30 ;; @480
L.GTLD-SERVERS.NET. 86400 A 192.41.162.30 ;; @496 L.GTLD-SERVERS.NET. 86400 A 192.41.162.30 ;; @496
M.GTLD-SERVERS.NET. 86400 A 192.55.83.30 ;; @512 M.GTLD-SERVERS.NET. 86400 A 192.55.83.30 ;; @512
;; MSG SIZE sent: 80 rcvd: 512 ;; MSG SIZE sent: 80 rcvd: 512
3.2. For longer query names, the number of address records supplied will 3.2. For longer query names, the number of address records supplied will
be lower. Furthermore, it is only by using a common parent name (which be lower. Furthermore, it is only by using a common parent name (which
is GTLD-SERVERS.NET in this example) that all 13 addresses are able to is GTLD-SERVERS.NET in this example) that all 13 addresses are able to
fit, due to the use of DNS compression pointers in the last 12 fit, due to the use of DNS compression pointers in the last 12
occurances of the parent domain name. The following output from a occurrences of the parent domain name. The following output from a
response simulator demonstrates these properties. response simulator written in perl [PERL] demonstrates these properties.
INTERNET-DRAFT November 19, 2007 RESPSIZE
% perl respsize.pl a.dns.br b.dns.br c.dns.br d.dns.br % perl respsize.pl a.dns.br b.dns.br c.dns.br d.dns.br
a.dns.br requires 10 bytes a.dns.br requires 10 bytes
b.dns.br requires 4 bytes b.dns.br requires 4 bytes
c.dns.br requires 4 bytes c.dns.br requires 4 bytes
d.dns.br requires 4 bytes d.dns.br requires 4 bytes
# of NS: 4 # of NS: 4
For maximum size query (255 byte): For maximum size query (255 byte):
only A is considered: # of A is 4 (green) only A is considered: # of A is 4 (green)
A and AAAA are considered: # of A+AAAA is 3 (yellow) A and AAAA are considered: # of A+AAAA is 3 (yellow)
preferred-glue A is assumed: # of A is 4, # of AAAA is 3 (yellow) preferred-glue A is assumed: # of A is 4, # of AAAA is 3 (yellow)
For average size query (64 byte): For average size query (64 byte):
only A is considered: # of A is 4 (green) only A is considered: # of A is 4 (green)
A and AAAA are considered: # of A+AAAA is 4 (green) A and AAAA are considered: # of A+AAAA is 4 (green)
preferred-glue A is assumed: # of A is 4, # of AAAA is 4 (green) preferred-glue A is assumed: # of A is 4, # of AAAA is 4 (green)
INTERNET-DRAFT February 2007 RESPSIZE
% perl respsize.pl ns-ext.isc.org ns.psg.com ns.ripe.net ns.eu.int % perl respsize.pl ns-ext.isc.org ns.psg.com ns.ripe.net ns.eu.int
ns-ext.isc.org requires 16 bytes ns-ext.isc.org requires 16 bytes
ns.psg.com requires 12 bytes ns.psg.com requires 12 bytes
ns.ripe.net requires 13 bytes ns.ripe.net requires 13 bytes
ns.eu.int requires 11 bytes ns.eu.int requires 11 bytes
# of NS: 4 # of NS: 4
For maximum size query (255 byte): For maximum size query (255 byte):
only A is considered: # of A is 4 (green) only A is considered: # of A is 4 (green)
A and AAAA are considered: # of A+AAAA is 3 (yellow) A and AAAA are considered: # of A+AAAA is 3 (yellow)
preferred-glue A is assumed: # of A is 4, # of AAAA is 2 (yellow) preferred-glue A is assumed: # of A is 4, # of AAAA is 2 (yellow)
For average size query (64 byte): For average size query (64 byte):
only A is considered: # of A is 4 (green) only A is considered: # of A is 4 (green)
A and AAAA are considered: # of A+AAAA is 4 (green) A and AAAA are considered: # of A+AAAA is 4 (green)
preferred-glue A is assumed: # of A is 4, # of AAAA is 4 (green) preferred-glue A is assumed: # of A is 4, # of AAAA is 4 (green)
(Note: The response simulator program is shown in Section 5.) (Note: The response simulator program is shown in Appendix A.)
Here we use the term "green" if all address records could fit, or Here we use the term "green" if all address records could fit, or
"yellow" if two or more could fit, or "orange" if only one could fit, or "yellow" if two or more could fit, or "orange" if only one could fit, or
"red" if no address record could fit. It's clear that without a common "red" if no address record could fit. It's clear that without a common
parent for nameserver names, much space would be lost. For these parent for nameserver names, much space would be lost. For these
examples we use an average/common name size of 15 octets, befitting our examples we use an average/common name size of 15 octets, befitting our
assumption of GTLD-SERVERS.NET as our common parent name. assumption of GTLD-SERVERS.NET as our common parent name.
We're assuming a medium query name size of 64 since that is the typical We're assuming a medium query name size of 64 since that is the typical
size seen in trace data at the time of this writing. If size seen in trace data at the time of this writing. If
Internationalized Domain Name (IDN) or any other technology which Internationalized Domain Name (IDN) or any other technology which
results in larger query names be deployed significantly in advance of results in larger query names be deployed significantly in advance of
EDNS, then new measurements and new estimates will have to be made. EDNS, then new measurements and new estimates will have to be made.
INTERNET-DRAFT November 19, 2007 RESPSIZE
4 - Conclusions 4 - Conclusions
4.1. The current practice of giving all nameserver names a common parent 4.1. The current practice of giving all nameserver names a common parent
(such as GTLD-SERVERS.NET or ROOT-SERVERS.NET) saves space in DNS (such as GTLD-SERVERS.NET or ROOT-SERVERS.NET) saves space in DNS
responses and allows for more nameservers to be enumerated than would responses and allows for more nameservers to be enumerated than would
otherwise be possible, since the common parent domain name only appears otherwise be possible, since the common parent domain name only appears
once in a DNS message and is referred to via "compression pointers" once in a DNS message and is referred to via "compression pointers"
thereafter. thereafter.
4.2. If all nameserver names for a zone share a common parent, then it 4.2. If all nameserver names for a zone share a common parent, then it
is operationally advisable to make all servers for the zone thus served is operationally advisable to make all servers for the zone thus served
also be authoritative for the zone of that common parent. For example, also be authoritative for the zone of that common parent. For example,
the root name servers (?.ROOT-SERVERS.NET) can answer authoritatively the root name servers (?.ROOT-SERVERS.NET) can answer authoritatively
for the ROOT-SERVERS.NET. This is to ensure that the zone's servers for the ROOT-SERVERS.NET. This is to ensure that the zone's servers
always have the zone's nameservers' glue available when delegating, and always have the zone's nameservers' glue available when delegating, and
INTERNET-DRAFT February 2007 RESPSIZE
will be able to respond with answers rather than referrals if a will be able to respond with answers rather than referrals if a
requester who wants that glue comes back asking for it. In this case requester who wants that glue comes back asking for it. In this case
the name server will likely be a "stealth server" -- authoritative but the name server will likely be a "stealth server" -- authoritative but
unadvertised in the glue zone's NS RRset. See [RFC1996 2] for more unadvertised in the glue zone's NS RRset. See [RFC1996 2] for more
information about stealth servers. information about stealth servers.
4.3. Thirteen (13) is the effective maximum number of nameserver names 4.3. Thirteen (13) is the effective maximum number of nameserver names
usable traditional (non-extended) DNS, assuming a common parent domain usable traditional (non-extended) DNS, assuming a common parent domain
name, and given that implicit referral response truncation is name, and given that implicit referral response truncation is
undesirable in the average case. undesirable in the average case.
4.4. Multi-homing of name servers within a protocol family is 4.4. More than one address records in a protocol family per a server is
inadvisable since the necessary glue RRsets (A or AAAA) are atomically inadvisable since the necessary glue RRsets (A or AAAA) are atomically
indivisible, and will be larger than a single resource record. Larger indivisible, and will be larger than a single resource record. Larger
RRsets are more likely to lead to or encounter truncation. RRsets are more likely to lead to or encounter truncation.
4.5. Multi-homing of name servers across protocol families is less 4.5. More than one address records across protocol families is less
likely to lead to or encounter truncation, partly because multiprotocol likely to lead to or encounter truncation, partly because multiprotocol
clients are more likely to speak EDNS which can use a larger response clients, which are required to handle larger RRsets such as AAAA RRs,
size limit, and partly because the resource records (A and AAAA) are in are more likely to speak EDNS which can use a larger response size
limit, and partly because the resource records (A and AAAA) are in
different RRsets and are therefore divisible from each other. different RRsets and are therefore divisible from each other.
4.6. Name server names which are at or below the zone they serve are 4.6. Name server names which are at or below the zone they serve are
more sensitive to referral response truncation, and glue records for more sensitive to referral response truncation, and glue records for
them should be considered "less optional" than other glue records, in them should be considered "more important" than other glue records, in
the assembly of referral responses. the assembly of referral responses.
4.7. If a zone is served by thirteen (13) name servers having a common 4.7. If a zone is served by thirteen (13) name servers having a common
parent name (such as ?.ROOT-SERVERS.NET) and each such name server has a parent name (such as ?.ROOT-SERVERS.NET) and each such name server has a
INTERNET-DRAFT November 19, 2007 RESPSIZE
single address record in some protocol family (e.g., an A RR), then all single address record in some protocol family (e.g., an A RR), then all
thirteen name servers or any subset thereof could multi-home in a second thirteen name servers or any subset thereof could have address records
protocol family by adding a second address record (e.g., an AAAA RR) in a second protocol family by adding a second address record (e.g., an
without reducing the reachability of the zone thus served. AAAA RR) without reducing the reachability of the zone thus served.
5 - Source Code 5 - Security Considerations
The recommendations contained in this document have no known security
implications.
6 - IANA Considerations
This document does not call for changes or additions to any IANA
registry.
7 - Acknowledgement
The authors thank Peter Koch, Rob Austein, Joe Abley, Mark Andrews,
Kenji Rikitake, Stephane Bortzmeyerand Olafur Gudmundsson for their
valuable comments and suggestions.
This work was supported by the US National Science Foundation (research
grant SCI-0427144) and DNS-OARC.
8 - References
[RFC1034] Mockapetris, P.V., "Domain names - Concepts and Facilities",
RFC1034, November 1987.
[RFC1035] Mockapetris, P.V., "Domain names - Implementation and
Specification", RFC1035, November 1987.
[RFC1123] Braden, R., Ed., "Requirements for Internet Hosts -
Application and Support", RFC1123, October 1989.
[RFC1996] Vixie, P., "A Mechanism for Prompt Notification of Zone
Changes (DNS NOTIFY)", RFC1996, August 1996.
[RFC2181] Elz, R., Bush, R., "Clarifications to the DNS Specification",
RFC2181, July 1997.
[RFC2308] Andrews, M., "Negative Caching of DNS Queries (DNS NCACHE)",
RFC2308, March 1998.
INTERNET-DRAFT November 19, 2007 RESPSIZE
[RFC2671] Vixie, P., "Extension Mechanisms for DNS (EDNS0)", RFC2671,
August 1999.
[RFC3226] Gudmundsson, O., "DNSSEC and IPv6 A6 aware server/resolver
message size requirements", RFC3226, December 2001.
[RFC3258] Hardie, T., "Distributing Authoritative Name Servers via
Shared Unicast Addresses", RFC3258, April 2002.
[RFC3901] Durand, A., Ihren, J., "DNS IPv6 Transport Operational
Guidelines", RFC3901, September 2004.
[RFC4472] Durand, A., Ihren, J., Savola, P., "Operational Consideration
and Issues with IPV6 DNS", RFC4472, April 2006.
[PERL] Wall, L., Christiansen, T., Orwant, J., "Programming Perl",
O'Reilly, ISBN 0-596-00027-8, July 2000.
9 - Authors' Addresses
Paul Vixie
Internet Systems Consortium, Inc.
950 Charter Street
Redwood City, CA 94063
+1 650 423 1301
vixie@isc.org
Akira Kato
University of Tokyo, Information Technology Center
2-11-16 Yayoi Bunkyo
Tokyo 113-8658, JAPAN
+81 3 5841 2750
kato@wide.ad.jp
Appendix A - Source Code
#!/usr/bin/perl #!/usr/bin/perl
# #
# SYNOPSIS # SYNOPSIS
# respsize.pl [ -z zone ] fqdn_ns1 fqdn_ns2 ... # respsize.pl [ -z zone ] fqdn_ns1 fqdn_ns2 ...
# if all queries are assumed to have a same zone suffix, # if all queries are assumed to have a same zone suffix,
# such as "jp" in JP TLD servers, specify it in -z option # such as "jp" in JP TLD servers, specify it in -z option
# #
use strict; use strict;
use Getopt::Std; use Getopt::Std;
INTERNET-DRAFT February 2007 RESPSIZE INTERNET-DRAFT November 19, 2007 RESPSIZE
my ($sz_msg) = (512); my ($sz_msg) = (512);
my ($sz_header, $sz_ptr, $sz_rr_a, $sz_rr_aaaa) = (12, 2, 16, 28); my ($sz_header, $sz_ptr, $sz_rr_a, $sz_rr_aaaa) = (12, 2, 16, 28);
my ($sz_type, $sz_class, $sz_ttl, $sz_rdlen) = (2, 2, 4, 2); my ($sz_type, $sz_class, $sz_ttl, $sz_rdlen) = (2, 2, 4, 2);
my (%namedb, $name, $nssect, %opts, $optz); my (%namedb, $name, $nssect, %opts, $optz);
my $n_ns = 0; my $n_ns = 0;
getopt('z', %opts); getopt('z', %opts);
if (defined($opts{'z'})) { if (defined($opts{'z'})) {
server_name_len($opts{'z'}); # just register it server_name_len($opts{'z'}); # just register it
skipping to change at page 10, line 4 skipping to change at page 12, line 4
} }
return $len; return $len;
} }
sub arsect { sub arsect {
my ($sz_query, $nssect, $n_ns, $cond) = @_; my ($sz_query, $nssect, $n_ns, $cond) = @_;
my ($space, $n_a, $n_a_aaaa, $n_p_aaaa, $ansect); my ($space, $n_a, $n_a_aaaa, $n_p_aaaa, $ansect);
$ansect = $sz_query + $sz_type + $sz_class; $ansect = $sz_query + $sz_type + $sz_class;
$space = $sz_msg - $sz_header - $ansect - $nssect; $space = $sz_msg - $sz_header - $ansect - $nssect;
$n_a = atmost(int($space / $sz_rr_a), $n_ns); $n_a = atmost(int($space / $sz_rr_a), $n_ns);
INTERNET-DRAFT February 2007 RESPSIZE INTERNET-DRAFT November 19, 2007 RESPSIZE
$n_a_aaaa = atmost(int($space $n_a_aaaa = atmost(int($space
/ ($sz_rr_a + $sz_rr_aaaa)), $n_ns); / ($sz_rr_a + $sz_rr_aaaa)), $n_ns);
$n_p_aaaa = atmost(int(($space - $sz_rr_a * $n_ns) $n_p_aaaa = atmost(int(($space - $sz_rr_a * $n_ns)
/ $sz_rr_aaaa), $n_ns); / $sz_rr_aaaa), $n_ns);
printf "For %s size query (%d byte):\n", $cond, $sz_query; printf "For %s size query (%d byte):\n", $cond, $sz_query;
printf " only A is considered: "; printf " only A is considered: ";
printf "# of A is %d (%s)\n", $n_a, &judge($n_a, $n_ns); printf "# of A is %d (%s)\n", $n_a, &judge($n_a, $n_ns);
printf " A and AAAA are considered: "; printf " A and AAAA are considered: ";
printf "# of A+AAAA is %d (%s)\n", printf "# of A+AAAA is %d (%s)\n",
skipping to change at page 10, line 36 skipping to change at page 12, line 36
return "red"; return "red";
} }
sub atmost { sub atmost {
my ($a, $b) = @_; my ($a, $b) = @_;
return 0 if ($a < 0); return 0 if ($a < 0);
return $b if ($a > $b); return $b if ($a > $b);
return $a; return $a;
} }
6 - Security Considerations
The recommendations contained in this document have no known security
implications.
7 - IANA Considerations
This document does not call for changes or additions to any IANA
registry.
8 - Acknowledgement
The authors thank Peter Koch, Rob Austein, Joe Abley, and Mark Andrews
for their valuable comments and suggestions.
INTERNET-DRAFT February 2007 RESPSIZE
This work was supported by the US National Science Foundation (research
grant SCI-0427144) and DNS-OARC.
9 - References
[RFC1034] Mockapetris, P.V., "Domain names - Concepts and Facilities",
RFC1034, November 1987.
[RFC1035] Mockapetris, P.V., "Domain names - Implementation and
Specification", RFC1035, November 1987.
[RFC1123] Braden, R., Ed., "Requirements for Internet Hosts -
Application and Support", RFC1123, October 1989.
[RFC1996] Vixie, P., "A Mechanism for Prompt Notification of Zone
Changes (DNS NOTIFY)", RFC1996, August 1996.
[RFC2181] Elz, R., Bush, R., "Clarifications to the DNS Specification",
RFC2181, July 1997.
[RFC2308] Andrews, M., "Negative Caching of DNS Queries (DNS NCACHE)",
RFC2308, March 1998.
[RFC2671] Vixie, P., "Extension Mechanisms for DNS (EDNS0)", RFC2671,
August 1999.
[RFC4472] Durand, A., Ihren, J., Savola, P., "Operational Consideration
and Issues with IPV6 DNS", April 2006.
10 - Authors' Addresses
Paul Vixie
Internet Systems Consortium, Inc.
950 Charter Street
Redwood City, CA 94063
+1 650 423 1301
vixie@isc.org
Akira Kato
University of Tokyo, Information Technology Center
2-11-16 Yayoi Bunkyo
Tokyo 113-8658, JAPAN
+81 3 5841 2750
kato@wide.ad.jp
INTERNET-DRAFT February 2007 RESPSIZE
Full Copyright Statement Full Copyright Statement
Copyright (C) The Internet Society (2006). Copyright (C) IETF Trust (2007).
This document is subject to the rights, licenses and restrictions This document is subject to the rights, licenses and restrictions
contained in BCP 78, and except as set forth therein, the authors retain contained in BCP 78, and except as set forth therein, the authors retain
all their rights. all their rights.
This document and the information contained herein are provided on an This document and the information contained herein are provided on an
"AS IS" basis and THE CONTRIBUTOR, THE ORGANIZATION HE/SHE REPRESENTS OR "AS IS" basis and THE CONTRIBUTOR, THE ORGANIZATION HE/SHE REPRESENTS OR
IS SPONSORED BY (IF ANY), THE INTERNET SOCIETY, THE IETF TRUST AND THE IS SPONSORED BY (IF ANY), THE INTERNET SOCIETY, THE IETF TRUST, AND THE
INTERNET ENGINEERING TASK FORCE DISCLAIM ALL WARRANTIES, EXPRESS OR INTERNET ENGINEERING TASK FORCE DISCLAIM ALL WARRANTIES, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE IMPLIED, INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE
INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED
WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.
INTERNET-DRAFT November 19, 2007 RESPSIZE
Intellectual Property Intellectual Property
The IETF takes no position regarding the validity or scope of any The IETF takes no position regarding the validity or scope of any
Intellectual Property Rights or other rights that might be claimed to Intellectual Property Rights or other rights that might be claimed to
pertain to the implementation or use of the technology described in this pertain to the implementation or use of the technology described in this
document or the extent to which any license under such rights might or document or the extent to which any license under such rights might or
might not be available; nor does it represent that it has made any might not be available; nor does it represent that it has made any
independent effort to identify any such rights. Information on the independent effort to identify any such rights. Information on the
procedures with respect to rights in RFC documents can be found in BCP procedures with respect to rights in RFC documents can be found in BCP
78 and BCP 79. 78 and BCP 79.
 End of changes. 42 change blocks. 
134 lines changed or deleted 164 lines changed or added

This html diff was produced by rfcdiff 1.34. The latest version is available from http://tools.ietf.org/tools/rfcdiff/