One document matched: draft-gudmundsson-dns-srv-iana-registry-04.xml


<?xml version='1.0' ?>
<!DOCTYPE rfc SYSTEM 'rfc2629.dtd'>

<?rfc toc="yes" ?> 
<?rfc symrefs='yes'?>

<?rfc compact='yes' ?>
<rfc ipr='trust200902' category='std'> 

<!-- ugly solution to having the date live in the file that changes with 
     each version -->
<front> 
    <title abbrev="IANA SRV registry"> 
    Creation of a Registry for DNS SRV Record Service Prefixes 
    </title> 
 
    <author initials="O." surname="Gudmundsson" fullname="Olafur Gudmundsson"> 
      <organization> 
        Shinkuro Inc. 
      </organization> 
      <address> 
        <postal> 
	  <street> 4922 Fairmont Avenue, Suite 250 </street> 
	  <city> Bethesda </city> 
	  <region> MD </region> 
	  <code> 20814 </code> 
          <country>            USA          </country> 
        </postal> 
        <email>          ogud@ogud.com        </email> 
      </address> 
      </author> 
 
    <author initials="A." surname="Hoenes" fullname="Alfred Hoenes"> 
      <organization> 
        TR-Sys 
      </organization> 
      <address> 
        <postal> 
          <street>  Gerlinger Str. 12 </street> 
          <city>    Ditzingen </city> 
          <code>    D-71254   </code> 
          <country> Germany   </country> 
        </postal> 
        <email> ah@TR-Sys.de </email> 
      </address> 
    </author> 
    <date month="October" year="2009" />
    <area> 
      Apps 
    </area> 
    <workgroup> 
      Network Working Group 
    </workgroup> 
    <keyword> DNS </keyword> 
    <keyword> SRV </keyword> 
    <keyword> IANA </keyword> 
    <keyword> Service location </keyword> 
    <abstract> 
      <t> 
       The DNS SRV record has been specified in RFC 2052 and RFC 2782. 
       These two RFCs did not specify an IANA registry for names of the 
       services using SRV records as defined by various protocols. 
       This document creates such a registry and populates it. 
      </t> 
    </abstract> 
 

 </front>
<middle> 
  <section title="Introduction"> 
 
    <t> 
      The SRV resource record (RR) was originally introduced in 
      <xref target="RFC2052">RFC 2052</xref> as an experimental extension 
      to the Domain Name System (DNS). 
      It proved a highly valuable addition for service location and 
      provisioning. 
      The SRV record was thus standardized in 
      <xref target="RFC2782">RFC 2782</xref>. 
      The main difference between these two versions is the use of the 
      underscore ("_") prefix character in names to avoid naming conflicts between 
      service names and regular names.   
     </t> 
 
     <t> 
       The presentation form of the SRV resource record (RR type 33), 
       including the recommended naming structure for its use, 
       is as follows <xref target="RFC2782"></xref>: 
     </t> 
 
     <t> 
       <list style="empty"> 
	 <t> 
	   _Service._Proto.Name   SRV  Priority Weight Port Target 
	 </t> 
       </list> 
     </t> 
 
     <t> 
       The "_Service" label indicates the name of the Service/Application that 
       is being offered. The "_Proto" label denotes the name of the transport 
       protocol to be used for the service. 
       "Name" is the domain name that is offering the service. 
       The PORT field is the port number over which the service is provided 
       at the Target name.  The Priority and Weight fields are used for 
       selecting among multiple SRV records at the same owner name. 
     </t> 
 
     <t> 
       RFC 2782 says that the source of names for "Service" and "Proto" is 
       "Assigned Numbers" (STD2) or a locally defined repository. 
     </t> 
 
     <t> 
     <list style="hanging"> 
       <t hangText="Note:"> 
	 The STD2 series of documents was obsoleted by 
	 <xref target="RFC3232">RFC 3232</xref> and IANA registration 
	 publication was handed over to on-line registries maintained by IANA. 
	 Unfortunately it is not explicitly explained in RFC 2782 which 
	 section of STD2 it is referring to, nor does RFC 3232 help. 
	 <!-- 
	     OLD:  By common knowledge, the corresponding IANA registry's 
	     are the two "Port Numbers" registries. 
	     Hmmm. 
	     There always was one combined port number registry for TCP & UDP! 
	     After studying 2782, the following text seems a better match: 
	   --> 
	 By common knowledge, RFC 2782 referred to the Keyword columns of 
	 the "Protocol Numbers" and "Port Numbers" IANA registries. 
       </t> 
     </list> 
     </t> 
 
     <t> 
       However, upon reflection, both alternatives do not seem to make 
       particular sense: 
     </t> 
 
     <t> 
     <list style="symbols"> 
       <t> 
	 As SRV records contain the port where each server provides 
	 the service, the outmost utility of SRV RRs is for services 
	 that do not have a registered port number.  Also, the number 
	 of ports available is small compared to the possible number 
	 of service names that could be registered. Therefore, the 
	 "Port Number" registry needs a more strict registration policy 
	 and is not the proper place for registering Service names for 
	 use with SRV resource records. 
       </t> 
       <t> 
	 Having locally defined lists of Service and/or Proto names 
	 would equally allow to list the full service information 
	 in such local databases and thus make the usage of SRV 
	 records redundant.  In any case, this scenario is not 
	 applicable for publicly available services where potential 
	 clients are not under the control of the authority offering 
	 the services, and hence most probably would have no access 
	 to the proper "locally provided" information. 
	 The reader should be reminded that locally maintained database 
	 solutions generally scale very poorly, and that this once was 
	 the major momentum for the deployment of the Domain Name System. 
       </t> 
     </list> 
     </t> 
     <t> 
       For these reasons, this document creates a separate IANA registry 
       for Services that allow or specify service discovery via SRV look-up. 
     </t> 
 
<!-- DISCUSS: 
     Should the *Protocol* be mentioned here as well ? 
     Preexisting practice was defining specific service/protocol 
     combinations only! 
     It also might be wise consider the new strategy being 
     developed in draft-ietf-tsvwg-iana-ports for registering 
     port+proto combinations. 
--> 
<!-- COMMENT: 
	OG thinks that the "Service" name should be the same and 
	unique for all "Protocol" even though an application only 
	envisions using TCP the registration should block any other 
        registrations for SCTP 
--> 
     <t> 
     <list style="hanging"> 
       <t hangText="Note:"> 
	 A couple of RFCs published in the past pretend to have registered 
	 with IANA particular SRV Service/Protocol combinations, but 
	 at the time of this writing, evidence shows that this did not 
	 happen actually.  Section 4.2 tries to collect this past wisdom 
	 to initiallly populate the new registry. 
       </t> 
     </list> 
     </t> 
     <t> 
       This Service Registry explicitly removes the constraint that 
       services need a port number registration. 
       The registration requirement for this new registry is set low 
       in order to make it relatively easy to register a new service name. 
     </t> 
 
     <t> 
       To a large extent, the requirement to register port numbers can be 
       eliminated by encouraging SRV for service discovery and location 
       in all new application protocols.  For this reason, there ought not 
       be a name conflict between what is registered in the SRV 
       registry defined in this document and the legacy service names 
       in the port numbers registry.  See 
       <xref target='Collision'></xref> for elaborations 
       on such name conflicts. 
     </t> 
 
     <t> 
       In the spirit of BCP 17, <xref target="RFC2219">RFC 2219</xref>, 
       this registry should also help 
       providing an orderly substitute for the poorly specified 
       Well-Known Network Service Alias names. 
     </t> 
 
     <t> 
       Note: <xref target="RFC5507">RFC 5507</xref>, discusses the 
       use of "underscore labels" in the DNS more generally, from the 
       architectural point of view of the IAB. 
     </t> 
     <section title="Terminology"> 
       <t> 
	 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 
	 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 
	 document are to be interpreted as described in <xref 
		   target="RFC2119">RFC 2119</xref> . 
       </t> 
     </section> 
  </section> 
 
  <section title="SRV Service Prefix Registry Considerations"> 
    <t> 
      Registering a new service should be easy and painless process; 
      all that is needed is a pointer to a service description. 
      In some cases, applicants may not want to release the service/protocol 
      specification, and that is fine. 
    </t> 
     <!-- OG: What I wanted to say is a new colum in the table should be hard 
	  to register but that is not realy material here so I shortened 
	  this and got rid of the transport protocol. 
       --> 
     <!-- AH: Hmmm. -- Should be reconsidered! 
	  That does not match preexisting practice of defining 
	  specific service/protocol combinations only! 
       --> 
 
     <section title="To _ or Not To _"> 
       <!-- AH: "transports" replaced by "protocols" below  --> 
       <t> 
	 The original SRV RFC used regular DNS names for service and 
	 transport names.  This was shown to cause some name conflicts. 
	 For example, it is impossible to have the name "TCP" delegation and 
	 also provide a service on http.tcp. For this reason, the current 
	 SRV RFC specifies the use of underscore in front of all names for both 
	 protocols and services. The argument for following this 
	 nomenclature for protocols is clear and needed. On the other 
	 hand, having a full DNS name space created on the left of each 
	 protocol name, the argument for name collision in the 
	 services does not apply.  Arguments that the presence of 
	 the underscore character makes it easier to spot that this is a 
	 service are not convincing. 
	 A possible future extension to allow IDN Service names may only 
	 work if the underscore prefix is not used. 
       </t> 
     </section> 
     <section title="Service Name Maintenance"> 
       <!-- AH: made text more precise w.r.t. characters vs. octets, 
		needed after introducing the idea of I18N and IDN; 
		I expects registrants will start thinking in 
		Unicode characters, and IDN will blow up 
		the number of octets needed significantly --> 
       <t> 
	 As Service names are a DNS label, the strings can be up to 63 
	 octets long.  This is a 
	 large enough space that reuse and reclaiming of names is not 
	 an issue, unlike for the port space. 
       </t> 
       <!-- AH added considerations for other maintenance operations --> 
       <t> 
	 However, applicants (or their provably legitimate successors) 
	 can later request updates to the contact information and/or 
	 references, and anyone can add another protocol for an 
	 existing service, based on additional specification. 
	 Such requests, when sent to IANA, must clearly be marked as 
	 change requests. 
       </t> 
 
     </section> 
 
     <!-- OG: Regular expressions are a bad idea I'm told 
	      so I'm dropping this section IANA hated this idea --> 
 
       <!-- Eliminating this section 
	    <section title="Regular Expression in Service Registrations"> 
	      <t> 
		There are cases were a registrant may want to register a 
		collection of names that all follow a pattern like 
		"my-updates-21yy" or "Producer-model_number". In the first case 
		the registration would be for "My-updates-21[0-9][0-9]" and in 
		the second case "Producer-.+". These registrations should be as 
		specific as possible avoiding cases like "http.+" that collides 
		with https. If a regular expression in a request collides with 
		an existing string, then it should be rejected.  Only 
		registrations with regular expressions at the end of the string 
		are allowed. 
	      </t> 
	    </section> 
	--> 
       <!-- AH: Idea: 
	    Should there be at least one hyphen to the left of the wildcard? 
	    Terminology (unresolved above): 
	    A specific character also is an (elemntary) regular expression. 
	    How about using "wildcard" or "multi-valued term(s)" above ? 
	 --> 
 
	 <section title="Name Collisions in Service Registrations" anchor='Collision' > 
	   <t> 
	     As this document allows registrations of NEW services without 
	     the underscore ("_") prefix, these registrations MUST NOT 
	     collide with pre-existing registrations that start with 
	     or without the underscore prefix. 
	   </t> 
	   <!-- AH: 
		I recently have learned that case-insensitive comparison 
		is a very subtle problem in Unicode.  I have proactively 
		changed the language to avoid phrasing that has been 
		challenged in other WGs. 
	   --> 
	   <t> 
	     A name collision is defined to occur if two 
	     labels compare equal under case-insensitive comparison 
	     after stripping of the underscore prefix (if any) from both. 
 
	     <!-- OG: IANA wanted this removed. 
		  Furthermore, to provide a grace period where existing names can 
		  be registered, for one calendar year after this document is 
		  published as an RFC, no registrations without "_" will be 
		  accepted. 
	       --> 
	     <!-- AH: I guess this remark is misplaced/outdated now.  --> 
	   </t> 
	   <t> 
	     In registering a new name in the SRV registry there MUST 
	     not be a name conflict with the "PORT NUMBERS" registry keyword field. 
	   </t> 
	   <t> 
	     New registraions in the "PORT NUMBERS"  MUST NOT create a 
	     name confict with the SRV registry. 
	   </t> 
	 </section> 
     </section> 
 
     <section title="SRV Protocol Labels" anchor="Subreg1"> 
       <!-- AH: addend Labels for balance with next section and IANA Cons. --> 
       <!-- 
	   AH: ATTENTION! 
	   As mentioned above, previous practice was to almost always 
	   specify only specific combinations. 
	   The most recent example, RFC 5389, is most convincing in 
	   that this is really needed. 
	   IMO, it should be considered to amalgamate this section with 4.*. 
	 --> 
       <!-- AH: striked one instance of Transport, 
	    and added text for better explanation. 
	    Added headline for transport protocols into table for balance 
	--> 
       <t> 
	 This section contains the known Protocol identifiers 
	 to be entered into the SRV Protocol Labels sub-registry. 
       </t> 
       <t> 
	 First, the IETF-specified transport protocols are listed, 
	 with the common labels also appearing in the IANA "Protocols" 
	 registry. 
       </t> 
       <t> 
	 In a number of RFCs there have been examples where 
	 services/protocols are defined to work over 
	 "Overlay" (or "Substrate") protocols such as xmpp and http. 
	 These are added next. 
       </t> 
       <t> 
         Table 1 lists the labels assigned to IETF standardized 
	 transport protocols, and those specified for other substrate 
	 protocols in existing RFCs published before this document. 
	 It represents the initial contents of the SRV Protocol Labels 
	 sub-registry. 
	 </t> 
       	 <texttable anchor='Transport Protocol registry'> 
	   <ttcol align='left'>Protocol</ttcol> 
	   <ttcol align='left'>References</ttcol> 
	   <c>Transport protocols </c> <c> </c> 
	   <c>_tcp  </c><c>  <xref target="RFC0793">RFC 793</xref></c> 
	   <c>_udp  </c><c>  <xref target="RFC0768">RFC 768</xref></c> 
	   <c>_dccp </c><c>  <xref target="RFC4340">RFC 4340</xref></c> 
	   <c>_sctp </c><c>  <xref target="RFC4960">RFC 4960</xref></c> 
	   <c>Overlay protocols </c> <c> </c> 
<!-- 
	   <ttcol align='left'>Overlay</ttcol> 
	   <ttcol align='left'>References</ttcol> 
--> 
	   <c>_http  </c><c>  <xref target="RFC4386">RFC 4386</xref></c> 
	   <c>_ipv6  </c><c>  <xref target="RFC5026">RFC 5026</xref></c> 
	   <c>_ldap  </c><c>  <xref target="RFC4386">RFC 4386</xref></c> 
	   <c>_sip   </c><c>  <xref target="RFC5509">RFC 5509</xref></c> 
	   <c>_ocsp  </c><c>  <xref target="RFC4386">RFC 4386</xref></c> 
	   <c>_xmpp  </c><c>  <xref target="RFC3921">RFC 3921</xref></c> 
	 </texttable> 
       <t> 
	 NOTE #1: For the purpose of the _Protocol labels and this registry, 
	 UDP-Lite <xref target="RFC3828"></xref> is treated 
	 as indistinguishable from UDP <xref target="RFC0768"></xref>. 
       </t> 
       <t> 
         NOTE #2: There have been a few underscore protocol labels defined 
	 for use by specific mail service databases such as 
	 "_domainkey" and "_vouch"; these are not registered anywhere. 
	 The related specifications do not employ SRV RRs however; 
	 therefore these labels are currently regarded as out of scope. 
	 It would be possible to register these in the registry above as 
	 RESERVED labels, to capture their use and avoid label collisions, 
	 if that is deemed useful. 
       </t> 
 
     </section> 
 
     <!-- OG: We used the names "SRV Application ..." and "SRV Service ..." -- 
       -- interchangably I changed all to SRV Service ... 
       --> 
     <section title="SRV Service Labels"> 
<!-- AH: should we swap the 2 sub-sections below, 4.1 and 4.2 ?  --> 
<!-- AH: since effectively _Service._Protocol pairs are registered, 
     I have changed the headline to match the IANA Cons. Section  --> 
       <section title="SRV Service Prefix Registration Template"> 
	 <t> 
	   Submit registrations to TDB@TDB 
	 </t> 
	 <t> 
	   [[ RFC Editor Note: final email address to be supplied by IANA! ]] 
	 </t> 
	 <t> 
	   Registration of SRV Service Prefix 
	   <list style='letters'> 
<!-- AH: should we add this one?  OG: YES --> 
	     <t> Kind of request: (new) or (update) </t> 
	     <t> Registrant name:  </t> 
	     <t> Organization: </t> 
	     <t> Contact e-mail and international phone number: </t> 
	     <t> Name of Service: </t> 
	     <t> Protocol(s) used: </t> 
	     <t> Reference to document(s) defining the service, (optional): </t> 
	     <t> Short Service description (optional): </t> 
	     <t> Delay publication: Y/N</t> 
	   </list> 
	 </t> 
	 <t> 
	   IANA will archive the accepted templates and make 
	   them available via links in the registry. 
	 </t> 
	 <t> 
	   <!-- OG: Stewart Chesire and IANA wanted this added --> 
	   <!-- AH: and I have added text to reinforce the uniqueness --> 
           IANA will accept and act upon applications for service identifiers, 
	   provided there is no Service name collision within the 
	   SRV Service Prefixes registry or with the Port Numbers registry. 
	   The publication of such assignments can be 
           deferred up to 30 days from the receipt of the application. Please 
           specify if delay is requested. 
	 </t> 
       </section> 
 
       <section title="Initial SRV Service Labels Registry Contents" 
	 anchor="Subreg2"> 
	 <!-- AH: matched title with IANA Cons. section --> 
	 <t> 
	   This section is expected to contain the most common service labels 
	   defined in the IETF that are in use with and without underscore prefix. 
	   It is expected that the template above will be used to register 
	   other service labels that are in common use. 
	 </t> 
	 <t> 
	   Table 2 specifies the initial contents of the SRV Service Labels 
	   sub-registry. 
	   This list is based on DNS server logs from September 2008 
	   and strings found in RFCs (published before September 2009). 
	 </t> 
 
	 <!-- 
	     I have harvested all RFCs and added much stuff to the table below, 
	     keeping it in collation order / fixing that where wrong !  A.H. 
	   --> 
	 <!-- OG: Cool thanks I harvested the recent 5xxx RFC's as well --> 
	 <!-- OG This table does not have full references and I prefer to keep it 
	      that way --> 
 
	 <texttable anchor='table_services'> 
<!-- OG: we do not need this 
	   <preamble> 
	     Registry format: 
	   </preamble> 
-->  
	   <ttcol align='left'> Service label </ttcol> 
	   <ttcol align='left'> Protocol(s) defined </ttcol> 
	   <ttcol align='left'> Reference(s): </ttcol> 
 
	   <c> _apex-edge </c>      <c> _tcp </c>         <c> RFC 3340 </c> 
	   <c> _apex-mesh </c>      <c> _tcp </c>         <c> RFC 3340 </c>  
 
	   <c> _beep</c>            <c> _tcp </c>         <c> RFC 3620 </c>  
 
	   <c> _capwap-control</c>  <c> _upd </c>	   <c> RFC 5415 </c> 
	   <c> _certificates </c>   <c> _tcp </c>         <c> RFC 4387 </c> 
	   <c> _crls </c>           <c> _tcp </c>         <c> RFC 4387 </c>  
 
	   <c> _diameter </c>            <c> _tcp, _sctp </c>   <c> RFC 3588 </c> 
	   <c> _dns-llq-tls </c>    <c> _tcp </c>         <c> </c> 
	   <c> _dns-sd </c>         <c> _upd </c>         <c> </c>  
	   <c> _dns-update </c>     <c> _upd </c>         <c> </c>  
 
	   <c> _dvbservdsc </c>     <c> _udp, _tcp </c>   <c> RFC 5328 </c> 
 
	   <c> _im </c>             <c> _xmpp, _sip </c>  <c> RFC 3921,  RFC 5509</c> 
	   <c> _iris-lwz </c>       <c> _udp </c>         <c> RFC 3981 </c> 
 
	   <c> _jabber </c>         <c> _tcp </c>         <c> </c> 
 
	   <c> _kerberos </c>       <c> _upd, _tcp </c>   <c> RFC 4120</c> 
 
	   <c> _ldap </c>           <c> _tcp </c>         <c> RFC 3088 </c> 
 
	   <c> _mihcs </c>       <c> _upd, _tcp , _scp</c>   <c> RFC 5679</c> 
	   <c> _mihes </c>       <c> _upd, _tcp , _scp</c>   <c> RFC 5679</c> 
	   <c> _mihis </c>       <c> _upd, _tcp , _scp</c>   <c> RFC 5679</c> 
	   <c> _mip6 </c>           <c> _ipv6 </c>        <c> RFC 5026, RFC 5555 </c> 
	   <c> _msrps </c>          <c> _tcp </c>         <c> RFC 4976 </c> 
	   <c> _mtqp </c>           <c> _tcp </c>         <c> RFC 3887 </c> 
 
	   <c> _pkixrep </c>        <c> _http, _ldap, _ocsp </c> <c> RFC 4386 </c> 
	   <c> _pres </c>           <c> _xmpp, _sip </c>    <c> RFC 3921, RFC 5509</c> 
	   <c> _pgp </c>            <c> _tcp </c>         <c> RFC 4387 </c> 
	   <c> _pgprevocations </c> <c> _tcp </c>         <c> RFC 4387 </c>  
 
	   <c> _rwhois </c>         <c> _tcp </c>         <c> RFC 2167 </c> 
 
	   <c> _soap-beep </c>      <c> _tcp </c>         <c> RFC 3288, RFC 4227 </c> 
	   <c> _sip </c>            <c> _udp, _tcp, _sctp </c> <c> RFC 3263, RFC 4168 </c> 
	   <c> _sips</c>            <c> _tcp, _sctp </c>  <c> RFC 3263, RFC 4168 </c> 
	   <c> _sipfederation </c>  <c> _tcp </c>         <c> </c> 
	   <c> _slpda </c>          <c> _udp, _tcp </c>   <c> RFC 3832 </c> 
	   <c> _stun </c>           <c> _udp, _tcp </c>   <c> RFC 4389 </c> 
	   <c> _stuns </c>          <c> _tcp </c>         <c> RFC 4389 </c>  
	   <c> _syslog </c>         <c> _tcp </c>         <c> RFC 3195 </c> 
 
	   <c> _tunnel </c>         <c> _tcp </c>         <c> RFC 3620 </c> 
 
	   <c> _xmlrpc-beep </c>   <c> _tcp </c>         <c> RFC 3529 </c> 
 
	   <c> _xmpp</c>            <c> _tcp </c>         <c> RFC 3921 </c> 
	   <c> _xmpp-client </c>    <c> _tcp </c>         <c> RFC 3920 </c>  
	   <c> _xmpp-server </c>    <c> _tcp </c>         <c> RFC 3920 </c>  
 
   <!-- AH: moved that note up into the prose. 
	     <postamble> 
	       List based on DNS server logs from September 2008 and strings 
	       found in RFCs (up to July 2009). 
	     </postamble> 
   --> 
	 </texttable> 
	 </section> 
	 <section title="RFC Examples and Pointers"> 
	 <t> 
	     There are a few instances where RFCs have what looks like 
	     SRV label names but are just examples and should not be 
	     registered. This section for completeness contains any 
	     such instances the editors are aware off. 
	 </t> 
	 <t> 
	     <list  style='symbols'> 
	       <t> _iris-*    RFC 3981 appendix A </t> 
	       <t> _mail      RFC 4985 </t> 
	 <!-- more or less an example there; is there a spec for that? --> 
	       <t> _ntp       RFC 4085 </t> 
	       <t> _pigen     RFC 3091 </t> 
	 <!-- An RFC is an RFC, independent of the publication *date* ! --> 
	       <t> _telnet    RFC 3620 </t> 
	 <!-- more or less an example there; do NTP folks have a draft for that? --> 
	       <t> _ssh       RFC 4592 </t> 
	 <!-- more or less an example there; do SSH folks have a draft for that? --> 
	     </list> 
	 </t> 
	 <t> 
	   There are a few instances where RFCs hint at external 
	   SRV usage/names but the editors have not been able to track 
	   down the documents. 
	 </t> 
	 <t> 
	     <list style='symbols'> 
	     <t>RFC 4123 says: ITU-T H.323 Annex O also defines SRV use. </t> 
	     <t>RFC 4280 points to 3GPP specs defining use(s?) of SRV. </t> 
	     </list> 
	 </t> 
	 </section> 
       </section> 
     <section title="Relationship to Other IANA Registries"> 
<!-- AH: removed indentation 
	 <list style="hanging"> 
--> 
	   <t> 
	     The SRV Service Labels registry is not a replacement for the two, 
	     more specific registries established by 
	     <xref target="RFC3861">RFC 3861</xref>, the  
	     "Instant Messaging SRV Protocol Label Registry" 
	     currently located 
	     at <eref target='http://www.iana.org/assignments/im-srv-labels'/> 
	     and the "Presence SRV Protocol Label Registry" 
	     currently located at 
             <eref target='http://www.iana.org/assignments/pres-srv-labels'/> 
	   </t> 
	   <t> 
	     Currently, there is one registration ("_xmpp") in each of these 
	     registries, performed by <xref target="RFC3921">RFC 3921</xref>, 
	     and another, more recent registration ("_sip"), performed by 
	     <xref target="RFC5509">RFC 5509</xref>, 
	     and this is reflected above. 
	   </t> 
	   <t> 
	     To avoid service name collisions, future registrations 
	     in the above two registries should always be accompanied 
	     by registration updates for the SRV Service Prefix registry. 
	   </t> 
	   <t> 
	     The IANA "Public Key Infrastructure using X.509 (PKIX) Parameters" 
	     "PKIX SRV Protocol Labels" sub-registry currently filed at 
             <eref target='http://www.iana.org/assignments/pkix-parameters'/> 
	     contains a single entry for the service label "_pkixrep" 
	     in combination with three protocols, based on 
	     <xref target="RFC4386">RFC 4386</xref>, 
	     which has been included in Table 2. 
	     Arguably, that sub-registry might be abandoned by a future 
	     update of <xref target="RFC4386"></xref>, 
	     in favor of making use of the new, general-purpose registry 
	     defined in this document, since it fulfills all requirements 
	     posed in sections 2 amd 4 of that RFC. 
	   </t> 
<!-- AH: removed indentation 
	 </list> 
--> 
     </section> 
 
     <section title="Security Considerations"> 
       <t> 
	 This draft creates a registry that should have been created a long 
	 time ago. This in its own does not have any security implications. 
	 However it is hoped that the registry will provide valuable 
	 information for administrators and security policy makers, to 
	 the benefit of the overall security community of the Internet. 
       </t> 
     </section> 
 
     <section title="IANA Considerations"> 
       <t> 
	 This document creates a new IANA registry with 2 sub-registries. 
	 The registry is named "DNS SRV Service Prefixes". 
	 The policy for creating new sub-registries is "IETF Review" 
	 <xref target="RFC5226"></xref>. 
       </t> 
 
       <t> 
	 The first sub-registry is: "SRV Protocol Labels". 
	 Allocation policy is: 
	 "IETF Review" <xref target="RFC5226"></xref>. 
	 The initial content of this sub-registry is specified by 
	 Table 1 (<xref target='Subreg1'></xref>).    <!-- Section 3 --> 
       </t> 
<!-- [AH]  next para commented out now (no discussion occurred) 
       <t> NOTE: If Overlay is split from the Protocol sub-registry 
	 that becomes an additional action. 
	 Do we need "Standard Action" policy 
         for "Transport" protocols vs. "IETF Review" for "Substrate" 
         protocols? 
       </t> 
--> 
 
       <t> 
         The second sub-registry is: "SRV Service Labels". 
         Allocation policy is: 
         "First Come First Served <xref target="RFC5226"></xref>, 
         but MUST NOT conflict with service names in the Port Number registry" 
	 (see <xref target='Collision'></xref>). 
         Rules for Labels to be registered are in Section 4. 
         The initial content of the registry is specified by 
	 Table 2 (<xref target='Subreg2'></xref>).   <!-- Section 4.2 --> 
         A Template for subsequent registrations is in Section 4.1. 
         IANA will archive and make available all accepted registrations 
         via links from the registry. 
       </t> 
       <t> 
	 This document updates the allocation procedure of the PORT NUMBERS 
	 registry located at http://www.iana.org/assignments/port-numbers 
	 such that service names for new registrations MUST NOT conflict with 
	 names registered as "SRV Service Labels". 
	 </t> 
     </section> 
</middle> 
 
<back> 
  <references title="Normative References"> 
    <?rfc include="reference.RFC.2782" ?> 
    <?rfc include="reference.RFC.5226" ?> 
  </references> 
  <references title="Informative References"> 
    <?rfc include="reference.RFC.0768" ?> 
    <?rfc include="reference.RFC.0793" ?> 
    <?rfc include="reference.RFC.2052" ?> 
    <?rfc include="reference.RFC.2119" ?> 
    <?rfc include="reference.RFC.2219" ?> 
    <?rfc include="reference.RFC.3232" ?> 
    <?rfc include="reference.RFC.3828" ?> 
    <?rfc include="reference.RFC.3861" ?> 
    <?rfc include="reference.RFC.3921" ?> 
    <?rfc include="reference.RFC.4340" ?> 
    <?rfc include="reference.RFC.4386" ?> 
    <?rfc include="reference.RFC.4960" ?> 
    <?rfc include="reference.RFC.5026" ?> 
    <?rfc include="reference.RFC.5507" ?> 
    <?rfc include="reference.RFC.5509" ?> 
 
  </references> 
</back> 
 


</rfc>

PAFTECH AB 2003-20262026-04-24 04:47:04