Talk:List of DNS record types

Zytrax Open Book
One of the "Open Books" on http://www.zytrax.com/ covers DNS and has been used as a base for a book, it seems to be pretty comprehensive: http://www.zytrax.com/books/dns/. It also contains a section about record types. --C167 (talk) 18:36, 26 November 2011 (UTC)

Copies of this information elsewhere
The SpamCopWiki holds a copy of the section of the Wikipedia DNS article that this article was based on. --Alvestrand 08:34, 22 August 2007 (UTC)

more record types (DNSSEC/IPSec/SPF)
As people can see from my recent updates, I've added/updated quite a few of the DNS record types, but I haven't touched several important classes, in particular the DNSSEC and IPSec records. Honestly, I do know know the state of either of these well enough to know which ones are current, which ones are obsolete, and which ones are irrelevant.

Should these record types be added? If so, does anyone know which ones?

I have also not added the type99 SPF record type as I was involved in that project. Wrs1864 (talk) 18:55, 12 May 2008 (UTC)


 * I believe I've added them all now - either to the list or to the list of what's obsolete and why. Please check! --Alvestrand (talk) 16:26, 18 May 2008 (UTC)


 * I double checked, and the only records that I found were missing were the IANA reserved ones (which I added).  I do have some uncertainty about some of the records though.  Wasn't the KX record obsoleted by the DS record?  Also, I thought A6 had also been obsoleted, but I guess from what I can tell it was only "moved from standard track to experimental".  I'm also not sure why RT is under the "not in current use" rather than with the rest of the AFSDB/X25/ISDN set. Wrs1864 (talk) 23:21, 18 May 2008 (UTC)


 * I think (working from memory) that KX is actually specified for some use in relationship with Kerberos, and has nothing to do with DNSSEC at all; DS was invented in an attempt to remove the need in DNSSEC version 1 for putting KEY and SIG records into glue records at the delegating point. A6 was intended as a more powerful and maintainable alternative to the AAAA record, but introduced so much possible complexity that it gathered a lot of opposition; as a compromise, it was relegated to experimental. RT is "not in current use" because I don't remember it well enough to say whether it's tied to an obsolete protocol or not. Someone (maybe me) will have to go read it. --Alvestrand (talk) 05:43, 19 May 2008 (UTC)

obsolete/irrelevant DNS records
There are a large number of DNS records that are either obsolete (MD/MF/MB/MG/WKS/X25/etc.) or have never really been used much (HINFO/GPOS/SINK/APL/etc.) which I have not included. I'm not sure if they should be mentioned in a small summary section, similar to how WKS is currently mentioned, or not mentioned at all, or if they should be put into the main table, maybe with a new column denoting obsolete/experimental/standard/whatever.

I am personally leaning toward the summary paragraph format but I am very interested in what other people think. Wrs1864 (talk) 19:05, 12 May 2008 (UTC)


 * For things that are out-and-out obsolete, such as WKS, or only used with standards that never saw deployment (such as the first generation of DNSSEC records), I think the summary paragraph is best. If there's any reasonable doubt about the status of the record, it should probably go into the list. --Alvestrand (talk) 13:42, 18 May 2008 (UTC)


 * I like your changes, thanks.  For what it is worth, I listed a record in the large list if there was already a wikipedia article.  By definition of what should be in wikipedia, all of those records are "notable".  Wrs1864 (talk) 15:09, 18 May 2008 (UTC)


 * Are you sure that HINFO is a hopeless case? RFC 4035 uses it in examples, folks sometimes talk about coordinates stored in HINFO, and IANA has registries related to HINFO (machine names + operating system names). --217.184.142.38 (talk) 21:37, 28 May 2008 (UTC)


 * I'm pretty sure it's dead - the examples in RFC 4035 are for a system running ITS and a system running TOPS-10. That sounds like Rob's sense of humor.... and the IANA registry at lists WANG but not WINDOWS-2000..... that's not a very up-to-date list.... --Alvestrand (talk) 22:16, 28 May 2008 (UTC)


 * I've actually seen a Wang about 25 years ago... ;-) Windows 2000 is apparently still covered by Windows NT5. Admittedly "last update 2002" sounds like dead, no XP, no Ubuntu, no Vista, how about another decruft "experiment" ? --217.184.142.28 (talk) 16:40, 30 May 2008 (UTC)

Both 217.184.142.38 and User:Dogcow have expressed doubts about whether some records should be labeled as dead, and I think they have a point. I'm sure that every single defined RR type is published *somewhere* on the internet, and there might even be a few things checking them, so technically, they are "in use". I changed the wording to say "not in use by any notable application", and if any of those records are used by notable applications, a wikipedia article can be written. However, that kind of points to whether some of the records in the main list are really notable, in particular things like CERT, DHCID, DLV, SSHFP, and TA seem least likely to be in any significant use to me. DNAME was once used to redirect 2001.ip6.int to 2001.ip6.arpa and even RFC 3364 which downgrades A6/bit-labels notes that DNAME usage has been proposed for things other than IPv6, but I don't know if it really is being used. The SSHFP option isn't even mentioned in the SSH article and  the TA record doesn't even have an RFC defined for it.

I guess one of my concerns is that we appear to be violating the WP:No original research and WP:NOTCRYSTAL guidelines by using our knowledge of DNS usage to decide what is worth getting a main list entry, and what gets shoved into the "obsolete" bucket. It isn't hard to find people who feel that IPSec, IPv6, and DNSSEC are all pipedreams (can you say DJB?) and therefore all of those records shouldn't be in the main list. On the other hand, I don't think we should throw every DNS record into the main list, that would just add clutter. Wrs1864 (talk) 19:55, 30 May 2008 (UTC)


 * what I'd suggest is to be pragmatic - let's just put things into the 2 categories using our best judgment, and if someone finds a RS saying one thing or the other, that should be cited, and category adjusted appropriately. FWIW - DNAME has to my knowledge never been deployed, but is still under serious consideration as a mechanism to use with non-ASCII TLDs (make them DNAMEs pointing to the "real" TLD rather than independent TLDs), and DLV/TA is a mechanism that is still in active controversy; ISC has (I think) made a "pseudo-root" available using DLV/TA, and is using that as a goad to get ICANN to sign the root.


 * WRT WKS, here's a citation from RFC 1123, a BCP document from 1989:
 * Not much has happened with WKS in the intervening 19 years. TXT, on the other hand, is definitely not obsolete. --Alvestrand (talk) 10:12, 1 June 2008 (UTC)
 * I had vague memories about DNAME being used, so I went poking and found this post from Bill Manning. (It was 2002.ip6.int and 2001:478::/32, not 2001.ip6.int as I mentioned above).  As far as WKS, I've added references to most of the obsolete articles, RFC 1127 is even stronger about WKS saying "Recommend against using WKS records from DNS."  I don't know about TA being used, but I do remember ISC launching to test root, I guess I was under the impression that it never went far.  Lots of stuff generates attention early on, but then fades out.  The problem I have with putting new records like TA/SSHFP/CERT in the list now is that once in the list, they likely won't be removed, they will just build up as cruft along with all other new records. Wrs1864 (talk) 12:03, 1 June 2008 (UTC)
 * The reason I changed the reference from 1127 to 1123 is that 1127 is informative and relatively unknown (it's mostly the author's pontifications upon the implications of 1123), while 1123, an IETF work product, is regarded as a part of the basic ruleset for Internet hosts. Agree that it's a bit of a philosophical problem what to do about new records, and don't have a strong opinion about what the right thing is (yet). --Alvestrand (talk) 13:58, 1 June 2008 (UTC)
 * Having read RFC 1127 now, RFC 1123 is in fact a better source. BTW, the hottest IETF Last Call debate I've ever seen was about 2821bis recently, and the history of WKS was discussed from all possible and impossible angles on the SMTP list. In the WKS case historical is not the same as irrelevant, but maybe too technical for Wikipedia, unless somebody bothers to write an article about the merits of MX +/- WKS / SRV / null MX / etc. Here WP:NOR should be no big issue when folks stay away from pushing their own ideas. --217.184.142.58 (talk) 19:53, 1 June 2008 (UTC)
 * My reading is that currently the WKS record is used chiefly as an example of how not to do things - when something is useful only when a majority of people do it, it has to be built in as mandatory from the beginning, or it will never get deployed; incremental additions have to be useful for the first people to deploy it. --Alvestrand (talk) 05:58, 2 June 2008 (UTC)
 * Well, this probably isn't the place to debate it, but IMHO, the far more critical flaw with WKS is that it is simpler just try the port to see if service is there. The information provided by WKS is less reliable and harder to use. Wrs1864 (talk) 15:06, 2 June 2008 (UTC)
 * Well, this probably isn't the place to debate it, but IMHO, the far more critical flaw with WKS is that it is simpler just try the port to see if service is there. The information provided by WKS is less reliable and harder to use. Wrs1864 (talk) 15:06, 2 June 2008 (UTC)

RP appears twice: "Not used by notable applications" and as the last item, describes its use "may be used for certain human-readable..."

Do humans count as 'notable applications'? I've run into registrars/zone operators who require RP records. Here's one: http://nic.southborough.ma.us/

I agree that RP records aren't as common as they might be. The web as a source of contact info and spam operators that harvest any source of email addresses have certainly contributed to their downturn in popularity. But I've still seen more RP records than LOC records.

In any case, the article should be consistent.

Also, the reference to the SINK record: I'm not sure it ever was intended to float. It does have an RR assignment (40) per http://www.iana.org/assignments/dns-parameters/dns-parameters.xhtml#dns-parameters-2 But if one reads the internet draft, it seems like an April Fools RFC (http://en.wikipedia.org/wiki/April_Fools'_Day_Request_for_Comments). The last version even has an April date -- though there were several drafts. At most, it seems more designed to make a point about the abuse of TXT records and the DNS as a universal database than as a serious proposal for a record type. I could be wrong.

Eastlake (the author) is still around - it might be worth contacting him to determine his actual intent. (That's not original research, which the entire list borders on, but a suggestion to approach a primary source.) Rather than publish an e-mail address here, I'll point out that he is listed on linked-in at https://www.linkedin.com/in/eastlake

Rectapedia (talk) 11:50, 25 March 2015 (UTC)

Two notes: 1) The RP record type is listed both in the main (actively used?) types and in the obsolete types.  Either it is obsolete or it is in use - NOT both. 2)  There was a lot of opposition to the deprecation of the SPF RRtype and disagreement with RFC 7208. If it is to be considered obsolete, this disagreement should be noted as it does still find some use. 2001:470:D:468:7455:7A:1C17:3DB4 (talk) 21:47, 8 April 2017 (UTC)

This list seems to be close to 100% original research, and most of it is wrong and/or very misleading. For example, NULL records are clearly not obsoleted in RFC 1035 and clearly are in use (e.g. in iodine, for the intended purpose). It seems most of the records in this section have not been obsoleted in any normal sense of the word, and there are usually no references to back it up, either - this is clearly made up misinformation.85.216.38.80 (talk) 20:32, 25 March 2018 (UTC)

AFSDB record
Is the AFSDB record used by only OpenAFS, or by all Andrew File System implementations? We should probably link to the main article if all of them use it. Wrs1864 (talk) 12:43, 24 June 2008 (UTC)


 * You have a point. The record is actually tied to the AFS protocol, rather than to any particular implementation. OpenAFS just happens to be the most widely used implementation. Goink

DNS Record Lookup Resource
This article ought to have a link so some online tool that allows you to view all the DNS records registered for any given name. This is a tool I have been looking for in the past and would find quite useful. If anyone knows of such a tool, please add it. 158.104.1.114 (talk) 03:02, 10 September 2008 (UTC)
 * if you're running a windows or unix/linux box, "nslookup" (a commmand line tool you can invoke from "cmd") is your friend. And no, I don't think this belongs in the article - it's context specific. WP:NOT is the relevant page to quote, I think. --Alvestrand (talk) 03:56, 10 September 2008 (UTC)
 * Such a tool shouldn't be on the main page. But in case anyone reading this needs it, here's one you can use: https://www.nslookup.io/ Ruurtjan (talk) 10:05, 16 September 2021 (UTC)

Unclear
"separate than that used in the SOA record" - this is plain English? Maybe "separated from ..." would mean something. --Doru001 (talk) 09:55, 14 June 2011 (UTC)

DNS Question / Idea
I have an idea about a feature that would change how the internet worked.

I propose that DNS have resource record types such as: porthttp,porthttps,portsmtp,portsmtps,portimap,portimaps,portpop3,portpop3s,portftp,portftps,etc...

I appears to me that such a move would change the way the world relies on IP addresses, since we appear to be running out. If you think how many servers there are in the world that have static IP addresses that only host one web page or even one service.

It seems like it would be fairly easy to implement a record type lookup to determine what port a website is using. The same thing could be accomplished for looking up what port a mail server is using. Yes the web browsers of the world could be adapted to do a ip and port lookup.

SSL Certificates could be adapted to include both the domain name and port.

DNS record types could be created and used to register what port a webpage is on.

For example: My webpage would have two DNS records. Record Type "A" host record, www.example.com = 127.0.0.1 Record Type "porthttp" port record, www.example.com = 80

This type of resource lookup would allow the internet to host a website or service on which ever port they wanted and would prevent the world from relying on common ports as the answer of where am I going to put my service.

Think about this. Currently one static IP address can only host one website service on a single static IP address. Routing and NAT take care of this. The only problem is there is no DNS record to lookup a port for the expected service. If a single IP address can have 65535 ports then a single IP NAT address could host 65535 websites on 65535 different ports. All is need is the ability to find the port that the http service is located on. — Preceding unsigned comment added by 68.106.179.249 (talk) 10:58, 28 July 2011 (UTC)

I also see this as a chance to solve the world's problem with ISP's blocking commonly used ports.

I don't have the programming skills to do the work, but if the idea was communicated to the world, the right people could make it happen.

09:55, 28 July 2011 (UTC) — Preceding unsigned comment added by 68.106.179.249 (talk)

Is this the place for this suggestion? Would be more helpful to send this up to the consortium that governs DNS...Nekksys (talk) 19:32, 26 August 2013 (UTC)

Seemingly contradictory DNAME explanation
The description of DNAME:

"DNAME creates an alias for a name and all its subnames"

seems to contradict the explanation given in http://en.wikipedia.org/wiki/CNAME_record#DNAME_record :

"An A record lookup for foo.example.com will fail because a DNAME is not a CNAME. However, a look up for xyzzy.foo.example.com will be DNAME mapped and return the A record..."

If these two statements are contradictory, one of them should obviously be corrected. If they are not contradictory, they need to be clarified. — Preceding unsigned comment added by 83.251.178.207 (talk) 14:28, 7 March 2012 (UTC)


 * You are right, Those do seem to be either contradictory or at the very least ambiguous. I have clarified the wording at CNAME record Please review and see if you can suggest any improvements. Thanks! --Guy Macon (talk) 05:49, 2 June 2013 (UTC)

lacking examples for people seeking technical use
cnames = really? i have no clue of what that does or what that's supposed to mean. — Preceding unsigned comment added by 75.135.156.3 (talk) 05:49, 29 December 2013 (UTC)

CSYNC
Missing CSYNC - RFC7477, record type 62, "Advises parent zone operator to update NS and Glue records"

Rectapedia (talk) 13:23, 26 March 2015 (UTC)

a bunch of explicit div declarations: use anchor instead?
I started editing the table section to add a call to the anchor template in order to be able to link to a specific RR (happens to have been "PTR"), and instead found a lot of div tags with id attributes. I suppose this serves the same purpose, the ability to (wiki)link to a document fragment with "#" in the URI. What would be the implications to changing these from div tags to calls to anchor? Is this the sort of thing which should be done?

DNAME redux -- not obsolete and never was
DNAME is not obsolete. It is not uncommon to use it in mergers and acquisitions.

RFC 2672, which the table claims obsoletes DNAME, does not. It clarifies its use and illuminates some operational pitfalls.

What may have confused folks is that one of its uses was incorrectly deprecated by RFC3363, which deprecation is reversed in 2672. On page 12 of RFC 6672, which is current: ''In [RFC3363], the following paragraph is updated by this document, and the use of DNAME RRs in the reverse tree is no longer deprecated.'' Note that this is the only use that was questioned. DNAME per-se was never deprecated or obsoleted.

The 2008 musings about "to my knowledge was never deployed" are not authoritative, and the extent that they may have been true at the time, certainly do not reflect current reality.

Please read 6672 carefully and correct this article.

A full list record type assignments and their official status is available at http://www.iana.org/assignments/dns-parameters/dns-parameters.xhtml#dns-parameters-4

Except as officially obsoleted or deprecated in an RFC, the use of "Obsolete" is inappropriate in this article.

If you can cite a reliable source for frequency of use, you could reasonably have an "infrequently used" section - assuming there's a reasonable threshold.

Considering the number of records currently in the DNS, a very small percentage can easily represent a large number of records.

Rectapedia (talk) 18:59, 15 December 2016 (UTC)

External links modified
Hello fellow Wikipedians,

I have just modified one external link on List of DNS record types. Please take a moment to review my edit. If you have any questions, or need the bot to ignore the links, or the page altogether, please visit this simple FaQ for additional information. I made the following changes:
 * Added archive https://web.archive.org/web/20080611185015/http://www.ops.ietf.org/lists/namedroppers/namedroppers.2004/msg00949.html to http://www.ops.ietf.org/lists/namedroppers/namedroppers.2004/msg00949.html

When you have finished reviewing my changes, you may follow the instructions on the template below to fix any issues with the URLs.

Cheers.— InternetArchiveBot  (Report bug) 01:37, 25 December 2017 (UTC)

RP RR type
Regarding the RP resource record... the data there are simply described as "Human Readable", but are actually pointers to another record (probably one or more TXT records) that contain the free-form text. Sort of like the MX record refers to another domain name A record.

Likewise, the other field is intended to contain an email address, encoded in the same fashion as the email address in the SOA record.

Louis Mamakos (talk) 20:38, 31 July 2019 (UTC) Louis Mamakos 