Talk:Comparison of DNS server software

Disclaimer: I started this article, but I pretty much have no domain knowledge here (so to speak), so if, e.g., the Features table is ridiculous because all DNS servers do authoritative mode, I won't object to it getting totally overhauled. —Fleminra 07:40, 2 February 2006 (UTC)
 * I think the article is a good start. Doing this article right is a one-week research project, but I know the basics for the freely available DNS servers from the top of my head.  Samboy 07:47, 2 February 2006 (UTC)

List or comparison
ATM the article is a List of DNS server software rather then a Comparison of DNS server software. The latter would provide extensive tables to really compare the products. ScotXW (talk) 17:49, 23 October 2013 (UTC)

Open source
Instead of a revert war over djbdns and “open source,” how about just being unambiguous about what the article’s table means by “open source” (e.g. using and  ). Or maybe there should be two columns: “license is OSI-approved” and “source code is freely available.” —Fleminra 02:38, 23 February 2006 (UTC)

Cleanup needed
This article lacks encyclopedic context; it needs some text to explain why it is here and the importance of the information it provides. Many of the entries are missing significant information, and the choice of attributes compared seems arbitrary, especially considering that one of the headers is a red-link. &mdash;donhalcon╤ 21:21, 7 March 2006 (UTC)
 * FYI, WP:POINT (see Special:Contributions/Donhalcon). —Fleminra 23:30, 7 March 2006 (UTC)

BIND works on Windows
I'm almost certain BIND can be run on Windows using cygwin. Can someone else confirm this for me before I do an edit?

There is even a native BIND9 port. http://www.isc.org/index.pl?/sw/bind/view/?release=9.4.2 —Preceding unsigned comment added by 194.97.7.90 (talk) 15:09, 8 February 2008 (UTC)

Author update
I'm the PowerDNS author and I just edited this page, I know this is frowned upon, but I fixed two mistakes. Ahu 20:38, 19 May 2006 (UTC) — Preceding unsigned comment added by Bert Hubert (talk • contribs) 20:38, 19 May 2006 (UTC)

Problem of DNSSEC support on Windows 2003 DNS
Windows 2003 DNS server does NOT support DNSSEC specified by RFC4033, RFC4034 and RFC4035(published in 2005). It partially implemented the DNSSEC specification in RFC2535 that is now obsoleted by RFC4033, RFC4034 and RFC4035.

So, at this time, it would be correct to say that Microsoft DNS does not support DNSSEC any more. It seems that Microsoft DNS cannot recognize DS RR, RRSIG RR, DNSKEY RR and NSEC RR, instead it can handle SIG RR, KEY RR and NXT RR(specified in RFC2535).

If above description has no problem, the contents of the table concerning DNSSEC support should be fixed, I think. --Ngtao 12:46, 13 June 2006 (UTC)

new "interface" column
Maybe the new "interface" column should be "management interface"? Just because the most heavily used interface of all of these servers is the interface that listens on port 53 and speaks DNS protocol. Or maybe I'm being pedantic. —Fleminra 20:39, 7 June 2006 (UTC)

BIND on Windows
There are windows binaries available on the BIND website, suggesting that it does work with windows!

From the Bind Website download page:

A binary kit for Windows NT 4.0 and Windows 2000 is at http://ftp.isc.org/isc/bind/contrib/ntbind-9.3.2/BIND9.3.2.zip http://ftp.isc.org/isc/bind/contrib/ntbind-9.3.2/BIND9.3.2.debug.zip

--Rikjeremy 22:54, 22 June 2006 (UTC)

Adonis updates
Regarding the recent Adonis-boosting changes: I think these new columns need to be qualified. The implication that Adonis is the only DNS server software that does error checking of any kind is pretty dubious. —Fleminra 23:12, 21 August 2006 (UTC)
 * The high-avability feature is also, IMHO, pretty dubious. Basically, DNS has high-avability built-in, so IMHO a DNS server doesn't need it. 68.107.75.220 19:56, 10 December 2006 (UTC)

High availability built-in
What exactly does this column mean? It sounds like marketing garbage that does not belong on Wikipedia. 71.96.189.16 02:46, 22 October 2006 (UTC)


 * Basically, a marketing department for a commercial DNS server added it to help promote their DNS server on this page. I have removed it. 68.107.75.220 19:50, 10 December 2006 (UTC)

Misc

 * BIND is included in most if not all Linux distros, shouldn't it be in the included column for Linux?
 * Yes, with a footnote explaining that it is a package for the major Linux distros (RedHat/CentOS, Ubuntu, SuSE, etc.) and may or may not be a part of a given non-mainstream distribution. Samboy 18:31, 6 January 2007 (UTC)


 * A column for GeoIP support would be handy.
 * For the three open source, currently supported DNS servers: PowerDNS has GeoIP support; MaraDNS and BIND do not. For the uninformed, GeoIP is the ability to give a different DNS reply based on the location one is querying a DNS server from. Samboy 18:31, 6 January 2007 (UTC)
 * BIND 9.10 now includes GeoIP. Cba.isc (talk) 04:02, 5 August 2014 (UTC)

Article cleanup
OK, I just did some article cleanup:


 * I removed three DNS appliances that use BIND as the resolver. This is a comparison of DNS software, not the marketing department for companies that add a web interface to BIND.
 * I removed MacDNS and Dents; both of these projects are dead, have been dead for years, and have no hope of ever coming back to life again. Dents, in particular, was not a usable DNS server when it died eight years ago.
 * I removed a red link to Comparison of DNS resolver software, as per Red link guidelines. If you are going to write this article, go ahead and add this link again.

Samboy 19:04, 6 January 2007 (UTC)

inherent bias in open source column
The open source column uses green/red colored backgrounds for yes/no. I feel that this creates an inherent bias against closed-source products simply by the human reaction and normal use of those colors. I'm removing the colored background for that column only. --mcpusc 02:34, 12 May 2007 (UTC)

What is Open Source?
What is Open Source? There are a lot of people who argue it just means "The Source code is available", and doesn't have to be OSI compliant. See this recent LWN article and this Slashdot flame-fest for the point of view of people who claim "Oh, 'Open Source' just means the source code is available". Now, can anyone find a pre-OSI use of the term "Open Source" to indicate the source code of code was available, and how it was used. Just go over to groups.google.com and find a real use of "Open Source" in 1996 or before.

Just some food for thought. And, yes, I feel "Open Source" means, simply, OSI-compliant, but I have gotten in to an edit war over a DJB advocate about whether we could call DjbDNS "Open Source" here, even though djbdns is not OSI-compliant. Samboy 05:31, 8 July 2007 (UTC)


 * OK, looking at the Slashdot flame fest, I found a reference to a 1996 use of the term "Open Source" Samboy 05:57, 8 July 2007 (UTC)

Should there be a column for DHCP integration?
dnsmasq can provide DHCP services and then serve DNS requests about the clients under it's DHCP control. I *believe* the Microsoft product does this too, but I have no verification of that. TonyAiuto 19:55, 23 July 2007 (UTC)
 * No, the Microsoft DHCP server doesn't provide integration with the DNS server. A client may try to update its dns entry with the IP address it received from a DHCP server, but this isn't integration at all RichardLetts (talk) 00:09, 24 March 2010 (UTC)

I would agree that the feature matrix does need a DHCP column for clarification.75.134.135.128 (talk) 19:36, 15 November 2011 (UTC)

Lose "Open Source"
This is a subjective term that extraordinarily complicates the table. If this page is to stay in the WP, the "Open Source" column needs to change to "freely available source" or some other neutral term. Not even Free Software advocates agree on the meaning or implications of "open source".

I'll wait for feedback for a few days, and then I'm going to go make this change.

--- tqbf 19:13, 20 September 2007 (UTC)

ANS
Thanks for scrubbing, but ANS is hugely notable; it's Nominum's server, and Nominum is Mockapetris' company. Nominum is the to the IETF what Blackwater is to the US Army. —Preceding unsigned comment added by Tqbf (talk • contribs) 19:19, 6 November 2007 (UTC)


 * If ANS is notable, then write an article about it. Comparison articles attract heavy amounts of spam, and the easiest way to tell if something is notable is whether it survives on its own as an article. AlistairMcMillan (talk) 01:28, 21 November 2007 (UTC)


 * Notability guidelines do not directly limit article content WP:NCC
 * Listing a viable product here is not spam, and he should not have to write an article for ANS to be included on this list. Please see topic below
 * --Dns9900 (talk) 10:52, 21 November 2007 (UTC)

Removing "unnotable" / "spam" products
Please don't remove DNS server products just because you feel they are "unnotable" or "spam".

Notability only applies to articles - it does not limit the content of articles. See WP:NNC

The suggestion about writing an article about a product to gain notability, in order to get listed here doesn't make sense. If a product isn't notable (and some of the removed ones probably weren't) - they can't have an article - see WP:N But again, that doesn't preclude such products from being listed in contents such as this.

The argument that listing products here is spam is also wrong. How can you have a fair comparison of products if you deliberately leave some out (or even delete some)?

Of course there is going to be "me too" in a comparison article. But as long as it is kept to a simple clean list without marketing jargon etc. - I don't really see any problem with that.

This article should list as many products as possible. Please add more instead of removing. —Preceding unsigned comment added by 87.56.149.202 (talk) 07:04, 21 November 2007 (UTC)


 * Thanks for writing this.


 * I tend to agree that the bar for inclusion in this article should not be "would survive as its own article".


 * I would tend to agree that this article is being a bit too sharply patrolled. Clearly I think CNS/ANS belong in the article, for instance.


 * At the same time, if you look at prior revisions of this article, it clearly had become a magnet for spam, with vendors changing the comparison axes to suit their own products.


 * I also strongly disagree that we benefit from having as many products as possible here. It is not valuable or illuminating to list the details of non-viable DNS servers on this page, unless there is something otherwise notable about them. WP is not a directory. WP is not a soap box.


 * Finally, you are required to assume good faith. Specifically, If criticism is needed, discuss editors' actions, but it is not ever necessary nor productive to accuse others of harmful motives. I don't work for a DNS server vendor, nor do I have any COI issues here, but I'm not required to make a case for that to you. --- tqbf  07:13, 21 November 2007 (UTC)

I agree that some recent edits were very obvious marketing attemts, and such should not survive. However just having a product on the list is not spam.

And I concur that "viability" is a much better criteria for inclusion than "would survive as its own article". Dead products should obviously be removed (have no web-site, no users, or hasn't been updated for a long time, etc.). But excluding otherwise viable products based on "notability" seems like discrimination to me. Why should the "big guys" get such a huge advantage? WP matters a lot!

As for WP is not a directory - that seems to bring into question the validity of the whole article. Does removing less "notable" products make the article less of a directory? I think it just makes it a biased directory WP:NPOV.

I apologize for previous "who do you work for?" remarks. This was not in good faith. It is just that deleting those DNS servers looked a lot like big guys trying to squash smaller competitors - which seemed hugely unfair.

I strongly believe that this article has no value and should rather be deleted completely if it only compares the "big guys" and doesn't include viable alternatives - WP:NPOV - even if such are not "notable".

--Dns9900 (talk) 09:01, 21 November 2007 (UTC)


 * I agree with Dns9900 - this should not be an exclusive showcase for big products.
 * We have been using one of the deleted products, Simple DNS Plus, for many years and this is certainly a very notable product used by thousands of Windows users - including some very major IPS.
 * It may not be notable in WP terms, but I don't believe a DNS server comparison is complete or fair without it.
 * And writing a WP article about a product doesn't make it more notable in my book.
 * --RicDeu (talk) 11:49, 21 November 2007 (UTC)

Software doesn't become notable by having an article. The point is that when you write an article you are supposed to justify the notability in the article, usually in the opening paragraphs. We can't have this in the comparison article, it needs to be in individual articles. AlistairMcMillan (talk) 13:16, 21 November 2007 (UTC)


 * You seem to be missing the point; Notability is NOT a criteria for being included in content. See WP:NNC.
 * Deleting those DNS servers is spam by omission.
 * If you want to delete some entries - argue the case for each one instead of just bulk erasing everything you don't like.
 * Please tell us why you think only Microsoft and ISC (BIND) are allowed here?
 * Do they have some special right to "spam" in WP?
 * Respectfully reversing - again.
 * --RicDeu (talk) 15:47, 21 November 2007 (UTC)

Awesome that there's some heat in this article again! Thanks to ip:87.56.149.202 for getting that started.

I have a question: is there some line in the sand we an draw about product inclusion? I've written 3 DNS servers, but you've never heard of them, because (in one case) I made little to no effort to publish it, and (in the other two cases) they were small parts of larger products. Clearly none of them belong in the article.

The original reasoning from 87.56.149.202 was that they had found some of the deleted servers highly ranked in a DNS survey. Maybe popularity is a reasonable measure?

I'm less concerned with "spam" than I am with "crudding up the article". —Preceding unsigned comment added by Tqbf (talk • contribs) 17:37, 21 November 2007 (UTC)


 * Excellent initiative Tqbf!
 * Dead products should obviously be removed. Further up this page there's a great sample of how that should be done (documented and argued for).
 * Usage and popularity would be great measurements but are updated and reliable sources available for this?
 * I think viable products that are actively maintained and/or marketed should be here. Don't know how to define "viable" though?
 * --Dns9900 (talk) 20:01, 21 November 2007 (UTC)


 * "Perfection" is not, unfortunately, available to us. I'd argue that in the same way that an article subject must credibly assert notability in its lede, a short description of each product could (perhaps implicitly) assert the reason it's interesting in the context of this article. I will admit that, for instance, in the case of SimpleDNS, going and looking it up so I could write a brief intro changed my mind; clearly it belongs here, it's a popular choice for Windows users who want to publish names for their newly acquired domain. --- tqbf  20:05, 21 November 2007 (UTC)


 * Reminder: you can very effectively (1) make your case that a piece of software belongs and (2) improve the article dramatically by writing a brief description of the product in question at: Talk:Comparison_of_DNS_server_software/Product_Descriptions. --- tqbf  20:07, 21 November 2007 (UTC)

Intros for each piece of software
I've started a new intro section for the article at Talk:Comparison_of_DNS_server_software/Product_Descriptions.

I did this for two reasons:


 * 1) Looking down the list of servers, I was irritated that I didn't know what many of these servers were or why I cared, and I'm pretty familiar with DNS.
 * 2) Forcing ourselves to write a short intro for each piece of software may clarify whether the server merits inclusion in the list.

Can some of you help fill in the blanks? Especially for those servers you insist be included in the article. If you can't even describe them in 2-3 sentences, they probably don't belong. :)

--- tqbf 17:50, 21 November 2007 (UTC)


 * Great initiative - but where is the new intro section? was it removed again?
 * --Dns9900 (talk) 20:08, 21 November 2007 (UTC)


 * The intro I'm working on is linked above, it's a subpage of the talk page (it will look like crap pasted in the article half-done). This article could use a much better 2-3 sentence lede (what right now reads, "this is a comparison of different DNS servers...", yeah, duh?). --- tqbf  20:41, 21 November 2007 (UTC)

There seems to be an unintentional bias towards BIND on this page. It states that BIND is the "de facto standard" and the "most widely deployed DNS server," which may be correct, but the two references used here are from 2005 and 2004, respectively. These two references also refer only to public DNS servers, which may at first be misleading to anyone who is comparing DNS servers for internal/private networks use. I'm not suggesting a debate about the truth of these statements, but that we may be able to clarify the context of the statement. - Samerde (talk) 19:23, 19 December 2011 (UTC)

Moved Packaging To Bottom
Two reasons:


 * 1) It's not the most important set of comparison axes (Wikipedia is not a buyer's guide; we have to assume readers are coming to this article to learn more about DNS through the lense of the characteristics of different software).
 * 2) It starts to resolve a structural problem with the article, that being: we need to introduce the concept of "authority" and "recursive" servers before we can even introduce the software, because otherwise the article is tossing around terms like "authoritative" and "recursive" before defining them.

--- tqbf 17:53, 21 November 2007 (UTC)

Remove "DNS Views Support" from Feature Matrix
There is no standard "DNS Views" feature mentioned in any RFCs or other standard documents. This column probably originates from the proprietary "Views" feature in BIND. I think this column should be removed because different DNS servers have different features similar to BIND Views - but not exactly the same - and therefore this column is now used for marketing jargon foot notes about how each product does this better. Comments please. --Dns9900 (talk) 20:15, 21 November 2007 (UTC)


 * I'm unclear on this too, but assume it means something along the lines of "split view" DNS, which actually is a very important (abstract) feature --- it means you can use a single server to host both your internal (secret) names and your external (published) names. --- tqbf  20:25, 21 November 2007 (UTC)

Yes, this is how I understand the feature too, and yes it is important. The problem is that different DNS servers implement this concept so differently that it is not really directly comparable. The result is "Yes - but..." for every DNS server. --Dns9900 (talk) 20:36, 21 November 2007 (UTC)


 * I can't back this up, but I'm guessing that's not really true. Sam follows this article: can you do split-view DNS with MaraDNS? Can you really do it with something like SimpleDNS? --- tqbf  20:40, 21 November 2007 (UTC)


 * You can do it when you run two processes of DNS Server whatever flavour and then use NAT/firewalling to direct network a to process a and network b and the rest to process b --User:ZaphodB ZaphodB Fri Feb 8 16:14:21 CET 2008  —Preceding unsigned comment added by 194.97.7.90 (talk) 15:14, 8 February 2008 (UTC)

SPAM
Would those that are insisting that we include everything please suggest how we stop this article from accumulating spam. AlistairMcMillan (talk) 20:48, 21 November 2007 (UTC)


 * I'm ambivalent, so I'm not who you're addressing, but why isn't "Intros for each piece of software" a good start? I'm with you on minimizing cruft and spam, but not with you on attempting to bulletproof the article. Again: let's just have product come with a short description of why it's interesting --- then, as they accumulate, we can evaluate the claims made in the description and have something very specific to talk about.


 * I feel like we'll get nowhere talking about this in the abstract. --- tqbf  20:51, 21 November 2007 (UTC)

I am not sure how you (AlistairMcMillan) define "spam"?

Marketing jargon and sales pitches etc. certainly is spam. I think everyone can agree on that.

There is probably no way to permanently stop an article like this from attracting that type of spam, but it is simple to remove it when that happens.

A way to cut down on it might be to remove the "DNS Views" column - see above.

However I don't think it is spam to simply list a viable DNS server product in an article titled "Comparison of DNS server software".

I do think it is spam (by omission) to delete such products - or try to force deletion by making them link to non-existing WP articles.

--Dns9900 (talk) 08:31, 22 November 2007 (UTC)


 * Not sure how removing valuable content from the article is a solution to spam. --- tqbf  02:29, 24 November 2007 (UTC)

Notable content / Notability vs. Articles
In his comments for article update 172973902 on November 21st 2007 user:AlistairMcMillan wrote


 * Fine then, if everyone is insisting these pieces of software are notable enough then write articles about them

Then he proceeded to remove all external product links and make new dead links to non-existing WP articles.

Two things:

1) Why do you insist that the individual products listed in this article must be notable?

The WP guidelines clearly state that article topics should be notable WP:N but also that this does not limit the content of the article WP:NNC - in this case the products listed.

2) Why do you insist that being notable equals having a WP article? (not that it really matters in this case - see 1)

I understand that something must be notable in order to have a WP article, but I don't see how not having a WP article makes something not notable?

Further I would argue that some of the existing WP product articles linked to by this article are so thin that they shouldn't be there at all.

By trying to force articles to be written about products - it seems you are only inviting more spam (see above) and lame articles.

And FYI - other users are asking similar questions about this and other topics on your user page User_talk:AlistairMcMillan

Respectfully - --Dns9900 (talk) 09:28, 22 November 2007 (UTC)


 * I'm insisting that individual products listed on this article must be notable, because people using Wikipedia to promote their own products really fucking pisses me off (please note that I'm not the only person who is sick of it). Please note that WP:NNC, which you keep linking to, says that although article content isn't covered by Notability guidelines it is covered by our rules on sourcing and linking to a product website isn't considered suitable sourcing (WP:PROVEIT).


 * And I never insisted that having an article proves notability. I'm insisting that we link to articles, because the articles can include proof of notability. And I agree with you wholeheartedly that some of the linked articles are thin and don't establish notability. The point is that if the article can't establish notability, then the software shouldn't be listed here.


 * And you may not like the WP:DIRECTORY policy, but it is official policy not even a guideline but official policy.


 * BTW I think it is wonderful that I'm being questioned by someone who is choosing to remain anonymous and whose only goal here seems to be to insist on us filling an article with external links. AlistairMcMillan (talk) 22:28, 23 November 2007 (UTC)


 * In fairness: I share some of his questions, though I'm glad we got rid of the external links. --- tqbf  22:38, 23 November 2007 (UTC)

We completely agree that WP should not be used the promote products.

But why then do you (AlistairMcMillan) insist that it is OK for Microsoft and ISC to promote their DNS server products here, but it is not for smaller competitors?

Microsoft's product may be notable (due to sheer size of the company), but how does that make their listing less of a promotion?

Following your argument (promoting products f... p... you off) - then shouldn't all these products be removed?

I am fine with the external links being removed (sourcing/linking), as long as the products stay on the list.

But frankly I think replacing the external links with the dead WP links is disruption WP:POINT. There is no policy that I could find saying that every content item has to link to another article.

I have no problem with WP:DIRECTORY. My argument was that this applies to the whole article.

If you want to delete the whole article for being a directory - that's fine with me.

But as long as the article is here, it should be fair and unbiased WP:NPOV listing all relevant products.

I have no desire to fill this or any other article with external links! Now it is my turn to invoke WP:AGF

My only goal here is to prevent big companies and special interests from having exclusive rights to promote themselves here at the expense of smaller lesser known companies/products.

Funny, I thought WP encouraged anonymous editing...

--Dns9900 (talk) 02:02, 24 November 2007 (UTC)


 * Please list your evidence that Microsoft etc are using this article on Wikipedia to promote themselves. AlistairMcMillan (talk) 02:18, 24 November 2007 (UTC)

Dns9900: it is not unreasonable to ask that, for the products on this page you are most invested in, you at least take the time to write 2-3 sentences describing what they even are. Do so here: Talk:Comparison_of_DNS_server_software/Product_Descriptions. You've now written many hundreds of words in defense of perhaps 50 characters worth of editing to this article. Put your edits where your mouth are and help us improve the article. --- tqbf 02:32, 24 November 2007 (UTC)


 * AlistairMcMillan: I have no evidence that Microsoft is doing so. Do you have any evidence that the products you are trying to delete are?


 * tqbf: I will be happy to - I'll do a bit of research and make some updates shortly.


 * --Dns9900 (talk) 03:15, 24 November 2007 (UTC)


 * There is definitely evidence of other vendors doing that. Read the history. --- tqbf  07:08, 24 November 2007 (UTC)

Removed entries
I have to vent here for a sec: it makes me feel like a bit of a jackass to have wasted Alistair's time bickering about keeping spam entries by siding with the "keep" votes here and lawyering about WP:NNC when it turns out most of these products aren't even themselves implementations of DNS, but rather repackaged versions of BIND.

It strikes me that in all the time spent bickering about this article, one of the proponents of, say, "IPControl", could have taken the time to read the IPControl data sheets and the 1 magazine review the product had, wherein it's disclosed that the DNS Server in the IPControl product is simply BIND. The same goes for IPMDNS --- it would be hugely notable if there was a major commercial DNSSEC+TSIG server in the gene pool, and I'd want it included here as well. But there isn't; this product is simply BIND on a Linux appliance. And "Nixu SNS"? Come on! This is a commercial Linux distribution, not a name server.

Also, it seems disingenuous to argue in favor of entries that you can't be bothered to take the time to complete. So, for instance, it turns out that DNSMASQ isn't a nameserver, but rather a forwarding proxy. And DNRD as well, though you wouldn't know from its entry in the article, which none of the DNRD proponents took the time to fill out. Just as well, it doesn't belong here.

--- tqbf 18:08, 2 December 2007 (UTC)

Comment: Hi, i'd like to comment on this item and kindly ask you to reconcider adding Nixu NameSurfer Suite back to this list. I work for this company, but my intention is not to advertise the product or the company, but i just feel that Nixu NameSurfer fills the requirements and is comparable to other DNS Server implementations. I can see that previously "Nixu SNS" was attempted to be added to this page, but clearly it's not a separate DNS Server implementation but an applianse, containing BIND. Nixu Namesurfer OTOH has a proprietary DNS Server implementation that for example is included in other commercial products as the "DNS Engine", such as Appliansys DNSBOX. As said above: "it would be hugely notable if there was a major commercial DNSSEC+TSIG server in the gene pool, and I'd want it included here as well." Well, Nixu NameSurfer supports DNSSEC (as of version 6.0) and TSIG. I try to be unbiased here, but it's also one of the few DNS Server implementations that use SQL backend natively instead of flat files (Postgres 9 currently). So for people looking for options (for BIND), it's a viable choice. The implementation is also one of the oldest, it's been available for customers since -95 if i remember correctly. Honestly, i don't see how you can keep for example CNR and Secure64 in the listings and deny Nixu NameSurfer. If you feel that it could be added, i'll make sure the way it's presented is non-promotive or just let someone else write the entry. Thanks. :) - Ville Kummu / Nixu Software — Preceding unsigned comment added by Vkummu (talk • contribs) 17:58, 20 October 2011 (UTC)

Ordering of entries
"The big 3"


 * 1) BIND: No-brainer
 * 2) Microsoft: #2 by popularity, and a good foil for BIND
 * 3) djbdns - Most popular "alternative" server

The second tier:


 * 1) simpledns: a weird call, but it's (a) popular and (b) a good foil for the big 3
 * 2) NSD: an extremely important server, but less useful in the article than simpledns
 * 3) PowerDNS: from what I can tell, the only other popular open source server, and WP's server

Then:


 * 1) MaraDNS: an actively maintained open source server
 * 2) ANS and CNS: Notable as Mockapetris' products
 * 3) Secure64: A DNS server with its own OS (!)
 * 4) Posadis: appears to be defunct


 * Note: the notion of "tiers" of DNS servers is WP:OR, not intended for the article itself.' --- tqbf  19:51, 2 December 2007 (UTC)

--- tqbf 19:18, 2 December 2007 (UTC)


 * actually, i think it should just be alphabetical. their respective popularity should be noted, but for ease of use of the list, straight alpha seems better to me. Anastrophe 19:42, 2 December 2007 (UTC)


 * Strong disagree. Alpha means that one of the first things you see in the lists is ANS, one of the least notable servers. It makes it unnecessarily hard to compare Microsoft to BIND, and BIND to djbdns, which are the two most important comparisons to be drawn out from the table. Alpha seems like the choice you'd make if you were conceding not to know how important the different servers are. But we do know.
 * Candidly, I can't see how coverage of, say, Posadis and MaraDNS can be justified if the list is alpha; it means that each server we add potentially hurts the quality and utility of the article. I know, I'm thinking about this too much. --- tqbf  19:50, 2 December 2007 (UTC)


 * it was more a mild suggestion. i'm not married to any particular format. Anastrophe 04:49, 3 December 2007 (UTC)


 * Nah, I'm being defensive because I know I'm going to lose this argument; but if we can leave it this way until it (inevitably) becomes an issue, I'll be happy. --- tqbf  04:52, 3 December 2007 (UTC)

RR type support
Just out of curiosity, does anyone know why LOC record support is noted, while other more important/common record types, such as SRV, NAPTR, TXT, SPF (the new type 99 records), etc. are not? I am guessing they all support things like A, NS, MX, SOA, PTR, etc. Wrs1864 —Preceding comment was added at 04:18, 3 December 2007 (UTC)


 * Because someone decided it was an important basis for comparison? Someone is clearly wrong, because pretty much all of the servers can claim to support LOC. --- tqbf  04:53, 3 December 2007 (UTC)

BIND9 Rewrite
BIND9 and BIND4, both notable, popular DNS servers, are profoundly different servers. BIND8 and BIND9 have substantially different codebases. I wouldn't want to break them into 3 servers, but the rewrite is relevant to the comparison.

--- tqbf 01:18, 16 December 2007 (UTC)


 * i don't disagree, but the specifics are relevant to the BIND article, rather than this one. so long as this one notes which version is being used as the basis for comparison, older versions (for any software, for the most part) aren't terribly relevant. BIND 4 and 8 are deprecated, so if someone were to go to the BIND website on the basis of this or the BIND article, they'd be directed to the BIND 9 download. seems implicit to me in that regard. if this article were to include BIND 4 and 8 for comparison, then the differences would be notable. Anastrophe (talk) 01:22, 16 December 2007 (UTC)


 * The article stated there were "four different versions" of Bind, being Bind4, Bind8, Bind9. I've done the math and changed that to 3.  For that matter, is anyone still running Bind4?  I have not personally seen it in use in over 10 years, even though it may exist as a code base for some proprietary or deeply firewalled products.  Fehrgo (talk) 11:27, 15 January 2008 (UTC)


 * i recall seeing assorted dns server surveys over the years, and there still is a base out there of hosts running BIND 4. the implications of that are horrible of course, but it's the reality. BIND 4 code is no longer available on the ISC website however, so i do wonder whether its useful to include it in a comparison list such as this, since nobody is able to perform a new deployment of it. it's more a historical artifact rather than DNS server software one might consider using. Anastrophe (talk) 17:14, 15 January 2008 (UTC)

Proprietary license
The article that we appear to be using to define "proprietary" licenses has a great big "This article does not cite any references or sources." tag at the top. Not entirely credible. Our article on Proprietary software which most of the software comparison articles pointed to states at the top "Proprietary software, coined by the free software community...". Comparing software by using the terms produced by one side of an argument is something we should be avoiding. AlistairMcMillan (talk) 01:49, 4 April 2008 (UTC)


 * okay, but then what shall the licenses be called? or should the column be changed, perhaps to "Cost", which would give a clearer representation. most people are interested in software that's free as in beer. the politics of free as in liberty is, regardless of what many may counter, not that important to the majority of software end-users. Anastrophe (talk) 05:28, 4 April 2008 (UTC)


 * There already is a column called Cost. These pieces of software must be released under some license. Is "EULA" the correct equivalent for GPL/BSD/etc? AlistairMcMillan (talk) 19:30, 4 April 2008 (UTC)

djbdns is faster?
I added a Citation Needed to the claim "djbdns components are considered faster than their BIND counterpart" although calling them weasel words may have been appropriate too. I'm in favor of removing this phrase completely since it's the only place in the article where a specific claim of performance is made, and it seems out of place to me. In my limited searching, I've only managed to find one performance comparison source, apparently unbiased, which actually claims the opposite: Domain Name Server Comparison (although this paper has a fairly negative tone on djbdns in particular). Gurnec (talk) 16:08, 5 May 2009 (UTC)


 * it is unreasonable to give other editors less than one day to find sourcing for a challenged claim. please restore the material you challenged - with the challenge tag - and allow our fellow editors at least a little time to investigate. a good rule of thumb is one month, minimum, for your fellow editors to look into the matter. that's why 'fact' tags are qualified with the month they were applied, as less than that is too fine a granularity for what is supposed to be a collaborative process. thanks. Anastrophe (talk) 05:40, 6 May 2009 (UTC)


 * it's also worth pointing out that the one paper you've referenced is from an unreliable source - in that the author has overtly misrepresented numerous facts about djbdns. see . that's certainly not a reliable source to hang this criticism on djbdns. Anastrophe (talk) 05:47, 6 May 2009 (UTC)


 * This edit, by Klose, removes unsourced text about djbdns' performance. Anastrophy, please realize that Gurnec only added a citation needed flag, he did not remove the text, that was another editor.


 * Why oh why does DJB cause so much strife in this world? There's always controversy surrounding him.  Anyway, my feeling is that on Wikipedia we should compare what we can compare reliably, and not compare things that are subject to debate.  The opening sentence currently states we will do a comparason of "features, platform support, and packaging", which i feel is about the limit of what we can do reliably. We should not compare security or performance of the software listed.  Firstly, it would probably be original research, which we can't do properly on wikipedia, so we don't try.  Secondly, the article would devolove into a terribly flamewar and nitpicking festival, something that happens all too often in technical discussions elsewhere.


 * Let us not talk about things we cannot reach a clear consensus on, and I think one of those things is performance, which is being brought up now, and also security. So that means that in general, we should remove text making performance or security claims, and only talk about things which are factual.  We may repeat claims of the makers of the software, but we should be careful we are not just advertising.  —fudoreaper (talk) 21:49, 6 May 2009 (UTC)


 * that's a fair and reasonable statement. thanks. i agree. Anastrophe (talk) 05:19, 8 May 2009 (UTC)

Dynamic (authoritative) DNS?
Isn't Dynamic DNS (RFC 2136, RFC 3007) a feature worth it's own column in the comparison table? Or is that too hairy and full of "yes, but..." ? Rootmoose (talk) 12:56, 29 April 2010 (UTC)

TreeWalk
TreeWalk has been around and still was getting recommendations from security advisors. I'm surprised it's not here. 97.85.185.160 (talk) 03:47, 31 May 2010 (UTC)

Creator
ISC did not "create" BIND as indicated in the Packaging table, Computer Systems Research Group did. To fix, I can see some options: Change entry to CSRG; change header to Maintainer (but not sure if that "breaks" any other entry); Change entry to CSRG/ISC; change to specify BIND9. --J Clear (talk) 16:19, 14 August 2010 (UTC)

pdnsd
Just a quick note that I’ve updated the comparison table, fixing (with a ref) the mistake claiming that pdnsd supports DNSSEC (while pdnsd now supports the record types used in dnssec, pdnsd doesn’t support DNSSEC yet; I added a reference for this), and have added pdnsd to the other two tables (pdnsd has been recently updated, is free [GPL], and IMHO deserves a row in all three tables) Samboy (talk) 18:34, 29 March 2013 (UTC)

Lack of Template:Proprietary / Template:Free and other essential information
I find the tables to be slightly difficult to read, with no licensing and other essential information such as programming language, latest version number, and initial release date. See Comparison of Internet Relay Chat daemons for a better comparison. 80.221.159.67 (talk) 02:52, 7 September 2016 (UTC)

DNS64
The feature matrix should include DNS64 support status. I think it would be very useful. Scineram (talk) 19:33, 22 February 2017 (UTC)

Extension Mechanisms for DNS (EDNS0)
The feature matrix should include EDNS0 support status. I think it would be very useful. grzchr15 (talk) 17:50, 21 March 2017 (UTC) https://tools.ietf.org/html/rfc7871 https://www.isc.org/docs/ECS.pdf https://dnsflagday.net/2020/


 * Do you know which of the DNS servers already support EDNS0? i.e. do enough already support it to add a column to the matrix? - Dyork (talk) 00:30, 22 March 2021 (UTC)

External links modified
Hello fellow Wikipedians,

I have just modified one external link on Comparison of DNS server software. 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/20080325113621/http://technet2.microsoft.com/windowsserver/en/library/6E5EF496-7411-44EE-9B83-578238B8EC241033.mspx to http://technet2.microsoft.com/WindowsServer/en/Library/6e5ef496-7411-44ee-9b83-578238b8ec241033.mspx

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) 16:20, 11 August 2017 (UTC)

Which operating systems do these all work on
I am very unimpressed that this article doesn't say which operating systems each DNS works on. Obviously, nobody gives a damn about DNS that won't work on their machine.

Rollback regarding Technitium DNS Server
Could you please explain this revert? You rollbacked the edit saying that it was an unexplained content removal, but I stated in my edit summary that it sources to GitHub, which is WP:USERGENERATED. There are 1000s of such open-sourced DNS servers on GitHub and not every single one of them are notable enough to be mentioned in Wikipedia. The Antediluvian (talk) 19:55, 5 April 2022 (UTC)
 * After I examined what this specific software is and the developer of this software, I undid my revert. I performed the revert in the first place because I recognized the first part of the name of the application. So Technitium is a company that is popularly known for its MAC spoofing application (called Technitium MAC Address Changer, or TMAC). TMAC is mentioned on PCWorld [here] and Digital Trends [here]. TMAC is more than 10 years old and is not posted on GitHub. I thought its DNS server was popular, but I stood corrected. I undid my revert after I found out that both Technitium and its DNS server were not very notable. — B. L. I. R. 00:31, 6 April 2022 (UTC)
 * After some searching I found it mentioned on ZDNet [here], on a ResearchGate paper [here], and also has several CVEs listed for it. 122.177.106.38 (talk) 08:36, 24 May 2023 (UTC)

Knot DNS and Knot Resolver
I think the two Knot lines should be merged. They are separate binaries, but written and distributed by the same source. The same is true for PowerDNS; their authoritative and recursive resolvers can be installed and run separately. (And BIND? And others?)

Combining the two Knot lines makes it look much more capable, and more inline with those that also serve as both authoritative and recursive resolvers, whether or not they can be run separately.