Talk:Root name server

From Wikipedia, the free encyclopedia
Jump to: navigation, search
WikiProject Computing / Networking (Rated C-class, Mid-importance)
WikiProject icon This article is within the scope of WikiProject Computing, a collaborative effort to improve the coverage of computers, computing, and information technology on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.
C-Class article C  This article has been rated as C-Class on the project's quality scale.
 Mid  This article has been rated as Mid-importance on the project's importance scale.
Taskforce icon
This article is supported by Networking task force (marked as High-importance).
WikiProject Internet (Rated C-class, High-importance)
WikiProject icon This article is within the scope of WikiProject Internet, a collaborative effort to improve the coverage of the Internet on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.
C-Class article C  This article has been rated as C-Class on the project's quality scale.
 High  This article has been rated as High-importance on the project's importance scale.

Empty string[edit]

re "The empty string after the final dot is called the root domain, and all other domains (i.e. .com, .org, .net, .uk, etc.) are contained within the root domain." I'm confused here: .com or .org are contained within an empty string ? Jerome Potts 04:15, 28 November 2005 (UTC)

Reply to above and rebuttal:

Rootaccess -- few issues to be determined and understood for general user purpose.

1) who has the authroity to determine who gets the top level global domain access -- is it money or influence? 2) who determines the global maximum limit at top level domain and why is there a limit? — Preceding unsigned comment added by Vdhananjay (talkcontribs) 14:57, 8 November 2013 (UTC)

An example:

1. The dot between google and com is actually the root domain itself. This dot points the server to a top level domain (ie. com, .org etc...). 2. The word "google" is the Sub-domain

And so forth. Also this.

All fully qualified domain names (FQDNs) on the Internet can be regarded as ending with this empty string for the root domain, and therefore ending in a full stop character (the label delimiter), e.g.,

This is all bullshit. Citation needed or else remove that utter crap. Who do you think you are? Resolution doesn't go directly to a root-server anyway.

Level 3[edit]

Why isnt Level3 Communications mentioned here? Don't they currently operate a few rootservers? - Scott110

I looked up "level 3 Communications". on this website. You will find out some information there.

-Tate Martinez

Level(3) does not operate a root server. The list represented on the page is accurate. See to confirm for yourself. The list of operators is up to date.

--Hannigan 17:26, 14 February 2007 (UTC)

Expand please[edit]

Please expand this article, this is a pretty fundamental Internet topic. —The preceding unsigned comment was added by MMuzammils (talkcontribs) 13:27, 13 February 2007 (UTC).

Feel free to edit at any time. I also left you a comment on your talk page regarding the banner placement related to cites and references. I think that the article is doing this properly, and in the case of some of the edits, you have subject matter experts making these edits and it is not within community standards to cite yourself.

How are Root Nameservers Updated[edit]

What is the process by which the root name servers are updated? Are they updated by domain registrars? Or is this a function of an authoritative DNS server to update the root server? This is not clear and should be clarified within this article.

I guess the whole "this is not a discussion forum" for the article doesn't apply to some peoples discussions of how to improve articles:) Woods01 (talk) 12:04, 20 February 2011 (UTC)

The root servers publish the ICANN root zone. So ICANN updates the root zone, and then publishes that to the root server operators by some private mechanism, which is not shared with the public. The root server operator then deploys that data on their servers, each in their own way. Looking at the article now, it doesn't seem like we really talk about data distribution, or where the root zone file originates. As mentioned the distribution mechanism is not well established, from what I know, but the idea that it originates from ICANN is quite clear. This should be talked about in the root zone file section, i think. —fudoreaper (talk) 03:55, 30 January 2013 (UTC)

Attempt at reference fix[edit]

External links contain almost all of the required references so I moved them to references. I will add pointers as time permits to improve readability. Suggestion on formatting welcome.

—The preceding unsigned comment was added by Hannigan (talkcontribs) 19:15, 20 February 2007 (UTC). --Hannigan 17:26, 14 February 2007 (UTC)

internet backbone[edit]

"As the root nameservers function as the Internet backbone, ..."

text like this leads to the widespread misconception that all Internet traffic passes through the DNS root name servers. I suggest saying something like

"As the root name servers perform a highly visible function, ..."

Dkarrenberg 12:13, 9 March 2007 (UTC)

Regarding that attacks on the root servers are called "DNS Backbone DDoS Attacks", I think it is best described as:

"As the root nameservers function as the DNS backbone, ..."

Mgoerner 22:13, 30 March 2007 (UTC)


Aren't the IP-adresses of the rootservers pretty much hardcoded (e.g. a DNS has their IP-Adresses written directly into their lookup table)? If so, this unchanging information might be interesting to add to the server table - no? Peter S. 19:01, 9 March 2007 (UTC)

I don't know when, but your idea was incorporated, we have the IP addresses of the root servers in the table now. You are right that changes to these addresses are infrequent, so it should be easy to maintain. —fudoreaper (talk) 03:57, 30 January 2013 (UTC)

Hypothetical Situation[edit]

This is pretty badly written (it took me a while to figure out what situation it was talking about - since it is its own section, it shouldn't start on a pronoun), and is clearly undocumented. I think that some mention of this 2%/hour failure estimation might be worthwhile, but it needs a reference, shouldn't be in a separate section, and should be cleaned up. If nobody can produce a reference for it soon, I'm in favor of taking it out. 12:37, 13 August 2007 (UTC)

Agreed. This is pretty badly written. There is no hypothesis! I presume that this refers to the simultaneous failure of all root servers? Maybe we can just add a sentence to that effect? --WibblyLeMoende 14:24, 14 August 2007 (UTC)

Incorrect Statement[edit]

"For example, queries with the source address (corresponding to anywhere and everywhere) make it to the root servers."

Hmm. As a root operator for almost 10 years now, and probably more obsessive about root server traffic than just about anybody else I know, I can't say that this is actually true. And that wasn't mentioned in RFC 4697 (which I co-authored)

This sentence is not totally without hope, there are a couple of changes to the above statement which could make it true.

Queries for reverse DNS lookups that are not correctly parsed end up hitting the majority of the root servers (i.e. wasn't looked up, instead, the query for "" was looked up instead. ) We did mention that in RFC 4697 (See section 2.9)

We get a lot of RFC 1918 sourced traffic at the root servers that is dropped.

(Just to be sure I ran a tcpdump on since I started writing this post and I haven't seen any queries with a source address of

I didn't make any changes to the page because I'm too shy. And it would look too much like original research. But mostly the shy part.

--Pietbarber (talk) 22:37, 4 February 2008 (UTC)

Please be WP:Bold and fix things. I have rewritten that particular section and added a citation to an older article that I knew of. I couldn't find anything about source IP addresses of so I deleted that part. I have not had a chance to do more than skim your RFC, but it looks useful. The Measurement Factory article didn't give solutions to the problems the found. Wrs1864 (talk) 12:13, 5 February 2008 (UTC)

US Centric?[edit]

The list of the thirteen rootservers seems to be heavily US-centric. Is this actually the case?LorenzoB (talk) 15:47, 26 June 2008 (UTC)

Yes, that's the way it is, because the Internet was still largely a US thing at the time the DNS was introduced. But many of the root nameservers now consist of multiple machines distributed around the world, so it's not as US-centric as it once was. --Zundark (talk) 17:20, 26 June 2008 (UTC)
Just to add a bit of subtlety: the root server operators are mostly from the US, but the servers themselves are now spread out globally, thanks to anycast, and responsible operators. Additionally, note that the US root server operators are a diverse group: the US military, US space agency, a domain registration company, an ISP, two universities, and a non-profit. They must all collaborate for the system to work, and their diverse interests is one reason for the stability of the root server system, for that is one of their common goals. —fudoreaper (2009-09-08T23:27:58)

According to the article the US Government department of commerce has to approve anything that's done that would affect operation of the root-server system. This is the first i've heard of this but if it's correct that would make it US centric regardless of where the servers are. Woods01 (talk) 12:07, 20 February 2011 (UTC)

Picture of K root server[edit]

The picture in this article claims to be of the K-Root server at AMS-IX, but upon looking closer, it's just a Cisco router and a switch. I think the picture depicts the router for the K-Root server but not the K-Root server itself? U-238 (talk) 13:23, 26 November 2008 (UTC)

Name server software[edit]

In the chart of nameservers, we list the software for each of the servers. Is there a reference for this? There is nothing cited on the page currently, but does someone know of a good source?

Using the unix command "dig CH TXT version.bind." queries the server directly for their version number, here are the results:

  • "This space intentionally left blank"
  • "9.3.2"
  • "9.4.3-P3"
  • "9.6.1-P1"
  • "9.2.3"
  • "9.5.1-P3"
  • ""
  • "NSD 3.2.3"
  • "contact"
  • "VGRS4"
  • "NSD 2.3.7"
  • "NSD 3.2.2"
  • "9.4.3-P3"

From this, A, G and I do not tell us anything, and J says 'VGRS4'. It is quite possible that verisign operates their own, custom DNS software, which is what this suggests.

B, C. D, E, F, and M all appear to be running BIND, those versions are recognizable as such. However, i cannot rely on the information provided, as it is easy to set this to an arbitrary value, which does not match the actual software version.

Overall, then, I am concerned about the factual accuracy of the software column in the server list table. Can anyone add more information? —fudoreaper (talk) 05:22, 9 September 2009 (UTC)

Replying to myself: says that they run NSD, and so does ISC says they are running BIND9, which is logical since the ISC writes BIND. I will add these refs to the chart. —fudoreaper (talk) 05:47, 9 September 2009 (UTC)
Has m switched to NSD as well ? From AS 3215, I get the following results. Note also B.
  • "This space intentionally left blank"
  • "4.8.1"
  • "9.6.2-P2"
  • "9.6.1-P1"
  • "9.6.1-P3"
  • "9.6.2"
  • ""
  • "NSD 3.2.5"
  • "contact"
  • "This space intentionally left blank"
  • "NSD 3.2.4"
  • "NSD 3.2.4"
  • "NSD 3.2.4"
Mro (talk) 06:09, 15 July 2010 (UTC)

Dnsstuff link/promotion[edit]

Can the usefulness of the dnsstuff nameserver response times be weighed since dnsstuff does not provide any information as to how it gathers such times nor how often the times are actually updated. Woods01 (talk) 12:04, 20 February 2011 (UTC)

.Com (or .edu, .org, .gov, etc.) being the root domain is incorrect[edit]

I am currently taking a class called "Intro to Networking" which is said to prepare me for the CompTIA Network+ Exam. According to my study materials (more specifically, TestOut LabSim version 3.1.57 with add-on "Network+ N10-004"): "

When you use the host name of a computer (for example if you type a URL such as, your computer uses the following process to find the IP address.
1. The host looks in its local cache to see if it has recently resolved the host name.
2. If the information is not in the cache, it checks the Hosts file. The Hosts file is a static text file that contains hostname-to-IP address mappings.
3. If the IP address is not found, the host contacts its preferred DNS server. If the preferred DNS server can't be contacted, it continues contacting additional DNS servers until one responds.
4. The host sends the name information to the DNS server. The DNS server then checks its cache and Hosts file. If the information is not found, the DNS server checks any zone files that it holds for the requested name.
5. If the DNS server can't find the name in its zones, it forwards the request to a root zone name server. This server returns the IP address of a DNS server that has information for the corresponding top-level domain (such as .com).
6. The first DNS server then requests the information from the top-level domain server. This server returns the address of a DNS server with the information for the next highest domain. This process continues until a DNS server is contacted that holds the necessary information.
7. The DNS server places the information in its cache and returns the IP address to the client host. The client host also places the information in its cache and uses the IP address to contact the desired destination device. " (End of quote)

So, you see the computer host doesn't actually check the root server first for the information needed to resolve an address even though the root server is first in line in the web address. The root server is actually the omitted dot or period at the end of a web address. For example, according to my study materials, the technical full name of a given web address is "" (without the quotes)(notice the period at the end of the web address, that is the root server(how it differs from a root server website, like the ones mentioned, I can't really figure out)and the even more full technical name of that given web address is "" (without the quotes)(notice again the period after the ".com", that being again the root server) if that particular web site/address uses the HTTP protocol. Do you understand? Please tell if you don't understand and/or what you need clarification on, if any.

So you take an INTRO class at the community college and come here to teach all about networking. Come back in a few years kid when you get some classes above the beginner level and a few years of experience working in the real world. ~~ThanksKid — Preceding unsigned comment added by (talk) 07:21, 21 March 2014 (UTC)

Potter76 (talk) 22:02, 12 October 2011 (UTC)

Hi Potter76. Your explanation of DNS resolution is basically correct, though there's a few details I would nitpick. You seem to be objecting to some content in the article that says ".com is a part of the root zone". I'm not sure what text on the article is incorrect, can you point it out clearly, if it still exists? Also, com. is in the root zone, as are all the other top-level domains. So you may be objecting to something that's actually correct. Hopefully you notice this talk page and can respond. Cheers —fudoreaper (talk) 04:10, 30 January 2013 (UTC)

Misleading coverage of canonical root server names[edit]

the article as it is written now misleaded me to do this false edit.. maybe it should be worth mentioning what is and who maintains this website (talk) 06:56, 27 January 2013 (UTC)

Hi. Yes, that edit wasn't correct. We do link to in the external links section, but I'm not sure how much we have to say about that site in the article. The site is very simple, I doubt there's a lot of information out there about it. What should we write about it, do you think? —fudoreaper (talk) 16:42, 28 January 2013 (UTC)
Most importantly whether has any official relation to ICANN or IANA, or not. In the latter case you're right, it isn't worth mentioning.
I was confused because the article has serveral links to this page (not only in the external links section, but also in the leftmost column of that table) leaving the impression that this page ( serves some official function (e.g. like the "official homepage" of the root servers, if such can exist). -- (talk) 13:17, 31 January 2013 (UTC)
The officialness of is hard to determine. It seems to be run by the root-servers operators association, but that's a fairly loose organization, i think. Nothing really official about it. So I'm not sure the answer to your question. —fudoreaper (talk) 19:54, 1 February 2013 (UTC)

Difference between reported number of TLDs[edit]

As of now there are 1038 TLDs listed in and 1058 listed in I'm going with the one listing more. On a quick comparison the difference is BL, BQ, EH, LIXIL, MF, SS, TP, UM, and the rest are XN--*. M.ollivander (talk) 14:25, 30 July 2015 (UTC)

In List of Internet top-level domains this URL is referenced for the number of TLDs: which is reporting 1034. Now I'm confused M.ollivander (talk) 15:09, 30 July 2015 (UTC)

Number of authoritative servers for TLDs[edit]

I don't know how to find this number. I kept the one from 2004 in my recent edit. If you know how to find this please feel free to edit the article M.ollivander (talk) 14:27, 30 July 2015 (UTC)