[governance] India's communications minister - root server misunderstanding (still...)

[Roland Perry]

In message <5020A4F9.7030401 at itforchange.net>, at 10:47:45 on Tue, 7 Aug 
2012, parminder <parminder at itforchange.net> writes

>"As already mentioned, there are hundreds of root server instances. 
>Each of these is an actual root server."
>
>Isnt this statement as or more untrue, in a discussion where we are 
>mainly speaking about actual 'control' over the root file. The hundreds 
>of root servers mentioned above are NOT 'actual root servers'.

In the quotation above, "an actual root server" means a box which has 
been deployed to a physical location. The purpose of the mirrors (or 
"instances" as they are generally called) is to provide a server which 
is closer to its users and therefore [generally] more efficient and 
resilient in terms of Internet connectivity.

'Who provides the data for the servers' is indeed a different layer.

I'm sure it would help if these two concepts were better understood to 
be separate; I'm working on that, as a simple factual exercise.

>An actual root server is a shorthand for an actual root server 
>operator, who exercises control (at least potentially) over the root 
>zone file that he publishes

Just as potentially, every ISP whose own DNS servers are queried by 
their customers could also exercise control over the results which are 
given. Sometimes that's by accident, but occasionally it's by design.

It works like this:

When I want my ISP to fetch some data from igcaucus.org, the first step 
is to ask my ISP to translate igcaucus.org into an IP address.

The only role the root server has is to periodically remind my ISP of 
the location of .org's name servers/zone files - because the ISP caches 
the result. So that's the first stage completed: getting a list of 
.org's name servers.

While knowing where the .org name servers are is fundamental, the second 
stage is much more important on a day to day basis: querying one of 
those .org name servers to find out where to obtain information on the 
location of igcaucus's server.

And what the .org name server tells us is: where to find a name server 
run by the organisation chosen by igcausus to 'host' their domain[1].

And finally, that third name server delivers us the required IP address.

 > [the Indian Minister] quite wrongly said that Internet traffic flows 
 >through 13 root servers (he should have said, internet traffic, in a 
 >way, gets directed by 13 root servers)

No. All that the root servers do is inform the world where to find the 
name servers[2] for each TLD.

And even those TLD name servers don't have a role in directing traffic, 
they just help tell us where the end point is. A whole different set of 
technology [referred to as "Routing"] decides how to get from here to 
there, principally using a protocol called BGP.

ps Remember the caching of the root zone file by the ISP? Well, all the 
other queries are cached as well. Which explains the delay when you move 
the hosting of a domain. It usually takes a while (often two days) for 
all the caching to expire, and for users to be directed to the new end 
point.

[1] For those who are still with me, that information for igcaucus
     happens to be kept by Tucows, in a record administered by Avri.

[2] Which host the zone files for...
-- 
Roland Perry

-------------- next part --------------
____________________________________________________________
You received this message as a subscriber on the list:
     governance at lists.igcaucus.org
To be removed from the list, visit:
     http://www.igcaucus.org/unsubscribing

For all other list information and functions, see:
     http://lists.igcaucus.org/info/governance
To edit your profile and to find the IGC's charter, see:
     http://www.igcaucus.org/

Translate this email: http://translate.google.com/translate_t