Really odd one: parts of global DNS just dropped off the map

Jonathan de Boyne Pollard J.deBoynePollard at Tesco.NET
Sat Dec 11 12:42:24 UTC 2004

KD> Also, I think the hierarchical nature of the *namespace* confuses
KD> some people into thinking that the *resolution*mechanism* should be
KD> hierarchical too, even though the two don't really have any direct
KD> relationship with each other.

It confuses some people into thinking that the resolution mechanism *is* 
hierarchical and *does* exactly parallel the namespace.  It's all too 
commonly explained that way, too.  When I first learned about DNS, I 
laboured under the burden of someone explaining the operation of the DNS 
in just such an incorrect manner.  Here's an example of such a poor 
explanation, that I came across a while back:

    j> If a named process cannot resolve an address locally it will call 
    j> higher authority. Ultimately it will attempt to contact the 
system that
    j> is authoritative for the zone in question, however, unless the 
    j> is cached or in the hosts local files then it will not know the 
    j> of the authoritative server. This problem is resolved by *recursive
    j> resolution* of requests, i.e. *any DNS server will pass requests 
it cannot
    j> handle to a higher level server and so on* until either the 
request can
    j> be handled (either by sending a message to the identified 
    j> host) or *until the root of the DNS name space* is reached.
        -- <URL:>

The problem appears to be caused by the word "recursion".  People seem 
to be unclear on the point that once one has followed the (possibly 
zero-length) chain of forwarding proxy DNS servers, from the original 
DNS Client, to reach the resolving proxy DNS server (that is actually 
going to perform the query resolution), the recursion depth at that 
point is always exactly 1.

More information about the bind-users mailing list