gethostby* users

Les Mikesell lesmikesell at gmail.com
Sun Oct 14 18:20:58 UTC 2007


Ulrich Drepper wrote:

>> But, if they
>> resort the list, half of them will pick the wrong first choice even when
>> it is reachable.
> 
> No.  The sorting performed is not a universal and complete ordering.
> For once, it depends on the addresses of the client machines.  Second,
> if all target addresses are equally "bad" (i.e., for IPv4, all have the
> same matching prefix length)

I've not seen 'matching prefix length' to have anything to do with 
optimal route paths in practice other than the obvious case where one 
interface of a client is on the same subnet as one of the supplied IP's, 
which you can only determine by taking subnet masks into consideration.

 > the sorting will not change the order in
> which the entries are returned.  Hence the RR DNS will not lose its effect.

I don't object to putting same-subnet results first, but I'd want to see 
some real-world statistics before jumping to the conclusion that any 
other change is for the better.  I'd venture a guess that the majority 
of public DNS results are returned by 'intelligent' DNS servers that 
have already made some effort to give you the right choices in the right 
order.  Maybe we could at least ask someone at google if we need to 
second-guess their server results.

-- 
   Les Mikesell
    lesmikesell at gmail.com




More information about the fedora-devel-list mailing list