Path: utzoo!utgpu!news-server.csri.toronto.edu!rutgers!uwm.edu!rpi!zaphod.mps.ohio-state.edu!usc!ucsd!ucbvax!mlb.semi.harris.com!del
From: del@mlb.semi.harris.com (Don Lewis)
Newsgroups: comp.protocols.tcp-ip.domains
Subject: Re: BIND bug list [Really:  client/resolver interface]
Message-ID: <9006062042.AA18811@thrush.mlb.semi.harris.com>
Date: 6 Jun 90 20:42:21 GMT
Sender: usenet@ucbvax.BERKELEY.EDU
Distribution: inet
Organization: The Internet
Lines: 22

On Jun 6,  2:47pm, Philippe Prindeville wrote:
} Subject: Re: BIND bug list [Really:  client/resolver interface]
> 	From: del@mlb.semi.harris.com (Don Lewis)
> 	} Subject: Re: BIND bug list [Really:  client/resolver interface]
> 	> 	If you use UDP to make queries, you can't distinguish between:
> 	> 		1) name server is dead
> 	> 		2) query packet or response packet was lost
> 	> 		3) name server is taking a long time to find the answer
> 	> 	
> Since cases 1 and 2 (above) are indeed problems, I will state again
> that VMTP would distinguish between 3 and the previous two.  So
> perhaps it is useful beyond the scope of recursive queries.

Case 3 doesn't really happen (unless the server is seriously overloaded)
with non-recursive queries since the name server has all the information
needed for the answer already squirreled away.  I would be more concerned
with the extra overhead associated with VMTP since timely responses will
be received for the vast majority of non-recursive queries.

Don "Truck" Lewis                      Harris Semiconductor
Internet:  del@mlb.semi.harris.com     PO Box 883   MS 62A-028
Phone:     (407) 729-5205              Melbourne, FL  32901