BIND 9.2.1 and TCP

[zack.nash at amd.com]

	If I understand you correctly the reason we are seeing the TCP request is because NS records are being returned for every query on top of the query.  Why do we not see this occur with bind 8?
	I appreciate you indicating this issue but these are internal only name servers and the world does not see them.
Thanks,
Zack

-----Original Message-----
From: Jim Reid [mailto:jim at rfc1035.com]
Sent: Thursday, November 07, 2002 10:23 AM
To: Nash, Zack
Cc: Mark_Andrews at isc.org; bind-users at isc.org
Subject: Re: BIND 9.2.1 and TCP


Thanks for the illustration of DNS truncation. I wonder why amd.com
feels it's necessary to advertise 22 NR records for their zone. The
root and com zones get along just fine with only 13. And BTW, those
zones/servers exploit a trick with label truncation to fit the 13 NS
records and their associated A records into < 512 bytes so they don't
hand out a truncated response.