ietf
[Top] [All Lists]

Re: Question for the DNS system.

2004-03-10 17:29:21
Dear Eastlake, 

You are absoutly right, but if considering about the end-to-end SIP call, this 
lookup time will make some influence with the calling establish time, as 
Michael mentioned,  "The performance of lookups scales linearly  (worst case) 
with the number of zone cuts ".  

So that's the reason why I want to know the real-data about the lookup time in 
current DNS system. :-) It may be a problem for real-time applications.

Yours Felix.



----- Original Message ----- 
From: "Donald Eastlake 3rd" <dee3(_at_)torque(_dot_)pothole(_dot_)com>
To: "Felix, Zhang" <lionking(_at_)huawei(_dot_)com>
Cc: <ietf(_at_)ietf(_dot_)org>
Sent: Thursday, March 11, 2004 6:30 AM
Subject: Re: Question for the DNS system.


Deep (many label) DNS names work fine. But for some reason the DNS
system has long suffered from lust for the root where people scramble
for DNS names with the minimum number of labels. This, coupled with the
marketing efforts of some TLD owners resulting in very wide zones that
require huge servers, are harder to deploy DNSSEC on, etc. It's wide
zones that are a potential problem, not deep names.

(There is some logic in going for short domain names interms of total
number of characters or ease of memorization. But based on that,
1.2.3.example would be superior to onetwothree.example. Yet people lust
after the second...)

Thanks,
Donald
======================================================================
 Donald E. Eastlake 3rd                       
dee3(_at_)torque(_dot_)pothole(_dot_)com
 155 Beaver Street              +1-508-634-2066(h) +1-508-786-7554(w)
 Milford, MA 01757 USA                   
Donald(_dot_)Eastlake(_at_)motorola(_dot_)com

On Wed, 10 Mar 2004, Felix, Zhang wrote:

Date: Wed, 10 Mar 2004 13:02:36 +0800
From: "Felix, Zhang" <lionking(_at_)huawei(_dot_)com>
To: ietf(_at_)ietf(_dot_)org
Subject: Question for the DNS system.

Dear all,

     According to the current Internet, in most cases, the
allocation/design of DNS is not more than 3-5 levels, such as
us.ibm.com etc. What's my problems is that "when using lots of DNS
names with more than 5 levels, Is there some problem with the whole
DNS system, such as some performance problem for searching,
dificulities to operate etc.....?  Could the current DNS tree
architecure bear large traffic or not?"  Is there somebody can provide
some real data about the current Internet, such as average search time
etc.

     As everyone knows, IPv6 is coming ..........   :-)

 Best regards.

 Yours Felix. Zhang.



<Prev in Thread] Current Thread [Next in Thread>