=?euc-kr?B?W8D8w7zIuL3FXSBSZTogW21vYmlsZS1pcF0gTW9iaWxlSVB2NiBW?==?euc-kr?B?UyBETlM=?=

[paul@etri.re.kr]

Title: [Àüüȸ½Å] Re: [mobile-ip] MobileIPv6 VS DNS

inline..

From: Warodom Werapun[ple@graduate.kmitl.ac.th]
To: Jaehoon Jeong
Cc:mobile-ip@sunroof.eng.sun.com; DNSOP WG
Subject: Re: [mobile-ip] MobileIPv6 VS DNS
Date: 2003/03/15 Sat 02:15

If we have change DNS Server IP with Anycast IP address.[in the future]
Then, MN doesn't need to configure any DNS IP address.
==> In the past, IPv6 wg tried to use anycast for DNS server discovery,
==> but because anycast has some problems, they gave up that approach.
==> Please look at the draft, http://www.ietf.org/internet-drafts/draft-ietf-ipv6-dns-discovery-07.txt,
==>to see the problem of anycast. The related paragraph is as follows;
-------------------------------------------------------------------------------------------------------------
...
  On top of the classic problems (TCP sessions, ICMP messages,...) using
   an anycast address would hide the real locations of the recursive DNS
   servers to the stub resolver, prohibiting it to keep track of which
   servers are performing correctly.
...
-------------------------------------------------------------------------------------------------------------
==> This draft suggests a new approach in use of well-known site-local addresses for recursive DNS server.

But one problem remain, some networks can't use same network prefix.
   
How do you think about this solution?
==> I think the approach of site-local address is better than that of anycast.
==> But the limitation of site-local address approach is that all the networks should run recursive DNS server with
==> well-known site-local address(es) for recursive DNS server(s) in the managed network.
==> Thanks for your suggestion.

==> /Jaehoon
Regards,
- Warodom

>[mobile-ip] MobileIPv6 VS DNSHello, I am the author of this draft.
>http://www.ietf.org/internet-drafts/draft-jeong-hmipv6-dns-optimization-00.txt
>(Title : The Autoconfiguration of Recursive DNS Server and
> the Optimization of DNS Name Resolution in Hierarchical Mobile IPv6)
>This draft is one of RA-based DNS discovery solutions.
>
>  
>
>>----- Original Message -----
>>From: Warodom Werapun
>>To: mobile-ip@sunroof.eng.sun.com
>>Sent: Monday, March 10, 2003 7:19 PM
>>Subject: [mobile-ip] MobileIPv6 VS DNS
>>    
>>
>
>  
>
>>Following this link:
>>http://www.ietf.org/internet-drafts/draft-jeong-hmipv6-dns-optimization-00.txt
>>When mobile node move, we only map to new DNS server.
>>    
>>
>
>  When mobile node moves into another MAP domain in hierarchical mobile IPv6 network,
>  it needs to find a new recursive DNS server  for the optimization of DNS resolution process
>  in the respect of name resolution time.
>
>  
>
>>We can reach mobile node in new location by using name in DNS.
>>    
>>
>
>  Yes, your're right. But when a mobile node tries to connect to an Internet service, such as web server,
>  first of all, it should resolve the service's DNS name into IP address.
>  In current DNS operation, the mobile node uses a fixed recursive DNS server normally placed in its home network.
>  If the mobile node can find a new recursive DNS server in the visited network,
>  it can use the new one for DNS name resolution instead of the previous one.
>  In the result, the time for name resolution can be saved. This is the main object of this draft.
>  I think the idea of the draft can be applied to IPv6 or mobile IPv6 network as well as hierarchical mobile IPv6 network.

>
>  
>
>>So, we don't need to set up any Home agent, are we?
>>/Warodom
>>Ps: I'm so newbies in MIPv6.
>> Thank you for your reply.
>>    
>>
>
>Thanks for your question.
>
>Regards,
>Jaehoon
>
>  
>