[6bone] Problems with big packets!?

[ZaferP@koc.net]

Guys, wyh don't we have some practical approach. Here is the ping output which completed sucessfully with 5000 datagram size. Although I just ping the other end of the IPv6overIPv4 tunnel there are many hops in between. I agree that MTU is bounded by the lowest link on the path however if the endpoints are capable of discovering this minimum and capable of fragmanting then there should be no problem. However sometimes ICMP packets that are destined for the source to tell him to reduce the packet size for a low MTU link can not reach there because of a firewall, access-list etc. In this case there will be no MTU discovery or fragmantation and packets simply will be dropped.

cam7507ler3#ping ipv6 
Target IPv6 address: 3FFE:2500:0:414::1
Repeat count [5]: 
Datagram size [100]: 5000
Timeout in seconds [2]: 
Extended commands? [no]: 
Type escape sequence to abort.
Sending 5, 5000-byte ICMP Echos to 3FFE:2500:0:414::1, timeout is 2 seconds:
!!!!!
Success rate is 100 percent (5/5), round-trip min/avg/max = 588/588/588 ms
cam7507ler3#

-----Original Message-----
From: Michel Py [mailto:michel@arneill-py.sacramento.ca.us]
Sent: Saturday, October 05, 2002 6:10 AM
To: 6bone@mailman.isi.edu
Cc: andy@ipng.org.uk
Subject: Re: [6bone] Problems with big packets!?


>> John Fraizer wrote:
>> Um, What is the problem?
>> You're _NEVER_ going to encounter a 1000 byte
>> packet, let alone a 5000 byte packet in "the wild".
 
> Andy Furnell wrote:
> I disagree,
> I regularly see packets of 1000 bytes and above on
> applications such as NNTP and FTP. Pinging with large
> packet sizes is often the best way to determine faults
> with various networking media. Although in a tunneled
> IPv6 context (especially given the size of the IPv6 header)
> you're likely to see some pretty funky things happen when
> pinging with packets much larger than 1200 bytes.

Agree with Andy.

Michel.

_______________________________________________
6bone mailing list
6bone@mailman.isi.edu
http://mailman.isi.edu/mailman/listinfo/6bone