Forums before death by AOL, social media and spammers... "We can't have nice things"
|    comp.protocols.tcp-ip    |    TCP and IP network protocols.    |    14,669 messages    |
[   << oldest   |   < older   |   list   |   newer >   |   newest >>   ]
|    Message 14,033 of 14,669    |
|    Rick Jones to rahul.dev.agg@gmail.com    |
|    Re: icmp echo to a host with smaller mtu    |
|    14 Jun 13 17:44:55    |
      From: rick.jones2@hp.com              rahul.dev.agg@gmail.com wrote:       > One more question:              > For directly connected hosts, what should be the PMTU ? Should the       > sender always assign pmtu = interface mtu of sender?              > Or, should it send an ICMP echo starting with its interface mtu. If       > it gets the reply, assign pmtu = interface mtu. Else, decrease the       > icmp echo size and try again till it gets the reply and use that       > value as pmtu ?              Given I start from the premise of "Every station in a given broadcast       domain MUST have the same MTU (framesize)" sending the ICMP Echo       Request is (should be) unnecessary. And I wouldn't suggest it.              While the conditions are not identical, there was an amplification       attack possible against some TCP stacks (one of which used to be near       and dear to my paycheck) which did something similar with non-local       destinations. When speaking with the remote destination, they would       send an ICMP Echo Request with the DF bit set in the IP header, while       the rest of the traffic being sent to the destination was sent with       the DF bit cleared. The idea was to try to see if there was effective       PathMTU discovery possible along the path to the remote. Only once       there was an ICMP Echo Reply recieved would the DF bit start being set       on the "real" traffic. A rather clever thing to do, save for one       thing...              Trouble was, someone sending say TCP SYNchronize segments with a       spoofed source IP address would get the receiving stack to send a full       local MTU-sized ICMP Echo request to the spoofed source IP along with       the TCP SYN|ACK.              rick jones       --       The computing industry isn't as much a game of "Follow The Leader" as       it is one of "Ring Around the Rosy" or perhaps "Duck Duck Goose."        - Rick Jones       these opinions are mine, all mine; HP might not want them anyway... :)       feel free to post, OR email to rick.jones2 in hp.com but NOT BOTH...              --- SoupGate-Win32 v1.05        * Origin: you cannot sedate... all the things you hate (1:229/2)    |
[   << oldest   |   < older   |   list   |   newer >   |   newest >>   ]
(c) 1994, bbs@darkrealms.ca