The IP over ATM Mailing List Archive by date

[gja@thumper.bellcore.com]

>>Is it true to say that the only way to build a robust client is
>>to have it issue an ATMARP request (for an arbitrary address) 
>>when it first connects to the ARP server and at regular intervals 
>>after that to ensure the information is refreshed ?

This would appear to be a reasonable assertion.

	[..]
>>If the ATMARP requests are necessary then the InATMARP
>>requests appear redundant for SVCs.

I believe Mark noted the possible redundancy in the midst of the email
avalanche in august or september. In answer to Uttam Shikarpur's
concern, I don't see that the InARP sequence is any more secure
than extracting info from the ARP packet. Any host that has been
compromised to the point of allowing falsified ARP Requests to be
generated can certainly arrange for false InARP replies too. The
ARP Server will then end up with trusted but false InARP derived
information.

gja