ACK of purge packets

gardo@vnet.ibm.com Thu, 16 November 1995 18:47 UTC

Sender: ietf-archive-request@IETF.CNRI.Reston.VA.US
From: gardo@vnet.ibm.com
Message-Id: <199511161813.NAA20208@guelah.nexen.com>
Date: Thu, 16 Nov 95 13:03:35 EST
To: gray@ctron.com
cc: rolc@nexen.com, genecox@vnet.ibm.com
Subject: ACK of purge packets
Precedence: bulk
Ref: Your note of Thu, 16 Nov 1995 13:04:29 -0500

Eric,

>> Is there a need or requirement that some clients not participate in
>> the purge portion of NHRP?
>>
>
>Interesting question.  Do you suspect that I ask to dispense with
>acknowledgement so that I can ignore the protocol?

I suspect some clients would rather not receive a purge.
A server must keep state associated with each
server in order to perform a purge.  If a client informed the server
that it does not want to receive a Purge, then the server could drop
any state information that he would normally keep for that client for
the purge protocol.  A server may set different holding times in Replies
to clients that do not support the Purge versus those that do support
the Purge.

What do you think?

This was just a thought...  However, if we did such a think, maybe this
would solve the disagreements associated with purge acknowledgements and
the purge mask.

>
>  Eric Gray
>

Have a nice day!
-- Russell