RE: DRAFT Nashua interim meeting minutes

["Mallikarjun C." <cbm@rose.hp.com>]

David,

>> There was discussion in the security portion of the agenda to the effect
>> that "if we mandate IPSec with it's data integrity check capabilities, is
>> there further need to specify error recovery"?
>
>That topic was picked up again in the error recovery session.  In essence,
>if ESP is in
>use, CRC failures at the iSCSI level will essentially never happen, allowing
>at least
>within-session and within-connection (and presumably) within-command
>recovery
>to be dropped, 

I want to add one comment here, though the right context to discuss
this would be while making a case for "CRC shim"/ESP-with-CRC.

Within-session recovery should continue to live on even with perfect
TCP, to transparently effect connection recovery on NIC failures.  The
Nashua error recovery slides explicitly make this point.  It however
may not have to be mandatory with a perfect/shored-up TCP.
--
Mallikarjun 


Mallikarjun Chadalapaka
Networked Storage Architecture
Network Storage Solutions Organization
MS 5668	Hewlett-Packard, Roseville.
cbm@rose.hp.com



>but the HMAC in ESP is computationally expensive leading to
>the
>interest in a CRC shim.  I've added explanation of this to the point in the
>error
>recovery session where the CRC shim was discussed.
>
>Thanks,
>--David
>
>---------------------------------------------------
>David L. Black, Senior Technologist
>EMC Corporation, 42 South St., Hopkinton, MA  01748
>+1 (508) 435-1000 x75140     FAX: +1 (508) 497-8500
>black_david@emc.com       Mobile: +1 (978) 394-7754
>---------------------------------------------------
>
>