RE: WARP RDMA Architectural Requirements Summary posted

["Douglas Otis" <dotis@sanlight.net>]

Costa,

Would a third-party proxy completion of a command represent the same
functionality as that found using VI for SCSI?  In other words, allow a
separate device to act on behalf of another with respect to exchanging data.
An allowable type of spoofing to assist collection of data.  SCSI already
provides for content directed placement, just not by a third-party as does
VI.  It would seem a field added to the present structures to identify the
child device together with the parent solves this issue.

I have already subscribed to this reflector months ago with no activity.  Do
I need to re-subscribe?

Doug



> A group of us have been working on an RDMA system called WARP. The
> RDMA layer aims to provide exact data placement for iSCSI, NFS, CIFS,
> as well as the VI interface. The RDMA layer is implemented on a generic
> network adapter which does not have to have specific upper-layer
> protocols
> implemented on it (like iSCSI or NFS).
>
> The requirements and a sketch of the design is now available at
> http://www.stanford.edu/~csapuntz/draft-jpink-warp-summary-00.txt.  The
> document has been sent in as an internet draft and should be available
> as draft-jpink-warp-summary-00.txt by the end of next week.
>
> The mailing list rdma@cisco.com has been set up for discussion on the
> topic of RDMA and WARP RDMA. Please use that list and NOT the ips
> mailing list for discussion and questions.
>
> You can subscribe to the rdma list by sending mail to
> "mailer@cisco.com" with a blank subject line and the body of the
> message "subscribe rdma".
>
> Thanks,
> Costa
>
>