Re: New List: rdma@cisco.com: to discuss RDMA

[Matt Wakeley <matt_wakeley@agilent.com>]

meth@il.ibm.com wrote:

> I think we should specify an RDMA mechanism for use with iSCSI, as this
> will help all interested parties to have interoperable RDMA mechanisms. I
> don't think RDMA should be mandated for iSCSI; I assume there will be
> implemenations on hosts and devices that won't have the resources to
> perform RDMA.
>
> - Kalman Meth

iSCSI does not need an RDMA mechanism because (from David Black's message):

> > Actually, RDMA is not needed in FCP because all protocol chips
> > implemented perform a real peer-to-peer DMA straight to the
> > data areas specified by the user's interaction with the operating
> > systems allocation algorithms.  The combination of the FCP/SCSI
> > pointer structure, task tagging, and the FC relative offset perform the
> > function you would otherwise have to use RDMA to accomplish.
>
> And this illuminates the design tradeoff that may motivate RDMA.  If
> one only wants to accelerate one protocol (SCSI/FCP in the above
> example) then having hardware understand its headers and doing
> the DMA on that basis is a fairly obvious way to go - HBAs for both
> parallel SCSI and Fibre Channel (SCSI/FCP) do this.

The issue is that TCP does not provide message framing such that the out of
order messages could be found and processed.

Besides, how can you "specify an RDMA mechanism for use with iSCSI" if the
mechanism in TCP has not been defined, much less standardized?

-Matt