Re: FCIP/iFCP : Guarantee In-Order delivery for FC N/NL_ports

[Santosh Rao <santoshr@cup.hp.com>]

Y P Cheng wrote:

> > It is not only the desire to inter-operate with legacy FC devices
> > that drives this requirement.
> > This is a mandatory requirement, since FC-FLA Table 3 mandates
> > that N/NL_Ports originating FLOGI SHALL request in-order delivery in
> > the FLOGI.  (by setting the "Sequential Delivery" bit in their class
> > specific parameters for class 2 & 3 to 1.).
> > - Santosh
>
> If in-order-delivery is a MUST, the next question is "How many open
> sequences?"

In-order is a requirement that spans across exchanges. (not just within an
exchange).

> On a 10 Gb/sec network with long latency, we need many open
> sequences to fill the pipeline.  This is the same problem as TCP window size
> of iSCSI.  To keep all open sequences in order, in the worst case, there
> will be a huge holding buffer in an iFCP and FCIP device.  We have just
> moved the TCP window problem from an end-point into a bridge or gateway
> device.

Agreed. The alternative is to use 1 TCP connection per FC (S_ID, D_ID) pair.
Which is why I raised the question in the context of multiple TCP connections
between the edge switch/routers per (S_ID, D_ID) pair.


> Hence, by just changing the protocol, the problem does not go away.

Not sure what you mean here. Changing which protocols ?

Regards,
Santosh
begin:vcard 
n:Rao;Santosh 
tel;work:408-447-3751
x-mozilla-html:FALSE
org:Hewlett Packard, Cupertino.;SISL
adr:;;19420, Homestead Road, M\S 43LN,	;Cupertino.;CA.;95014.;USA.
version:2.1
email;internet:santoshr@cup.hp.com
title:Software Design Engineer
x-mozilla-cpt:;21088
fn:Santosh Rao
end:vcard