[storm] iSER update request from Alex Nezhinsky on iSER Hello/Reply messages

[Mike Ko <[e-mail]>]

Alex sent me the following iSER update request:

> No implemented ISER Hello - ISER Reply messages.
>
> Obviously there is no substantial solution on the spec level.
>
> In practical terms, it is necessary to push iSER-Hello implementation,
> in a way which is graceful to the legacy initiators. This is quite
> problematic, though. For example, if no ISER-Hello message is
> received, the target may ignore the fact and initialize the associated
> values with some default, legacy-oriented ones. The reverse situation
> is harder to handle, because targets not prepared to receive iSER
> Hello message may behave in a non-deterministic way upon such receipt.
> It should be noted that it’s more difficult to replace initiators than
> targets in the field, as they are numerous and less controlled, thus a
> “half-painful” practical solution may be achieved.
>
> There is no much sense in stipulating formally in the spec, that such
> behavior MUST NOT be allowed for iSER initiators and targets after a
> certain spec version. The problem with this is of course the fact that
> the version is communicated through the ISER Hello message itself,
> when the initiator is non-compliant. And, there is actually nothing to
> do at all, when a non-compliant target meets a compliant initiator, as
> described above.

Are iSER Hello/Reply messages implemented in practise?

Mike