[saga-rg] Minutes of conference call 14/12/05

[Andre Merzky]

The streaming API was discussed on the SAGA list because the
set of people interested in that API are also participating
in the SAGA effort, and we felt that eventually it might be
a candidate for inclusion into SAGA.  It is not relevant to
SAGA _as_of_yet_, as we don't have any use cases submitted
to SAGA-WG.  

It looks similar to the stream API, but in fact cannot be
folded into it.  The point is, that the stream API operates
on byte level, and does not require any protocol on top of
the transport layer (tcp, udp etc).

The message API has additional semantic requirements, which
DO require a protocol between the endpoints on top of the
transport layer.  Hence, the message API could be
implemented on top of the streaming API, but cannot be
folded into it.

The optical switching stuff is only relevant as some use
cases deal additionally with that - the API itself is
independent from the network layer.

For now te messagung API should be handled similar as CPR
and RPC I think: not considered for the API right now, but a
good candidate for later inclusion (if we get use cases
submitted).

Hope that clears the confusion somewhat...

  Cheers, Andre



Quoting [Tom Goodale] (Dec 14 2005):
> 
> Present:
> 
> Shantenu Jha and Tom Goodale
> 
> We didn't discuss much - I asked Shantenu why a messaging API had been 
> discussed when we already have a streaming API which looks very similar, 
> and he said that he thought that this was actually a discussion about an 
> API for optical light path stuff and not actually supposed to be a SAGA 
> API (in which case why was it discussed on this list ?).  I suggested that 
> it looked close enough to the SAGA streaming API that perhaps we could 
> fold differences in and have a streaming API which would work for the 
> optical light paths.
> 
> We decided that it is probably not possible to have another call before 
> Christmas, so the next call will be sometime in the first half of January.
> 
> I would again urge everyone who hasn't already done so (i.e. people other 
> than Shantenu, Pascal and myself) to send constraints out so we can come 
> up with a day and time for calls which are not too excrutiating.  We may 
> even need to rotate the time by eight hours each call so that only one 
> time-zone has a midnight meeting.
> 
> Please send your constraints to the mailing list.
> 
> Cheers,
> 
> Tom



-- 
+-----------------------------------------------------------------+
| Andre Merzky                      | phon: +31 - 20 - 598 - 7759 |
| Vrije Universiteit Amsterdam (VU) | fax : +31 - 20 - 598 - 7653 |
| Dept. of Computer Science         | mail: merzky at cs.vu.nl       |
| De Boelelaan 1083a                | www:  http://www.merzky.net |
| 1081 HV Amsterdam, Netherlands    |                             |
+-----------------------------------------------------------------+