[nm-wg] last and next phone call

[Leese, MJ (Mark)]

Hi folks,

There was a problem with the phone system last week, and as a result we had just five people on the call. We discussed some things (brief summary below) but decided it was better to rearrange the call. To give some notice, the next call will be Tuesday 5th April, at the usual time. Eric or Susan will send out the details.

We were going to propose holding the call one hour earlier (as an experiment) to encourage people in Asia (China, Japan, Korea...) to join us. This would make it 23:00 in Beijing, and midnight in Seoul and Tokyo, which is just about do-able. The good news is that since the US and Europe are now one hour ahead (because they've just moved into daylight saving time) we can achieve the same timing without shifting the call...

	08:00	San Francisco...
	11:00	Detroit...
	16:00	London...
	17:00	Paris, Berlin...
	23:00	Beijing
	00:00	Seoul, Tokyo...

I'll send out a short review of the last GGF in the next few days.

Cheers,

Mark.



NM-WG phone call, Thursday 24th March 2005, 16:10-16:30 GMT
-----------------------------------------------------------

Present: Mark Leese (Daresbury Lab); Dan Gunter (LBL); Martin Swany and Jason Zurawski (Uni of Delaware); Richard Hughes-Jones (Uni of Manchester)


Talking about the V2 schemas, Martin outlined the idea of introducing a "message type" into the schemas, so that we can move away from the fairly fixed model that every request results in a corresponding response being sent to the requestor. We may not always want this. In the abc-123 example below, Martin makes a request for data to Dan, but wants the data sent to Mark, he doesn't want the data himself...


     Martin     msgType="request"     Dan     msgType="statement"     Mark
        O----------------------------->O------------------------------->O
               "give data to Mark"                "RTT=10ms"


Another example of something not fitting the request-response model is a system that has registered/subscribed to receive updates, whether the updates are periodic or in response to some event. SNMP traps is probably the best analogy for this. 

Martin will put together a couple of simple examples for the next call.
 

Richard and Mark also gave a brief review of NM-WG at GGF13 (13-16th March). A separate summary will be sent by email in the next few days.