On 4/7/16 11:02, Susan Hares wrote:
I2RS:
I2RS will hold its second meeting from 16:20 – 17:20pm. This meeting
is after the NETCONF meeting. We will talk about protocol-strawman with
any NETCONF participants.
Sue, at the meeting today you started an example talking about running
config with a priority of 1 and an I2RS Client with a priority of 10 (or
reverse, I don't remember). Then you said that the I2RS Client's write
would win. Agreed.
But you didn't mention anything about the "shattering" pane of glass.
The fact that running has a priority implies that any write to running
would still sit under the Client's pane of glass. I'm not sure if
that's what you intended to convey, but I like it. This makes sense to me.
What I anticipate in the case of a last-write-to-running winning is
config churn. Let me use your thermostat example.
I like it cooler in the room. When I visit relatives (that typically
like it warmer) I will set the thermostat to 16 C or so. I expect this
to stay until I leave. So I'm the high priority I2RS client overlaying
the config that will remain. If my relative makes a change to the
underlying "config" of the thermostat, thus destroying my intended,
ephemeral config, I will just go back and re-add my ephemeral config.
In the networking world, this will be churn that is likely not needed.
Joe
_______________________________________________
i2rs mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/i2rs
