For this we may want to implement a peering protocol by
which different scheduler instances can communicate amongst
themselves. In this way we could connect several different
clusters in different locations around the 'net. We would
potentially want to allow stuff like wide-area job migration
and distribution.
Implementation would be theoretically simple: just add a new
peering communication plugin and possibly override the
backfill scheduling plugin with a WAN-enabled version.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Logged In: YES
user_id=362364
For this we may want to implement a peering protocol by
which different scheduler instances can communicate amongst
themselves. In this way we could connect several different
clusters in different locations around the 'net. We would
potentially want to allow stuff like wide-area job migration
and distribution.
Implementation would be theoretically simple: just add a new
peering communication plugin and possibly override the
backfill scheduling plugin with a WAN-enabled version.