You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
When using a storage storage service attached to a data service node that also has local application endpoints/sockets, bundles that get pulled from storage end up on the ingress queue. Therefore it becomes common to get bundles in the ingress queue which are actually destined for the same data service.
These bundles will traverse the routing table and will be delivered back to the egress queue of the same service.
This is really just an optimization concern, the flow works as-is but might be a little more expensive/roundabout than it needs to be. The data service ingress routine could check for this possibility, and turn it around immediately, without going to the central routing table.
This is low priority for now, because it does work as is, and things may change again before this becomes a performance concern.
The text was updated successfully, but these errors were encountered:
When using a storage storage service attached to a data service node that also has local application endpoints/sockets, bundles that get pulled from storage end up on the ingress queue. Therefore it becomes common to get bundles in the ingress queue which are actually destined for the same data service.
These bundles will traverse the routing table and will be delivered back to the egress queue of the same service.
This is really just an optimization concern, the flow works as-is but might be a little more expensive/roundabout than it needs to be. The data service ingress routine could check for this possibility, and turn it around immediately, without going to the central routing table.
This is low priority for now, because it does work as is, and things may change again before this becomes a performance concern.
The text was updated successfully, but these errors were encountered: