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
Stream priority is very useful for some QoS aware responses. Considering such a situation: several batch requests need bulk data transfer, together with some high concurrency latency sensitive requests with small packet sizes are multiplexed, without a stream priority, the latency sensitive requests will be blocked by those batch ones. Currently, I could only see a simple function definiton of writePriority within the project which just set the priority flag, corresponding multiplexing processor has not been found. Could such useful feature be delivered by grpc-go ?
The text was updated successfully, but these errors were encountered:
This is definitely possible, but it would take a moderate amount of redesign to implement. I don't imagine this is something we would be able to get to in the next ~6 months, but it is something we'd like to do eventually. In the meantime, one workaround is to create multiple connections and send different workloads on each by priority, and leverage external QoS to prioritize them.
Note that we don't transmit RPCs serially in the order in which they are initiated. Small requests that start after large ones should typically be able to complete before the large requests.
Stream priority is very useful for some QoS aware responses. Considering such a situation: several batch requests need bulk data transfer, together with some high concurrency latency sensitive requests with small packet sizes are multiplexed, without a stream priority, the latency sensitive requests will be blocked by those batch ones. Currently, I could only see a simple function definiton of
writePriority
within the project which just set the priority flag, corresponding multiplexing processor has not been found. Could such useful feature be delivered bygrpc-go
?The text was updated successfully, but these errors were encountered: