Thanks, I understand. The sequential locking isn't a requirement per-se, and the assumption is that the N-masters are in the same region. It seems like with clever multiplexing, you could relax the requirement the N-masters are in the same region.
Definitely, even without proper multiplexing, if the connections are not blocking, it is possible to just lower the latency to MAX(latencies) sending the commands in a loop to all the instances, and later reading them.
Redis clients should be capable to do that, by allowing to separate (on demand) the command delivering from the reply reading moments.
Comments
Thanks, I understand. The sequential locking isn't a requirement per-se, and the assumption is that the N-masters are in the same region. It seems like with clever multiplexing, you could relax the requirement the N-masters are in the same region.
Definitely, even without proper multiplexing, if the connections are not blocking, it is possible to just lower the latency to MAX(latencies) sending the commands in a loop to all the instances, and later reading them.
Redis clients should be capable to do that, by allowing to separate (on demand) the command delivering from the reply reading moments.