5 ms·
Latency has always been part of the CAP theorem. There is a strictly stronger generalization of it that says in a consistent system the latency of a transaction
by kcolford 5y ago
Latency has always been part of the CAP theorem. There is a strictly stronger generalization of it that says in a consistent system the latency of a transaction is bounded by the latency of the links in the network. In the case of a partition the network latency is infinite, so too is the transaction latency, thus no availability.