Choosing a regional topology
The terminology for regional topology is pretty cut and dry, but when we get down into the weeds it is common to find lots of variations in the overall interpretations and implementation details. In the context of serverless systems, we have two main options, primary/secondary and active/active, with some variations. Let's look at these in the order we might adopt them.
Primary / Hot-Secondary
Primary/Secondary is the most common regional topology for traditional systems. We route users to the primary region (east) and we failover to the secondary region (west). The failover process usually involves manual intervention, so it is more correct to call this a Primary/Cold-Secondary topology.For our serverless systems, we employee a Primary/Hot-Secondary topology. We are deploying all our services in both regions, and they are ready to go, because we are replicating the data in near real time, as we will see in the Replicating across regions section. We...
