Data replication:
- Once you’ve absorbed that the strategy should fit the kind of environment you are placing your cluster in.
- If you want to stay in one region of Amazon web services? You can either set yourself up as one dc per availability zone (AZ), or you can set up your cluster to use one dc and treat the AZs as racks.
- Then you can replicate so that if an AZ goes down you won’t have a system problem, just some downed and unavailable servers.
- If an AZ goes down (or a rack goes down) that all the copies of an important piece of data will be available.
- You also are stuck with latency issues if you want to cross into a different region.
- Your data has to replicate into the new region and if you went ‘simplestrategy’ you are going to have massive traffic trying to keep the database in sync.
