Summary

In this chapter, we covered three topics: idempotency, replication, and recovery. First, we discussed four scenarios where idempotency is useful and how it can be achieved with reference implementation.

Then, we briefly mentioned how to replicate data and services. We brought up the CAP theorem, in which trade-offs need to be considered for each system. We also delved into three models of replication, namely primary-secondary, partitioned and distributed, and quorum-based.

Finally, we covered some common mechanisms of recovery, outlining how a newly launched node can become operational in the context of distributed systems.

In the next chapter, we’ll cover the audit and monitoring aspects of a distributed system.