Replication
The topic of this lesson is replication. We often replicate data and services for various reasons: improved performance, higher availability, and improved fault-tolerance. Replication is often the key to achieving those properties. Replication sounds deceptively simple, but as we will see during the lecture, ensuring that all replicas (copies) are up-to-date in a consistent way is hardly trivial. For some applications, strictly consistent replicas are not required -- the high availability property is more important. One such system is called Gossip. We will briefly introduce it during the lecture, leaving the details of the architecture and system up to you to study on your own.
Download the lecture notes as a PowerPoint file (you can easily print handout views using PowerPoint, or just use these). You will also want to view the Chapter 15 slides from the book.
Reading
Intensive reading: 15.1 - 15.4.1
Extensive reading: 15.4.2 - 15.4.3
Exercises
Primary set: 15.1, 15.2, 15.5, 15.7 - 15.10, 15.12 - 15.14
Secondary set: 15.3, 15.4, 15.6, 15.11