We will cover a diverse state of the art topics related to data management in the cloud. These will include:
Making the impossible possible: starting from the basic impossibility results of consensus, we will move on to Paxos and its varieties.
Geo-Replication: fault-tolerance in the presence of natural disasters requires global replication. However, as geographic distances increase, so does latency. We will explore how different systems have dealt with this challenge.
Web scale caching: many global applications need reliable distributed caching systems. We will discuss various systems and how they address the challenges of high latency, fault-tolerance and diverse workloads.
Other topics include blockchain, edge clouds, IOT, etc.
Prerequisites: A foundational course in Distributed Systems, eg, CS171 or cs271.