Sensing, Signals & Communication Seminar - Vivek Cadambe

Modern data services store data in distributed storage systems. In such systems, it is important to ensure that the data is available to the users even though the system components can be unreliable, for instance, the servers can crash. In addition to reliability, a common requirement, especially in applications to databases and distributed computing systems, is the following property known as consistency: when the data is being constantly updated, a client that reads from the system should obtain the latest version of the data. The design of algorithms for consistent distributed storage has been studied extensively in distributed computing theory, and forms the basic building block of numerous cloud-based key-value store services. The goal of this talk is to use ideas of information theory to understand the storage cost of consistent distributed storage systems.