Presentation: Consensus: Why Can't We All Just Agree?

Track: Modern CS in the Real World

Location: Mountbatten, 6th flr.

Duration: 10:35am - 11:25am

Day of week: Monday

Level: Advanced

Share this on:

Abstract

Reaching agreement is never easy and distributed systems are no exception to this rule. In this talk, we take a journey though the history, to the current reality and look ahead to the future for distributed consensus. We start over three decades ago, when the field of distributed consensus began with a proof of its impossibility. We will journey to today’s data centres where algorithms such as Paxos, Zab and Raft provide a fundamental service to widely-adopted distributed systems. Along the way, we'll organizing parliament for a Greek island and debunk widely held beliefs about distributed systems. At the final stop on our journey, we will explore how our recently revised understanding of consensus enables a new generation of scalable yet resilient distributed systems, many of which were previously thought to be impossible to achieve.

Speaker: Heidi Howard

Distributed Systems PhD Candidate @CambridgeComputerLab

Heidi Howard is a PhD student in the System Research Group at Cambridge University, Computer Lab, under the supervision of Prof. Jon Crowcroft. Her research interests are consistency, fault-tolerence and performance in distributed systems, specializing in distributed consensus algorithms. Heidi is most widely known for her generalization of Leslie Lamport’s Paxos algorithm for solving consensus, known as Flexible Paxos. Heidi can often be found on her blog: read, write & execute (/hh360.user.srcf.net/blog/) or as @heidiann360 on twitter.

Find Heidi Howard at

Tracks