Leader Election for Replicated Services Using Application Scores

Abstract

Abstract. Replicated services often rely on a leader to order client requests and broadcast state updates. In this work, we present POLE, a leader election algorithm that select leaders using application-specific scores. This flexibility given to the application enables the algorithm to tailor leader election according to metrics that are relevant in practical settings and that have been overlooked by existing approaches. Recovery time and request latency are examples of such metrics. To evaluate POLE, we use ZooKeeper, an open-source replicated service used for coordinating Web-scale applications. Our evaluation over realistic widearea settings shows that application scores can have a significant impact on performance, and that just optimizing the latency of consensus does not translate into lower latency for clients. An important conclusion from our results is that obtaining a general strategy that satisfies a wide range of requirements is difficult, which implies that configurability is indispensable for practical leader election.

Improving Wide-Area Replication Performance through Informed Leader Election and Overlay Construction

“…Becker et al [3] have shown how the leader election influences the response time for cases in which the request distribution is skewed. In our example, if all requests are issued at replica s 1 , replica s 1 is a better leader than, for instance, replica s 2 .…”

“…Electing a good leader for replication in WAN has been studied by Becker et al [3]. In their work, they propose a bullying [8] leader election configurable with user-defined scores.…”

Leader Election in a Cluster using Zookeeper

“…Other research proposes a model based on watchers in ZooKeeper and define a watch as a trigger that causes an event to be dispatched to the client whenever the watched resource changes its state. Due to the fact that the processes are asynchronous and as a consequence, the network latency gives rise to multiple possible orderings of network messages; so the model was improved in order to enable consistency [2,3]. Another proposal for a leader election algorithm for replicated services that are based on a leader, updates propagation and client request was POLE (Performance-Oriented Leader Election) [5], the algorithm selects the leader depending on an application specificity.…”