Apache ZooKeeper is a coordination service built to manage large distributed systems. It coordinates the activities of different hosts and the use of common data with robust synchronization techniques.

While presenting itself externally as a single service, ZooKeeper forms a cluster of multiple nodes and server instances. This so-called Zookeeper Ensemble organizes itself by choosing a master node which takes the lead in synchronizing the cluster and managing consistency.

In a cluster, ZooKeeper provides the following services:

• Naming service to identify and address nodes in a cluster
• Cluster management to add or remove individual nodes
• Synchronization service to manage the saving and changing of distributed data
• Redundancy service to ensure high availability of data and services despite individual node failures
• Information service to provide real-time node status information
• Configuration service to provide real-time node configuration data
• Procedure for appointing a Master Node

Originally developed by Yahoo, ZooKeeper became a sub-project of Hadoop at Apache before becoming a standalone project in 2008. Today, ZooKeeper is a kind of standard for organizing distributed services and is used by HBase, Hadoop and similar frameworks.