Author: Zen and the Art of Computer Programming
1 Introduction
Data sharding
In a distributed database, data sharding refers to splitting a large table into multiple small sub-tables or partitions according to business rules or certain rules, and then storing them on different physical servers to improve query efficiency and expansion. properties, etc., and each small sub-table can be called "sharding". This process is data sharding. Generally, different sub-tables are assigned to different machines for storage and processing, so that hardware resources can be effectively used to improve query performance.
Distributed Coordination Service
Distributed Coordination Service (DCS) refers to multiple independent nodes forming a cluster. Various components in the cluster work together to complete the management, coordination and configuration of distributed systems. Currently, the most mainstream DCS include Apache Zookeeper, Etcd, Consul, etc.
Apache Zookeeper
Apache Zookeeper is an open source distributed coordination service. It is a distributed consistency solution for distributed applications, built on the atomic broadcast protocol. Its design goal is to perform atomic updates in a distributed environment to ensure consistency between data copies of each node. In addition, Zookeeper also provides a relatively simple synchronization mechanism. The client can send requests to the Zookeeper server and obtain updated results. For example, a distributed lock can be implemented through Zookeeper. Zookeeper is an independent server cluster. The client does not need to connect to all Zookeeper servers at the same time, but only needs to connect to one of the servers.