For building any computing system High Availability is a desirable and often necessary property. In simple terms, high availability, in the context of

Synchronous Replication with High Availability — A New Approach

submited by

Style Pass

2025-01-22 23:30:04

For building any computing system High Availability is a desirable and often necessary property. In simple terms, high availability, in the context of a database or a cloud application, means that there is almost never downtime. The practical result of a high availability system is that when you go to Google on your phone or computer you never get an error page.

Computers, networks, and power grids are pretty reliable. Computers can run for months or years without restarting, but processes do crash, hardware does fail, and planned restarts are occasionally needed. Networks and power grids work most of the time but sometimes they don’t.

Because all of the components, both hardware and software, in a computing system are subject to occasional failure, redundancy is essential for building high availability systems.

In a typical cloud computing system, an individual server is hosted in a data center, which will have redundant power, ideally with multiple grid providers, a battery backup system to handle short term outages, and diesel generators for long term outages. Networking will be provided by multiple upstream providers, which will we accessed through multiple switches, and each server will have multiple network interfaces that connect to different switches. Every component of the system has redundancy built into it in order to provide the highest level of availability possible.