Mastering Apache Pulsar

Appendix C. Geo-Replication

Geo-replication is the process of replicating (copying) data across multiple regions or multiple datacenters. I didn’t talk much about geo-replication in the book because there is a considerable amount to cover on Pulsar and fitting this in was hard. I decided to include it in this appendix because I still feel it’s important. Typically, geo-replication has one of two motivating factors:

• Preparation for disaster recovery scenarios

• Localizing data or making global data available everywhere

Applications that run on the internet are subjected to many conditions by which they can fail. Applications run on physical hardware that sits in a datacenter. Datacenters¹ can lose power, lose connectivity to the internet, be flooded, or have their hardware fail. The way datacenters combat these problems is by enabling applications to run across multiple datacenters (regions), or multiple places within a datacenter (availability zones). The datacenter can guarantee some level of reliability for each zone and also each region. By spreading your application across regions and zones, you can decrease the likelihood that you will encounter a failure that results in your applications being unavailable.

Many companies that start on the internet start by serving one geographic audience. Typically, it is the audience in the country of their founding. As the company gains traction and looks to expand its footprint, it may open offices and offer its product outside its ...