Scalable Human Blog

Ever wondered how some applications can perform even without an internet connection? This post will detail how some applications handle this challenge.

In handling offline operations between devices, the Multi-Leader Replication architecture is suitable to achieve this.

For example, consider a calendar application on your laptop, mobile phone and other devices…

What type of operations does this need to support?

• You need to see your meetings
  • Which means making read/write requests at any time
  • Regardless of having an internet connection or not
  • If you make any changes when offline, they will need to be synced with the server and other devices

Multi-Leader Replication architecture to the rescue…

• In this type of solution
  • Every device has a local database, that acts as a leader, meaning they all accept write requests
  • There is an asynchronous multi leader replication process that will synchronises between replicas of your calendar app on all devices
  • The lag can be hours or even days for waiting for internet access, which make this an attractive solution
• The reason why this architecture is similar to the Multi-Leader Replication
  • Each device is a data centre
  • To be mindful of, the network connection between these devices can be extremely unreliable

Be warned about the past…

There is a plentiful history of broken calendar implementations… 😑 Multi leader replication is a tricky thing to get right!

Not all is lost…

There are tools out there that make this easier to apply multi leader configuration. A recommended technology that is designed to handle this mode of operation is CouchDB, which is also open source document-oriented database (NoSQL) and implemented in Erlang. Worth checking out, if your application requires this.

📚 Further Reading & Related Topics

If you’re exploring clients with offline operations in distributed data-intensive systems, these related articles will provide deeper insights:

• Distributed Data-Intensive Systems: Replication vs. Partitioning vs. Clustering vs. Sharding – Learn how partitioning and replication strategies are critical in maintaining offline data consistency and synchronization in distributed systems.

• How Does Partitioning Work When Requests Are Being Routed? – Explore how request routing strategies, combined with offline operations, manage data consistency and availability in distributed systems with limited connectivity.