# From Global Relay Network to Global Nostr Nodes

![Relayable Network]()

I have just finished deploying a new bare metal server in Sydney, Australia to complete the Version 2 of the Relayable Relay Network. With lessons learned from Version 1 that has been up for the last 6+ months. When I first started this project it was to improve the architecture and design of relays to be more highly available (HA) and redundant in terms of data storage in many geographic locations.

With the finalizing of Version 2 of the network many improvements have been made.

## Some include:

- Faster strfry (relay software) performance

- Better syncing of data across geographic regions

- Better syncing of configurations and site content

A big change is the move away from a singular cloud provider to a more heterogeneous model with 80% of relays not being on AWS or any cloud provider. With numerous backbone providers.