Design

ReadRPC design

Diagram

Quick Overview

There is a database (ScyllaDB) in the middle of the diagram.

Below, you can see different indexers (two different types). There might be a lot of instances, though. Please note the semi-transparent placeholder for ?-indexer. We can potentially introduce additional ones for different data.

Please read more about the indexer on the [Components].

At the top of the diagram, we see a simple Load balancer pointing to the swamp of ReadRPC instances. A straightforward case for Web 2.0 is a load balancer before the web server.

There are already existing services on the sides of the diagram:

• NEAR Lake infrastructure that is needed for the indexers
• nearcore JSON RPC node(s) to:
  • proxy “send transaction” requests to
  • proxy some other methods we haven't implemented yet or don't plan to implement on the ReadRPC side