Data Quality

When working with multiple RPC providers and thousands of nodes, data quality can become problematic. Especially as nodes fall behind or, even worse, are maliciously serving data. We recognise these are issues when being an RPC router/aggregator and have built systems to combat these issues.

RouteMesh has two systems to ensure data quality and consistency:

1. Replay checks. Every nth request in our system is executed asynchronously against multiple other nodes and results are compared.
2. Lag checks. Every nth request on a chain, a latest block number call is made to check how far from the tip that node is. This is done through comparison with other nodes.

Both replay and lag checks are done on a chain basis and will run every hour if there are not enough requests on that chain.

If a node fails either check, they have a few strikes before they go into what we call a "staging arena" where they have to prove themselves before they are allowed to serve real traffic again.