We hate latency
So we went ahead and created a new type of region : Truncated region.
This one hosts only catchup nodes and truncated indexers but is able to handle between 85% to 95% requests locally. Request that cannot be fulfilled locally are forwarded to the closest full region.
This optimization allows us to deploy the service in regions where baremetal is either not available or too expensive.
In May we’ll be adding new regions , starting with:
- Singapore (MainNet, Front-End only)
- US-West (MainNet, Front-End only)
For now this change only benefits MainNet front-end endpoints.
No change is required to use the new regions as endpoints are proximity balanced.
See API page for details.
Happy low latency querying.