You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Right now a medium-end CPU can roughly handle 7tps. While this is more than enough right now, the situation can rapidly change and a solution should be already in place.
Current Implementation:
The node is so slow to process the transactions caus of different implementation pitfalls:
txs and blocks are not saved atomically in one single DB transaction.
pending txs are undo + block txs are unconfirmed and then confirmed (3 steps)
Every action hits the Database to ensure data consistency
The txs checks cannot be parallelize due to the way the core is written
Ways to improve
Moving the data consistency checks to a faster db (Redis?)
Avoid undoing all unconfirmed transactions before applying blocks txs.
Use redis as single source of through (all writes and reads are there) and sync to postgres later (asynchronously)
horizontal scaling through multi-core
The text was updated successfully, but these errors were encountered:
Right now a medium-end CPU can roughly handle 7tps. While this is more than enough right now, the situation can rapidly change and a solution should be already in place.
Current Implementation:
The node is so slow to process the transactions caus of different implementation pitfalls:
Ways to improve
The text was updated successfully, but these errors were encountered: