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
The IAVL store (Cosmos’ app level storage tree) is the legacy implementation that serves as both the merkle root integrity structure and the general purpose database structure for reads. Though the structure is seemingly efficient at generating merkle roots for the ABCI AppHash, Pocket Network observes a considerable downgrade in performance for historical queries, block processing, and overall disk size management. The idea behind Pocket Network’s Persistence Replacement is to preserve the efficient state commitment root generation while providing an alternative implementation for DB reads. The idea is to prune N-X IAVL versions in order to enable merkle tree generation, while maintaining a completely separate storage layer for efficient storage, historical queries, and block processing. The idea and design is very similar to Cosmos’ ADR-40.
Required Goals:
Simpler Codebase
Non-Consensus Breaking
Comparable syncing performance
Comparable disk usage
Additional Goals:
Faster historical queries
Faster current queries
Smaller disk usage
Better syncing performance
Slower disk usage rate
The text was updated successfully, but these errors were encountered:
The IAVL store (Cosmos’ app level storage tree) is the legacy implementation that serves as both the merkle root integrity structure and the general purpose database structure for reads. Though the structure is seemingly efficient at generating merkle roots for the ABCI AppHash, Pocket Network observes a considerable downgrade in performance for historical queries, block processing, and overall disk size management. The idea behind Pocket Network’s Persistence Replacement is to preserve the efficient state commitment root generation while providing an alternative implementation for DB reads. The idea is to prune N-X IAVL versions in order to enable merkle tree generation, while maintaining a completely separate storage layer for efficient storage, historical queries, and block processing. The idea and design is very similar to Cosmos’ ADR-40.
Required Goals:
Additional Goals:
The text was updated successfully, but these errors were encountered: