Hierarchical Consensus for Distributed Database Throughput

The patent divides the distributed database into multiple shards, where each shard is managed by a subset of nodes rather than requiring all nodes to communicate with each other. This segmentation reduces the communication complexity from O(N^2) in traditional BFT to approximately O((N/k)^2) where k is the number of shards, thereby improving throughput while maintaining fault tolerance within each shard

The patent introduces a committee structure where a selected subset of nodes acts as intermediaries (committees) for each shard. These committee members handle the consensus process on behalf of the entire network, reducing the number of direct node-to-node communications required while maintaining the security guarantees of BFT consensus

By organizing nodes into multiple shards with dedicated subsets handling each shard, the system allows linear scaling as N increases. Each shard operates independently with its own consensus process, so adding more nodes primarily increases the number of shards rather than the communication overhead per shard, enabling better network performance with increased decentralization

The patent adds a dimensional organization to the node structure by introducing shard assignments and committee selections. Instead of a flat N-node structure requiring O(N^2) communication, the system creates a multi-dimensional hierarchy (nodes → shards → committees) that reduces communication complexity while maintaining the benefits of having more nodes in the system