Offline Tree Data Preparation for Bulk Insert Bottlenecks

The patent applies preliminary action by sorting the bulk data items before insertion and pre-determining their positions in the tree structure. The server system sorts the received data items and calculates insert positions in advance, rather than determining positions during the insertion process itself. This preliminary preparation significantly reduces the processing time required during actual insertion operations.

The patent reduces processing power consumption by performing data sorting and position calculation in advance before the insertion operations begin. The server system prepares the sorted data and determines all insert positions beforehand, which minimizes the computational resources required during the actual bulk insertion process.

The patent applies segmentation by dividing the bulk data into smaller batches or groups that can be inserted in sequential segments. Rather than performing a single long-running insertion operation that blocks read access, the data is processed in multiple smaller insertion operations, allowing read operations to potentially occur between batches and reducing the overall blocking impact.

The patent merges multiple small data insertion operations into a single bulk operation. By combining all the data items into one bulk insert request rather than sending multiple individual insertion requests, the patent reduces network overhead and utilization. The sorting and position determination are performed on the server side after receiving the bulk data, eliminating the need for multiple round-trip communications.