Materialization of data cube in flash memory based on condensed data cube

In On-Line Analytical Processing applications, we can shorten the query response time and increase the efficiency of online analytical processing by the implementation of the memory data cube or other methods. According to the specific characteristics of NAND flash memory, combined with condensed data cube tuple storage characteristics and the three level storage structure of "Disk-Memory-NAND flash memory", Multi Level Dynamic Perfect Hash index structure is proposed, and then we have fully designed and implemented the details of the hash function. Implemented by the memory of coarse-grained cuboids, in the NAND flash memory on the implementation of fine-grained meta-group, thus the construction of the entire index was completed. It will be serialized into a write sequence of operations to the way each additional solution caused by the data insertion "frequently write" problem. Final experiment results showed that: the index structure to ensure order and perfect (no conflict) to add bulk update, for a single read and write operations, the overall disk storage and better than Static Hashing algorithm, it fits the NAND Flash application environment very well.