The Research About Multidimensional Data Indexing Architecture

With the rapid development of network applications, the data generated by the Internet showed explosive growth. It has become a major problem to implement such massive data index efficiently, it is even more difficult for multidimensional data indexing. Though cloud computing is an emerging computing infrastructure, it can be allowed much of data in a resource pool so that multiple computers in parallel to handle these large data, but as a important part of cloud computing platform infrastructure-cloud data storage system is especially critical in applications of multidimensional data. Since most of the current cloud storage systems are based on the way of distributed hash (DHT) to build data indexing, data is organized to store by the form of Value (Key-Value). Therefore, this type of cloud storage systems only support the inquiry of key, and it is not very satisfactory for the range query and nearest neighbor query problem in multidimensional data indexing of big data systems. Therefore, this paper attempts to put forward a new large multidimensional data indexing system architecture Skip-Octree under cloud environment in order to make some reference for such problems.This paper builds a new type of multidimensional data indexing architecture Skip-Octree on the basis of comprehensive studying the multidimensional data index and storage problem.This architecture adopts the structure of Octree to store data and establishes the corresponding index mechanismas on it so that it can make full use of the idea of different dimensional space partitioning to achieve a simple index for data. Secondly, on the basis of the architecture Skip-Octree, this paper also designs the algorithm about point associated inserting, deleting, querying, range querying andOctree tree splitting, merging algorithm about Octree. Finally, it tests the algorithms through the simulation experiment in order to prove the feasibility and high efficiency of the architecture, this experiment make test Respectively from the data point of inserting, deleting, querying, range querying, splitting, merging and evaluation about double layer structure. The Skip-Octree architecture is compared with the traditional Octree structure in the experiment, the results show that the Skip-Octree structure is better than Octree structure in multiple aspects and performance has been improved a lot. The appearance of this architecture not only simplifies the adjustment for the balance of the tree, but also take full advantage of Skip lists’ tiered release characteristics to achieve that the Octree is also released up by1/2probability. Through this way, the querying for dimensional space sudden becomes the indexing of linear structure. At the same time, it improves the efficiency and reduces the index storage space. From the experimental results, the multidimensional data indexing system architecture Skip-Octree is feasible and efficient; simultaneously, it also will play a very good reference for the applications for the enterprise.