The Design And Implementation Of The Storage Engine Of A Distributed And Structured Storage System For POI

With the gradual application of augmented reality and the explosive growth of spatial data in life, there are more stringent requirements proposed to deal with the storage of spatial data, especially in availability, scalability, and efficiency and so on. Traditional distributed data storage scheme does not meet the requirements of data storage space, and most of the spatial database can neither effectively store large amounts of spatial data, norachieving docking with the Internet, or accomplishing a combination of the online and offline, so the distributed structural storage for spatial oriented system is imperative. Distributed storage system for the POI is such a storage system for managing POI data.Firstly, the thesis generally introduces the augmented reality and the current status of cloud computing research, conducts a survey of several domestic and international mainstream distributed storage systems, and describes the advantages and disadvantages of each system. The application of augmented reality in the various technologies becomes mature, to solve the storage of massive spatial data, the paper designed and implemented distributed structured storage system for the POI(CBASE), and focused on the research of data storage engines.CBASE uses a distributed structure, with MySQL being the underlying storage database of every storage node, because MySQL spatial extensions can be applied to realize the main support for spatial data.CBASE design a good spatial index mechanism by using the POIID parsing rule which merged a unique identifier and POI’s range into POI’s ID, this rule can hide the internal query mechanism from customers, improving the transparency of the system. The secondary indexes was used to improve the efficiency of the system, in which the first level index implemented by the R tree can locate the spatial data according to the POI’s coordinates, and the second level index implemented by the index table can find the storage nodes according to the spatial location of POI data.The duplicate strategy was employed in CBASE to implement fault-tolerant mechanism, and it can enhance the consistency of the data. System will do the synchronization among storage nodes by Lazy Synchronization strategy as long as the data be updated.System take the majority strategy to modify data, this strategy will send result to users unless more than half storage nodes return success result. It can improve the efficiency of the system at the same time guarantee the validity of the data.CBASE designs the four-split area strategy to manage load balancing of system, four-split strategy will split the region if there are too many POIs in it, which can ensure there won’t be too much access in this region. CBASE designs the domain migration and domain split to migrate the data from one region to another region, the two process also be useful for storage nodes and can prevent the data’s hot issues during system’s running-time, improve the efficiency of the system.Finally, the thesis describes the realization of the system, accomplishes a functional and performance test, and generates and analyzes the results.