Research On Spatial-Temporal Data Offline Cache Management Technology

Spatial-temporal data having lots of features, such as multi-source heterogeneous, huge amount of data, uneven spatial distribution, and so on, have caused the problems of the information system of the client, such as poor visualization, low efficiency. In order to solve the bottleneck problem of I/O disk data, spatial-temporal data visualization and reduce network data access time and improve the effect of data visualization, this paper makes a research on the offline cache management technology on spatial-temporal data, the offline cache management model for fast access, improve the internal and external storage replacement algorithm and pre-scheduling method. The main work of this paper is as follows:1) Design for the offline cache management framework based on the visualization of spatial-temporal data. Focus on the analysis of offline cache management structure and cache strategy; formulate relevant cache data structures, such as memory cache pool, page table management, the core index file and offline data package etc.2) Analysis of cache contents and traditional spatial-temporal data SA, LFU and LRU memory replacement algorithm, propose four fork tree caching strategy, four fork tree buffer structure and its rationality, through increasing or reducing the number of access times of adjacent data to change the data page, to improve the spatial correlation of temporal and spatial data, so as to meet the characteristics of spatial-temporal data.3) This paper analyzes the characteristics of spatial and temporal data scheduling techniques, analyzes the scheduling method based on visual correlation, user behavior and spatial structure, and realizes the pre scheduling method of two kinds of spatial-temporal data.4) The design and implementation of the offline cache management model and visualization application system based on the temporal and spatial data of marine meteorological data, the feasibility and effectiveness of the data cache and scheduling, and the results of the system running results are completed.