Research On The Method Of Establishing The Forest Resources Monitoring Cloud Platform Based On Glusterfs

With the development of forestry information construction, there are more and more new methods and techniques applied in forestry. It is necessary to collect the monitoring data or forestry resources for a deep research on the field of forestry. At present, the monitoring work of forestry in China has various targets (including climate, soil, hydrology, quality of water and videos) and a wide distribution, great amount of various data need to be collected with not uniform structure. Only if the valuable forest monitoring data that are hard to get are well and effectively saved, managed and utilized, the analysis and the solution of forestry problems would be supported objectively and completely. This thesis will deal with the large data of monitoring forestry resources, storage of different their types and their parallel storage and computation. The research goes on as follows.Firstly, according to the wide distribution of forestry resources monitoring data, relative isolation between each other and large quantity of monitoring data, this thesis introduces the construction of cloud platform for monitoring forestry resources based on Glusterfs; by founding this cloud platform, we can remove the relative isolation of the monitoring data in all the areas and put all the monitoring data into the unified virtual storage resource pool we constructed, realizing logical unification of the data, helping the data shared and improving the utilization rate of stored resources.Secondly, according to the diverse formats and types of forestry resources monitoring data, by analyzing the logic structure of forestry recourses monitoring data this thesis roughly divides the data into structured one and unstructured one and design a the mixed storage mechanism, that is saving the data into the relational database or the NoSQL database according to their different types, solving the problem that different data types of data are stored in a same system.Thirdly, to deal with the equilibrium of data storage in the system, this paper applies the consistent hashing algorithm, making the storage and computation of data more balanced in the whole system; to the parallelization in dealing the data, this thesis designs the forestry resources monitoring cloud platform based on Glusterfs, applying MapReduce calculation model and technique of parallel database as well as realizing the parallel computation and storage of data.Finally, this thesis tests the simple Glusterfs cloud platform on its reliability scalability, flexibility and properties of eliminating the metadata etc. The experimental results show that the overall performance of the system is excellent. At the same time, issues of the locking and the file traversal still need to be improved continuously.