Research And Design Of Transmission Method Of The Massive Remote Sensing Data Based On Reactor Model

The increase of the spatial resolution of the remote sensing image increases the total amount of remote sensing data,and the innovation of the new sensor technology in the remote sensing field also puts forward new requirements for the data processing capability of the remote sensing platform.The future development of remote sensing is becoming more and more diversified.The structure and composition of remote sensing data are becoming more and more complicated,and the storage and distribution of data are facing new challenges.Only the introduction of efficient data distribution technology,the development of remote sensing science can better serve the public.The key to the rapid distribution of remote sensing data and the high efficiency and stability of system data transmission in the face of a large number of concurrent requests is to reduce I/O processing time and select efficient data transmission nodes.Most of the existing network communication technology development using I/O blocking flow and multi-thread technology for large file I/O operation processing,it is difficult to competent in the resource requirements,high concurrent distributed network of massive data transmission tasks.In this paper,through a large number of analysis of remote sensing data transmission status,proposed data transmission strategy based on Reactor model,combined with the already mature network transmission technology.The model uses NIO technology in the node connection when using synchronous non-blocking I/O operation,will not cause connection blocking,while efficient use of CPU resources,reduce system latency.In addition,when the node chooses,it analyzes the different factors that affect the transmission performance and the distribution of the downloaded resources according to the bandwidth utilization rate,the load degree and the network delay of the current node,and allocates the download task reasonably.The main work and achievements of this paper are as follows:(1)Research on the distributed network communication data distribution technology and the domestic and foreign research results in this field.(2)Proposed based on the NIO Reactor master and slave thread model,when the client sends a connection request,the use of two-way non-blocking Channel,use the Selector to polling the Channel to access multiple clients to connect,improve the I/O implementation efficiency.(3)Analyze the optimal node factor ?i by analyzing the bandwidth utilization,network load and current network delay of the resource node through the theoretical bandwidth,the "usage ratio method" and the Ping value.According to the download task of the resource node,the algorithm is improved to find the optimal node selection scheme,adjust the download node,avoid the network congestion,and finally improve the parallel transmission performance of the large amount of remote sensing data.