Research Of Key Technologies In High Speed Acquisition System

With the continuous development of network technology,Internet has become an indispensable part of people's lives.The number of network users and the amount of data in the network link show a trend of explosive growth.Various types of data-related requirements have been increasing rapidly.The swift growth of the flow puts higher demands on the performance of the data acquisition system.An excellent high-speed acquisition system needs to cost reasonable system resources,under the premise of high accuracy and outstanding real-time features.Data acquisition and parsing are helpful for improving network management and monitoring efficiency,as well as getting valuable network information.In this thesis,the topic of research is focus on the key technologies in the development of the high-speed acquisition system.Firstly,the basic framework of the acquisition system is introduced and lead to the key points in the system development.The main modules in the framework will be descripted in detail.In part of the memory management module of the acquisition system,in order to improve memory management efficiency,a self-adaptive variable block size memory management mode which can manage the hotspot domain and the cold domain separately will be designed,based on the analysis of the related memory management structure and the combination with the memory allocation and release methods of the current acquisition software.As for the classification of retransmission and out-of-order types for TCP packets,the thesis will introduce the current classification benchmarking algorithm which is based on two-way TCP packets.Then a rational one-way TCP classification algorithm will be designed,using statistical characteristics training method inferred by analyzing the characteristics of TCP packets.The size of uncompressed data affects the performance of the entire system significantly.In many sence,the time consumed for uncompressing data is relatively large.The thesis will design a parallel data uncompress method and apply the parallel processing model to Data uncompress process.Finally,the thesis summarizes the research topic and prospect the future research directions.