Design And Implementation Of Data Stream Management System For Network Data Flow

With the rapid development and widespread popularization of the Internet, the network has become an indispensable part of production and our daily life. Via the Internet, billions of data sources are generated in every second. Additionally, due to the improvement of hardware devices, various fields of departments are totally capable of producing sequences of high-speed mass data automatically, continuously, and precisely, which we also name it stream data for short. Because of those characteristics of the stream data, it is difficult, or say it, impossible to analysis and manage these data. The design and implementation of management system is of great significance.In this paper, we design a stream data management system for network data flow, for the actual needs of real-world. The system is based on GNU Linux system, and developed by the C/C++language, which can provide users with specific IP-based real-time queries and frequent item statistics-, In this thesis, we describe the important six modules of the system digitally.The system discussed in this paper changes phenomenon that we use traditional static database to manage the stream data and reduces the waiting time greatly when people query frequent items in a large amounts of data. What’s more,the system also meets the need of those who want the result with real-time analysis data immediately under certain circumstances.