Protocol Identification Of Satellite Based On The Data Mining

The 21 st century is the era of information explosion, the network as a medium of information dissemination has also been rapid development, from the Internet to a wireless network, all undergoing rapid changes. Satellite network as a means of communication, has many concern, with the development of the satellite business forward, satellite is ranging from the use for military expansion to civilian fields, which is involving mobile communications, television broadcasting, navigation, data transmission, also, the number of satellite protocol agreements have rapid increased, but tools for the satellite protocol analysis has a small number, especially when in the situation of information warfare, the first step is to access their protocol information, which is also a critical step in understanding the other protocol information in order to continue the follow-up work. Protocol identification has began when protocol started be bring out, but most of protocol identification skills are concentrated in the Inte rnet application- layer, from means of port number to later identification of the static characteristics in the field of load, are still targeted at the application- layer protocol identification, but few means of identification focus on the satellite protocol, this paper will do some research on the problem of the satellite link- layer protocol’s identification, with the means of data mining and the perspective of the static characteristics. The main content is divided into three parts.First, we discusses the application- layer protocol identification of using the algorithms called Apriori_App Layer which is based on the skill of data mining, its main function is to dig the frequent pattern in application- layer protocol. In order to improve the Apriori_App Layer algorithms’ s efficiency when faces the Long frenquent pattern, we propose Apriori_App Layer algorithm with a reverse connection, the aim is to reduce the number of protocol data traversing, so it can achieve the efficiency purposes.Secondly, in the consideration of the characteristics of the satellite link- layer protocol, we proposed a algorithms called Apriori_Bit based on the data mining, whose functions includs synchronous byte and frame size digging, frequent bit string pattern digging, association rule digging, and also analysis algorithm accuracy in error environment; proposed the weight of Apriori_WBit algorithm to solve the problem between " redundant itemsets " and interesting pattern in order to improve the efficiency of the algorithm.Finally, in order to deal with the large volume of data, we put Apriori_Bit algorithm into the efficient and stable Hadoop cloud computing platform. Meanwhile in order to futher improve the efficiency of Apriori_Bit algorithm which running on Hadoop cloud platform. We analysis some strategies which are used to improve the efficiency. According to the perspective that reducing transmission cost between different nodes and reducing I/O cost. We present four strategies including compressed intermediate data, modify job attributes, adding Combiner intermediate functions and data preprocessing.