Research Of Network Protocol Robustnes Test Platform Key Technology

In various applications of the Internet, the robustness of network protocol nodes is an important aspect to guarantee the normal operation of the network. In order to guarantee the robustness of the network protocol, an important method is to construct the protocol robustness test case, enter the corresponding data and find the defects of the network node. Because there are too many types of protocol defects, so it is necessary to write enough test cases to find out the defects of the protocol. Because of low efficiency and high cost of manual test, the protocol robustness testing platform automation technology has been a hot research.Automated protocol testing can improve the efficiency of the the protocol test. Completely using computer generated test data, it is difficult to carry out effective testing, the result of complete computer test is low, and the job is inefficient. So how to effectively build test cases for the automated test, according to the protocol characteristics and the protocol standard, with rich error type, so as to improve the efficiency of automated testing, is an important research topic.Protocol testing platform automation involves these related key technologies:1, protocol standard automatic analysis; 2, test case and data automation genarated; 3, automation of test case execution. This article carry out research work from these three aspects.The main contributions and innovations of this paper are as follows:(1) An improved TF-IDF algorithm is proposed to analyze the protocol data, which improves the efficiency of protocol data format extraction, and this paper verifies the algorithm. There are many algorithms for automatic analysis of protocol data. The TD-IDF algorithm originally used in the analysis of Internet data. In this paper, we improve the TF-IDF algorithm for protocol data format analysis by combining term frequency and document frequency, and dynamic adjustment of the frequency threshold parameters, in order to achieve high accuracy automatic analysis of the protocol data. The accuracy of the key words extraction of the protocol is 70%-80%.(2) VOPN based on Petri net is proposed in this paer, which provides a formal method for the construction of defect test cases. In order to be able to test the protocol, the testcases should be designed according to the protocol’s state with formal method. In order to complete the testing process, this paper proposes a defect testing model based on extended Petri net to solve the problem in this paper.(3) A robust test data generation method based on matrix perturbation is proposed to improve the efficiency of test data generation. The method improve the generation speed of defect data and diversity of error types to make progress of the test automation protocol data generation efficiency, through the protocol message mapping to integer matrix and the combination of matrix operation and defects of data generation.Experiments show that the test of the SIP server is successful. The proportion of the total number of test cases which found defects was about 27%.(4) Research on the automatic test platform based on TTCN-3. In the platform based on TTCN-3, in addition to the test cases mentioned above, the encoding and decoding of the test data is the key problem to be solved. In this paper, a general TTCN-3 testing data encoding and decoding module is constructed, which is based on the format description file.Finally, in order to evaluate the solution, this paper has tested some network protocols. Using the method proposed in this paper, good experimental results are obtained. It is proved that the theory and method presented in this paper is feasible and effective.