Search

Various types of network security devices generated a large number of redundant alarm information with the high rate of false alarms. Alarm information is very trivial which is more difficult to analysis and understand the alarm. In order to research this problem, this paper proposed an alert aggregation method for multi-source heterogeneous alarm. By analyzing alarm type, source IP, destination IP, destination port and time interval, it summed up four rules, dynamically updated the time interval threshold and improved degree of accuracy. The experimental results show that this method can efficiently reduce the number of heterogeneous alarm information, get simplified super alarm data, and realize the real-time processing ability of the alarm information. [ABSTRACT FROM AUTHOR]

This paper proposed a labeling to data stream binding method independent of the contents of the packet, which based on packet delay modulation, and introducing the packet inter-arrival delay as the new labeling vector. The method used Himming code to achieve the label error control, designing the data packet randomized blocks, and modulated the packet delay time according to the label binding rules, then realized the implicit binding of security label to data stream. Based on this method, it solved the targeted attack problems caused by explicit binding method and its inflexibility. Theoretical analysis and experimental results demonstrate that the proposed method is robust against the network random timing perturbation and safety the binding security label when is transmitted in the network. [ABSTRACT FROM AUTHOR]

The ranges of state variable value of transition system are large, and the state space of system transition, but the transitions of the system are often unpredictable. This paper defined the concepts: initial safety states, safety transition condition, safety state, safety action, safety run and safety transition system, and proved the all states in safety transition system were safe. Then it used it to remold the network transition system, and changed it a safety transition system. [ABSTRACT FROM AUTHOR]

Courier tracking number stolen and express list information are not protected carefully that may cause privacy leakage. So this paper proposed a new K-anonymity model to protect express information. Unlike most anonymity methods, this method anonymized data by randomly breaking links among attribute values in records. By data randomization, this method maintained statistical relations among data to preserve knowledge, whereas in most anonymization methods, knowledge was lost. Further more, it proposed the algorithm for extra privacy protection to tackle the situation where the user's prior knowledge of original data may cause privacy leakage. Simulation results show that the new K-anonymity method can enhance privacy protection and improve the accuracy of the knowledge protection. [ABSTRACT FROM AUTHOR]

This paper proposed a multi-authority attributed based signature scheme based on distributed ABE, which was suitable for distributed system. It compared with traditional schemes, our scheme allowed an arbitrary number of authorities to independently maintain attributes and distribute secret attribute keys. The master authority was responsible for the distribution of secret user keys, without need to supervise each attribute authority. Attribute authorities were responsible to verify whether a user is eligible of a specific attribute, and every attribute is associated with a single attribute authority. Our construction was very efficient, as it required only a small amount of pairing operations during sign and verification. Our scheme has a good flexibility, and has a great of practical value in practical applications. [ABSTRACT FROM AUTHOR]

In P2P network, performances of many nodes in Chord algorithm are inconsonant. The problem that nodes leave and join frequently will restrict the performance of system seriously. Therefore, this paper presented grouping improved algorithm based on related degree of information. It introduced the concept of the node information relevant degree in the algorithm, in which the original Chord was adjusted by group on the related degree of information. Then the super group was connected through the super nodes. It selected two super nodes from each group and constitute a super group. In the meantime, every node was increased a counter-clockwise route. Then they could choose the shortest path to find two super nodes in both clockwise and counterclockwise direction. The experiments show that performance and adaptability of system are strengthened by the improved algorithm. The search efficiency of Chord in P2P network is improved. [ABSTRACT FROM AUTHOR]

People have different understanding about domain knowledge and system goals, and expressed demand for software is not standardized, so that the domain requiremented are often inconsistent and have to be changed frequently. This paper presented a method of domain requirements description based on limited language, by designing the restricted language requirement template to descript domain application needs. Then it checked the legality of vocabulary and matching the sentence patterns to verify the correctness of domain requirement description. And it saved the domain requirement in XML files. Finally it built instances to verify the effectiveness. [ABSTRACT FROM AUTHOR]

Aiming at the problem of ranking keyword search results in relational database based on data graph, this paper presented two factors based results ranking method which combined result's structure weight and content matching. It often used content matching in field of information retrieve. Firstly, it used result's path weight to measure the compact degree among keywords, results could get a rough ranking according that. Then for results with the same structure weight, keyword frequency in information tuple and amount of information that contain keywords were used to further ranking. Experiment and analysis shows that this ranking method can improve result precision by making highly relevant results in front of result set. [ABSTRACT FROM AUTHOR]

After analyzing the sparsity of traditional collaborative filtering algorithm and the merit and demerit of current solutions, this paper proposed a collaborative filtering recommendation algorithm based on user characteristics and item attributes which was mainly based on the item-based collaborative algorithm. It predicted the unrated items by analyzing different users' interests to various attributes of items and integrating the attributes of rated items to reduce the sparsity of data sets, and then to improve the accuracy of items' similarity calculation. The experimental results based on MovieLens data set show that the sparsity of extreme data sets can be reduced effectively, to a certain extent the proposed algorithm alleviates the cold starting problem in collaborative filtering algorithm and achieves better prediction accuracy. [ABSTRACT FROM AUTHOR]