Research On Data Security Of Cloud Storage Based On Immune Mechanism

Compared with previous computing model such as distributed computing, parallel computing and grid computing, cloud computing have many strengths in the aspects of low cost, data storage, and computing power, which can enormously push the transformation of information technology(IT) industry. Nowadays, more and more users choose to store data in the cloud data center, but application prospects and challenges are the two sides of a coin. One of the biggest problems is how to secure data. C loud computing uses a distributed data storage block, which storing and computing almost simultaneously, so that encryption in every piece of data and decryption in calculation area is a very time-consuming process, which can dramatically reduce the efficiency of data processing in spite of bring security. With the rapid development of biosciences, heuristic algorithm inspired by biology provides us a better solution.In this paper, a proposed method that based on biological immune mechanism has been introduced to solve the problem of passive and inefficiency caused by data encryption. The characteristics of immune system include features adaptive, robust and self-organization, which is more suitable for intrusion detection in cloud environments. In the beginning, we need transform user data into autologous and detector, adaptive learning and dynamic updates that can be recognized by the immune system. The main contributions of this paper state as follows:(1) In order to obtain the protection of artificial immune system, pre-processing operations that include discrete data continuously, standardization, normalization and low maintenance are pivotal to turn user data into autologous that ca n be recognized by the immune system. The algorithms of PCA and K-means have been applied to reduce data dimension and improve computational efficiency correspondingly.(2) The quantity and quality of detectors in the immune system directly determine the detection efficiency of the system. In the process of conventional detector generating, immature detector will be deleted, and then randomly generate a certain amount of initial detector until the number of the detector reach the set value. Such approach is inefficient, we propose a way to shift mutation through direction-vector, can improve the generating efficiency of the detector.(3) Another important indicator of the detector generating is coverage rate. If coverage rate is too low that is bound to produce a certain number of missing rate. For overcome such drawback, this paper takes advantage of evolutionary computation, and genetic algorithm in the detector generating process, so that detector can evolve by selection, crossover, and mutation in each iteration with the widest range of coverage of spare space. Additionally, in order to ensure the diversity of the population, a randomized calculate fitness approach be taken in the process, rather than the traditional sort of way in accordance with inclusi ve fitness.(4) The detector consists of two important factors that are adaptive learning network environment and dynamically updated mechanism. This paper will increase two more properties, that is, age and num. Within a certain age, if the number of detected intrusion smaller than the threshold, the detector will be mutated, otherwise it will be converted to "memory" detector. Meanwhile, according to the calculation of fitness, the detector will extract individual with high fitness through mutation and negative selection after replacing individual with low fitness. In addition, the paper also introduces the thought of "vaccine bank", that is, the detected intrusion individual will be added to the bank, so that the next generation of detectors to be generated more targeted.(5) This paper proposes an automated industrial production lines based on dynamic clone selection algorithm in the aspect of cloud data security, and detailed describe the working process of the production line. In terms of security defense, we use biological immune system to recognize and destroy the alien viruses. For fast-efficient extraction of data, we use biological immune "second response" mechanism. Meanwhile, for the scalability of clo ud environmental resources, the production line will train immune function, such as a detector set and autologous collection with previously copied to the new scalable cloud server which is similar to nascent field of individual organisms acquired from the mother "innate immunity."At last, we summarize the full text and look forward to the future research.