Research On Dynamic Encryption Strategy Based On Data Sensitivity

In today's era,data,as an important factor of production,had brought huge improvements to the creation of economic value.With the rapid expansion of the market scale of the big data industry,the open sharing,exchange and circulation of big data had become a trend.However,the issue of personal privacy leakage is hindering the development of the digital economy.With the disclosure of data,the image of users on the Internet was gradually enriched,and the risk of personal privacy leakage was increasing.In order to protect the security of personal information of network users,some privacy protection algorithms have been proposed by researchers.However,these privacy protection algorithms were not profitable and harmless.Anonymity-based privacy protection algorithms need to be at the cost of data availability.Encryption-based privacy protection algorithms need to use a lot of computing resources.Privacy protection algorithms based on differential privacy require add noise to the original data entropy.Excessive noise will sacrifice the training effect of machine learning and bring accuracy deviations.In summary,no matter what kind of privacy protection algorithm,the intensity of privacy protection and the cost of use are directly proportional.It was necessary for researchers to estimate the degree of privacy risk required by the current data before data privacy protection,then how to measure the privacy requirements of data and how to use privacy protection algorithms according to privacy requirements were two crucial issues.Therefore,for the above problems,the main work of this paper is as follows:(1)To measure the privacy requirements of data,a privacy measurement algorithm based on mutual information was proposed.Firstly,the algorithm put forward a standard model to quantify privacy leakage based on the theory of information entropy,which can effectively calculate the amount of private information contained in data and measure the degree of privacy risk of sensitive information after public data was disclosed.At the same time,the algorithm used clustering entropy to adjust the value of sensitive attributes in equivalence classes after clustering,so that the value of sensitive attributes were distributed as evenly as possible in each equivalence class,and reduced the gap between public and sensitive attributes in a single equivalence class.Reduced the risk of sensitive data privacy leakage.(2)Using the calculation result of the data packet privacy weight in(1),designed a two-channel data dynamic encryption strategy.Selective encryption of data during transmission aimed to maximize the sum of data packet privacy weights within a limited time.First of all,the data packets were roughly classified according to the privacy weight of the data.Then,the weight ranking table was calculated by the data packet privacy weight and the encryption time and sorted in descending order,and the first packet was encrypted for transmission until at the end of the transmission time.Finally,check the remaining time inside the channel and adjust the transmission path of some data packets until the remaining time was less than the encryption time of any packet.Experimental results show that TDES can achieve higher efficiency in a shorter calculation time,and can well balance data security and equipment performance.