Research On Way Of Evaluating Cloud End User Behavior's Credibility Based On Ordinal Regression

As the cloud computing is used more and more widely in the market,some kinds of mode application which are similar with cross-domain resource sharing application modes appears constantly,with the problems of credibility and security in cloud computing increasing prominently.Network has the characteristic such as,open,dynamic and resource growth,autonomy,and diversity,all of which make the credibility of the cloud computing environment be affected by the randomness,fuzziness,unpredictability and various uncertainties.The development of cloud computing technology and the popularization of cloud services have a great impact on the industry.On the one hand,cloud technology enhances network's operation efficiency and reduces the cost.On the other hand,the cloud resource can be accessed by any network equipment.It increases the chances that the identity of user is misrepresented and then led to many security problems.Therefore,the actual needs of security can't be fully satisfied with controlling the malicious user access to the cloud resource by login authentication that relies solely on current user identity.User is the requester and provider of cloud resources.User behavior's credibility relates to the safety of cloud directly.So it's very important to evaluate whether user behaviors can be trusted or not on cloud.The factors which affect user behavior credibility are thoroughly discussed by analyzing Cloud terminal user behavior data,which is followed with the establishment of indexes system for cloud end user's behavior.The way of factor analysis is also utilized in order to further get a small number of unrelated comprehensive indexes based on reducing-dimension processing.Finally,the credibility evaluation of cloud end user's behavior research is realized by ordinal regression.In order to get the points of the evaluation theory method of cloud end user's behavior credibility,plenty of literature researches are adopted in the early stage.Then,based on the proposed evaluation model and the condition which have been modified by eliminating system of omission and false positives,user's behavior credibility is assessed smoothly by combined with user's historical behavior and the current behavior.The evaluation result of user's behavior credibility plays a double role.It helps system to judge whether to refuse the user's access directly or to monitor user to check again the credibility of identity quickly to make sure that legitimate accounts of end cloud users have not been used illegally,etc.This way benefits to cloud service providers through the way of protecting the security of cloud platform,improving the identification of the end user's credibility and reducing the overall cost of monitoring all the cloud end user's behaviors which is caused by lower distrust.