Proposed Improving Protection of Cloud Computing Environments Based on Machine Learning Techniques

Abstract

Cloud computing contains a huge amount of data, which makes it a common target for cyberattacks to access confidential data using various illegal methods. One of these attacks is the Structure Query Language injection attack SQLIA, It is categorised as one of the most prevalent threats to obtaining, modifying, or destroying data by the Open Web Application Security Project (OWASP). Therefore, it has become necessary to create a model to detect attacks on data on the cloud to protect it, in order to increase trust between individuals and institutions and not make this data available to people who are not authorized to access it. To solve these problems, this study presents a proposal to improve the protection of the cloud computing environment through two contributions.The first contribution is developing a machine learning model, known as a logistic regression framework, that serves as a mediator between the client and the server. Its goal is to ascertain the kind of requests that are received from the customer layer and whether or not they include hazardous or typical payloads. The second contribution is illustrating how dangerous it is for the cloud computing infrastructure and for consumers and organisations to rely on false forecasts regarding the confidentiality, integrity, and real-time availability of data.. The results obtained from applying the proposed model showed a very high accuracy of 99.82, and showed low rates of false negatives and positives. In addition, the time it takes to determine the type of request sent is 0.1514 seconds.