Automating the Translation of Cloud Users’ High-Level Security Needs to an Optimal Placement Model in the Cloud Infrastructure
Amin Bagheri, Alireza Shameli-Sendi · IEEE Transactions on Services Computing · 2023
Deploying network security functions in the cloud to protect the application, dynamically and with different goals, has been one of the most important topics. An example of these functions can be firewall, DPI, or IDS, which can be placed between different layers of applications as needed. So far, many works have been done regarding the placement of network security functions in the cloud, considering various goals, such as minimizing the consumption of resources, energy consumption, and the cost of users or the service provider. One of the required tasks in this field is to automate the translation of users’ security needs into a placement model to be executed by a solver. In all works, this translation is not considered due to the lack of a model or pattern, and it is assumed that the user has a specific requirement, and as a result, more focus is placed on optimizing the placement or placing the functions in the right place. In order to execute the optimal placement algorithms, we need appropriate goals and related constraints. In this paper, for the first time, a model is proposed to translate high-level user needs to the goals and constraints of the optimal placement algorithm. To ensure the accuracy of the generated placement models, three different scenarios have been considered from the cloud user’s point of view, and the generated placement model has been implemented in the popular Fat-tree topology with the number of nodes from 36 to 540,800. The real scenarios presented in the paper and the accuracy of the output produced by automation model can be welcomed by cloud service providers.