The use of network applications to manage network operations by the controller in SDN architecture introduces a threat that makes the controller to be susceptible to several network attacks. This is possible because the network applications operate without any access control mechanism that authenticates or dictates what operations they can execute in the network. This consequently makes the network applications to take advantage of their ability to manipulate, change or modify network state to compromise network operations and resources. In order to address this problem this paper introduces a token-based authentication method that enables the controller to authenticate the various network applications. The application of this method builds an access permission zone where only legitimate network applications with the correct token credentials can have access to the network prior to implementing any network changes. This paper contributes in providing an authorisation method Boolean Access Matrix that enforces permission constraints on what the network applications can access or execute within the network. The authorisation method helps limits the unprecedented access the network applications have over the control layer resources, core services and the network operations. The paper introduces a novel method of evaluating the trust between the controller and the network application based on Subjective Logic Reasoning (SLR) which is a belief learning model. SLR is an advanced learning algorithm that is derived from Probability Calculus and Statistics. Experiments demonstrate the efficiency and scalability of the proposed algorithms in a large scale test environment.

SDN体系结构中的控制器使用网络应用程序管理网络操作会带来威胁，使控制器容易受到多种网络攻击。这是可能的，因为网络应用程序在运行时没有任何访问控制机制来验证或指示它们可以在网络中执行的操作。因此，这使网络应用程序可以利用其操纵，更改或修改网络状态以损害网络操作和资源的能力。为了解决这个问题，本文介绍了一种基于令牌的身份验证使控制器能够验证各种网络应用程序的方法。此方法的应用程序将建立一个访问许可区域，在该区域中，只有具有正确令牌凭证的合法网络应用程序才能在实施任何网络更改之前访问网络。本文有助于提供一种授权方法布尔访问矩阵，该矩阵对网络应用程序可以在网络中访问或执行的内容实施权限限制。授权方法有助于限制网络应用程序对控制层资源，核心服务和网络操作的前所未有的访问。本文介绍了一种新的基于控制器的信任度评估方法。主观逻辑推理（SLR）是一种信念学习模型。SLR是从概率微积分和统计派生的高级学习算法。实验证明了该算法在大规模测试环境中的效率和可扩展性。