The convergence of software‐defined networking (SDN) and the Internet of Things (IoT) provides a scalable method for handling the considerable volumes of data produced by IoT devices. However, the lack of appropriate security measures can lead to unauthorized access to sensitive data, potential breaches, and privacy violations, as well as time‐consuming and inefficient data retrieval methods in SDN‐IoT systems that require decrypting the entire dataset. To address these challenges, this article proposes the time‐attribute‐based access control scheme in SDN‐IoT (TAAC). The TAAC scheme combines ciphertext‐policy attribute‐based encryption with a novel time‐attribute‐based access tree to ensure fine‐grained access control on time and attributes, enabling secure ciphertext interaction and information sharing across domains. Furthermore, the TAAC scheme also incorporates searchable encryption, which enhances the efficiency of data retrieval. By implementing searchable encryption techniques, the data receiver can generate trapdoors to search and retrieve specific encrypted data without the need to decrypt the entire dataset. In summary, the TAAC scheme improves storage efficiency and computation, enhances scalability, and provides robust security, offering an efficient and secure solution for ciphertext sharing in SDN‐IoT environments. Experimental results have demonstrated that the TAAC scheme shows excellent performance and outperforms other attribute‐based searchable encryption algorithms.

The Software-Defined Internet of Things (SD-IoT) utilizes the centralized control and programmability of Software-Defined Networking (SDN) to enhance network performance optimization and efficient resource utilization in IoT. As the network scale expands, the multiple-controller architecture becomes crucial for ensuring reliability and scalability in SD-IoT. However, the dynamic changes in traffic patterns often lead to imbalanced loads among controllers. Existing solutions primarily focus on switch migration schemes, but traditional schemes primarily rely on real-time data to assess controller loads, which fails to predict future controller loads and leads to unnecessary switch migrations. Meanwhile, existing schemes frequently encounter the challenge of overloading the target controller, leading to reduced migration efficiency. Furthermore, conventional schemes tend to overlook the issue of isolated nodes that arise from switch migrations, thereby compromising network reliability and security. To address these challenges, we propose bidirectional switch migration based on load prediction (BSM-LP), which utilizes an ATT-GRU model to accurately predict controller loads based on historical load data, thereby preventing unnecessary switch migrations. Moreover, we introduce a bidirectional switch migration algorithm that enhances migration efficiency while avoiding overloading the target controller. Additionally, we present an algorithm for identifying and integrating isolated nodes to reduce their occurrence. Finally, we validate the effectiveness of BSM-LP, and the experimental results demonstrate that it reduces the load imbalance rate by an average of 22.3% and the response time by 30.5% compared to existing schemes.

The proliferation of cloud computing technology has led to a significant number of users opting to store image data on the cloud. To ensure the confidentiality of image data, it is recommended to apply encryption techniques prior to uploading them to cloud servers. However, the traditional encrypted image retrieval schemes prove inadequate for practical application scenarios due to limitations in supporting multi-owner and multi-user settings, inefficient access control, and a lack of dynamic verifiability. To address the challenges presented by practical application scenarios, in this paper, we propose a Dynamic and Controllable Image Retrieval scheme in the Multi-owner multi-user settings (DCIRM). First of all, we realize the usability in multi-owner and multi-user scenarios through re-encryption settings. Then, we realize the verifiability of dynamic data by applying the dynamic authentication constructed by the chameleon hash function. Finally, we realize lightweight access control through the polynomial-based access strategy. The DCIRM scheme has been demonstrated to be privacy-preserving, efficient, and feasible through rigorous security analysis and experimentation with authentic datasets.