Abstract With the emergence of fog computing, new paradigms for data processing and management for IoT devices have been established in the quickly changing world of teaching/learning. This study addresses the complex issues brought about by the infiltration of diverse data sources by investigating novel approaches to strengthen data security and enhance access control mechanisms in fog computing environments. The commonly used cryptographic technique known as CP‐ABE is renowned for providing accurate access control. Unfortunately, current multi‐authority CP‐ABE methods have difficulties when implemented on low‐resource IoT devices. These techniques are not appropriate for resource‐constrained IoT devices since the sizes of the secret key and ciphertext grow in proportion to the number of attributes. In this paper, a novel multi‐authority CP‐ABE approach, called MA‐based CP‐ABE, efficiently tackles these issues by optimizing the length of secret keys and ciphertext. Users' secret keys are always the same size, no matter how many attributes they own. Moreover, MA‐based CP‐ABE ensures that the size of the ciphertext scales linearly with the number of authorities rather than characteristics, which makes it a sensible option for devices with restricted resources. A Generative AI approach has also been integrated along with CP‐ABE to make sure that the IoT data is secure and privacy is maintained. As per the security and experimental analysis, the proposed approach is considered secure and suitable for IoT‐based applications.

It is well known that cognitive radio (CR) techniques have great potential to deal with the problem of radio spectrum scarcity. Spectrum sensing technique plays a critical role in enabling unlicensed secondary users (SUs) to utilize spectrum holes in cognitive radio networks (CRNs). However, a licensed primary user (PU) can be appropriately protected by simultaneously performing spectrum sensing and data transmission i.e., by using full-duplex (FD) mode. In this paper, we have proposed and analyzed a strategy which includes spectrum sensing and spectrum access mechanisms both. We consider heterogeneous traffic of real-time and non real-time SUs based on their different delay tolerance characteristics. We address the issue of false alarm rate (FAR) associated with FD sensing. Spectrum handoff and call buffering strategies with$p$-Retry policy are employed jointly so that SUs that would otherwise be blocked or forcibly dropped could be buffered and possibly served later. To evaluate the performance of the proposed strategy, five-dimensional continuous time Markov chain (CTMC) model is developed and the queueing-theoretic approach is utilized. Numerical results demonstrate the influence of spectrum sensing errors on the performance of such CRNs. Results also reveal that the provision of buffers under retrial policy increases the overall network resource utilization while decreasing blocking and dropping probabilities.