Outsourcing is an essential strategy for enterprises, and leveraging it can offer advantages such as saving costs, improving efficiency, and allowing them to concentrate on their primary business activities. Therefore, announcing and managing service requests is paramount for obtaining suitable quotes and partnering with the right outsourcing service providers. To achieve this, enterprises can employ digital signatures to ensure the authenticity and security of these service requests. However, the irrevocable nature of traditional digital signatures poses challenges, particularly when canceling the service requests is necessary. A withdrawable signature scheme provides a novel property for signers to initially create “withdrawable” signatures, which they can later confirm into additional conventional signatures. Withdrawable signatures can be regarded as withdrawn when the signer has not confirmed them yet. This property provided by the withdrawable signature mechanism can then be adopted as a solution to this problem, allowing the enterprise to retract announced signatures on service requests. However, to meet the demands in the outsourcing system, existing approaches on withdrawable signatures require further development since they only allow signers in outsourcing systems to confirm withdrawable signatures on service requests. This limitation prevents enterprises from accepting quotes for service requests simply by confirming the withdrawable signature. This paper focuses on this problem by applying the withdrawable signature to the outsourcing framework, identifying the limitations of existing withdrawable signature schemes, and proposing solutions. We demonstrate how revisiting of withdrawable signatures can lead to flexible outsourcing systems that ensure the cancellation of announced service requests. Additionally, we provide a performance evaluation demonstrating that our revised withdrawable signature scheme achieves acceptable efficiency and security within the outsourcing system.

With the appearance and flourishing development of the Internet of Things (IoT), wireless sensor networks technology has been attracting increasing attention. Network coding is an indispensable technology in the wireless sensor networks, which can improve network transmission throughput. However, pollution attacks is a serious security problem that must be faced in the process of data coding. Although the homomorphic network coding signature schemes can solve this troublesome, the high signature generation and verification cost of these schemes will reduce the transmission efficiency. In this article, we propose an efficient identity-based linearly homomorphic network coding signature scheme for wireless sensor networks to guarantee data integrity and authenticity. In our scheme, the computation cost of signature generation and verification are both independent of the size of the data packet. The scheme is proved secure against existential forgery under adaptive chosen identity and adaptive chosen subspace attacks in random oracle model. Using Java pairing-based Cryptography Library (JPBC), the simulation results illustrate that our scheme is more efficient in practical application.

Edge Artificial Intelligence (AI) is a timely complement of cloud-based AI. By introducing intelligence to the edge, it alleviates privacy concerns of streaming and storing data to the cloud, enables real-time operations where milliseconds matter, and brings AI services to remote areas with poor networking infrastructures. Security is a significant problem in Edge AI applications such as self-driving cars and intelligent healthcare. Since the edge devices are empowered to process data and take actions, attacking and compromising them can cause serious damage. However, the wide deployment of computationally limited devices in edge environments and the increasing happening of side-channel (or leakage) attacks pose critical challenges to security. This article thereby aims to enhance the security for Edge AI by designing and developing lightweight and leakage-resilient authenticated key exchange (LRAKE) protocols. Compared with available LRAKE protocols, the proposed protocols in this article can be effortless applied in some mainstreaming security and communication standards. Moreover, this article realizes prototypes and presents implementation details; and a use case of applying the proposed protocol in Bluetooth 5.0 is illustrated. The theoretical design and implementation details will provide a guidance of applying the LRAKE protocols in Edge AI applications.