Secure, Private and Fast Memory Allocation of Processes in Post-Quantum Clouds

OpenAlex İngilizce

Over the past two decades, clients have increasingly offloaded their processes, comprising both code and data, to powerful cloud infrastructures for execution. However, side-channel attacks targeting sensitive workloads, particularly allocation-based attacks that exploit process size information, pose a serious threat to data privacy in cloud environments. Consequently, both clients and cloud providers have strong incentives to conceal the actual sizes of their processes and available memory resources. To address this challenge, this paper first introduces a set of cryptographically secure building blocks and then proposes a novel secure and privacy-preserving memory allocation algorithm for cloud computing. The proposed algorithm ensures that both the client's process size and the cloud server's available memory blocks remain private during execution. Leveraging cryptographic primitives that are resistant to quantum attacks, the proposed solution is inherently suitable for the postquantum era. The security of the scheme is formally proven through rigorous mathematical analysis. Furthermore, theoretical evaluation and experimental results demonstrate that the proposed approach achieves superior security and privacy guarantees compared to existing methods in the literature, while maintaining practical efficiency.