Device-to-device (D2D) communication technique is used to establish direct links among mobile devices (MDs) to reduce communication delay and increase network capacity over the underlying wireless networks. Existing D2D schemes for task offloading focus on system throughput, energy consumption, and delay without considering data security. This paper proposes a Security and Energy-aware Collaborative Task Offloading for D2D communication (Sec2D). Specifically, we first build a novel security model, in terms of the number of CPU cores, CPU frequency, and data size, for measuring the security workload on heterogeneous MDs. Then, we formulate the collaborative task offloading problem that minimizes the time-average delay and energy consumption of MDs while ensuring data security. In order to meet this goal, the Lyapunov optimization framework is applied to implement online decision-making. Two solutions, greedy approach and optimal approach, with different time complexities, are proposed to deal with the generated mixed-integer linear programming (MILP) problem. The theoretical proofs demonstrate that Sec2D follows a $[O(1∕V),O\left(V\right)]$ energy-delay tradeoff. Simulation results show that Sec2D can guarantee both data security and system stability in the collaborative D2D communication environment.

设备到设备（D2D）通信技术用于在移动设备（MD）之间建立直接链接，以减少通信延迟并增加基础无线网络上的网络容量。现有的用于任务卸载的D2D方案着重于系统吞吐量，能耗和延迟，而没有考虑数据安全性。本文提出了一种用于D2D通信（Sec2D）的安全和能源感知协作任务分载。具体来说，我们首先根据CPU内核数，CPU频率和数据大小建立了一个新颖的安全模型，用于测量异构MD上的安全工作量。然后，我们制定了协作任务卸载问题，该问题可在确保数据安全性的同时将MD的平均时间延迟和能耗降至最低。为了达到这个目标，Lyapunov优化框架用于实施在线决策。提出了两种求解方法，分别是贪婪法和最优法，它们具有不同的时间复杂度，以解决生成的混合整数线性规划（MILP）问题。理论证明表明Sec2D遵循$[Ø（\mathrm{1个}∕V），Ø（V）]$能量延迟权衡。仿真结果表明，Sec2D在协同D2D通信环境中可以保证数据安全性和系统稳定性。