The traffic burden at each node in Internet-of-Things (IoT) communication networks becomes prohibitively high especially when involving exhaustive computation. Mobile edge computing (MEC) makes this complicated computation feasible while alleviates the traffic burden by providing the corresponding node with powerful computing resources through wireless transmission between the node and the MEC for offloading computation. However, the transmission via the varying wireless channel requires considerable energy consumption and imposes delay. In this paper, we study the trade-off between the energy consumption and the delay performance in IoT network due to the offloading computation and the wireless communication. An optimization problem involving the offloading ratio for MEC as unknown parameter is established by minimizing the total energy consumption subject to a delay constraint. The problem is then solved by analyzing the convexity of the cost function and the constraint. Moreover, the scaling law of both energy cost and delay performance of IoT networks is investigated with respect to the number of nodes employing the MEC. It is discovered that the delay performance decreases in the logarithm with increasing the number of nodes while the energy cost grows linearly with the increase of the number of nodes. Numerical simulations verifying the performances of the proposed method in the studied IoT networks with MEC are provided.

物联网（IoT）通信网络中每个节点的流量负担变得过高，尤其是在涉及穷举计算时。移动边缘计算（MEC）使这种复杂的计算成为可能，同时通过节点与MEC之间的无线传输为相应的节点提供强大的计算资源来减轻计算负担，从而减轻了流量负担。然而，经由变化的无线信道的传输需要相当大的能量消耗并且施加了延迟。在本文中，我们研究了由于卸载计算和无线通信而导致的物联网网络能耗与延迟性能之间的权衡。通过最小化受延迟约束的总能耗，建立了涉及将MEC的卸载比作为未知参数的优化问题。然后通过分析成本函数和约束的凸度来解决该问题。此外，针对采用MEC的节点数量，研究了IoT网络的能源成本和延迟性能的缩放定律。发现随着节点数量的增加，延迟性能对数下降，而能量成本随着节点数量的增加线性增加。数值仿真验证了所提方法在研究的物联网中利用MEC的性能。此外，针对采用MEC的节点数量，研究了IoT网络的能源成本和延迟性能的缩放定律。发现随着节点数量的增加，延迟性能对数下降，而能量成本随着节点数量的增加线性增加。通过数值仿真验证了所提方法在研究的物联网中利用MEC的性能。此外，针对采用MEC的节点数量，研究了IoT网络的能源成本和延迟性能的缩放定律。发现随着节点数量的增加，延迟性能以对数形式下降，而能量成本随着节点数量的增加而线性增长。数值仿真验证了所提方法在研究的物联网中利用MEC的性能。