The workload generated by the Internet of Things (IoT)-based infrastructure is often handled by the cloud data centers (DCs). However, in recent time, an exponential increase in the deployment of the IoT-based infrastructure has escalated the workload on the DCs. So, these DCs are not fully capable to meet the strict demand of IoT devices in regard to the lower latency as well as high data rate while provisioning IoT workloads. Therefore, to reinforce the latency-sensitive workloads, an intersection layer known as edge computing has successfully balanced the entire service provisioning landscape. In this IoT-edge-cloud ecosystem, large number of interactions and data transmissions among different layer can increase the load on underlying network infrastructure. So, software-defined edge computing has emerged as a viable solution to resolve these latency-sensitive workload issues. Additionally, energy consumption has been witnessed as a major challenge in resource-constrained edge systems. The existing solutions are not fully compatible in Software-defined Edge ecosystem for handling IoT workloads with an optimal trade-off between energy-efficiency and latency. Hence, this article proposes a lightweight and energy-efficient container-as-a-service (CaaS) approach based on the software-define edge computing to provision the workloads generated from the latency-sensitive IoT applications. A Stackelberg game is formulated for a two-period resource allocation between end-user/IoT devices and Edge devices considering the service level agreement. Furthermore, an energy-efficient ensemble for container allocation, consolidation and migration is also designed for load balancing in software-defined edge computing environment. The proposed approach is validated through a simulated environment with respect to CPU serve time, network serve time, overall delay, lastly energy consumption. The results obtained show the superiority of the proposed in comparison to the existing variants.

基于物联网（IoT）的基础架构生成的工作负载通常由云数据中心（DC）处理。但是，近来，基于物联网的基础架构的部署呈指数级增长，已经加剧了DC的工作量。因此，这些DC不能完全满足IoT设备的严格要求，因为它们在配置IoT工作负载时具有较低的延迟以及较高的数据速率。因此，为了加强对延迟敏感的工作负载，称为边缘计算的相交层已成功平衡了整个服务供应格局。在这个物联网边缘云生态系统中，不同层之间的大量交互和数据传输会增加基础网络基础架构的负载。所以，软件定义的边缘计算已成为解决这些对延迟敏感的工作负载问题的可行解决方案。此外，在资源受限的边缘系统中，能耗被视为主要挑战。现有解决方案在软件定义的边缘生态系统中无法完全兼容，无法以能效与延迟之间的最佳平衡来处理物联网工作负载。因此，本文提出了一种基于软件定义的边缘计算的轻量级节能型容器即服务（CaaS）方法，以提供对延迟敏感的IoT应用程序生成的工作负载。考虑到服务水平协议，为最终用户/ IoT设备与Edge设备之间的两阶段资源分配制定了Stackelberg游戏。此外，还有一个用于分配容器的节能组合，整合和迁移还旨在在软件定义的边缘计算环境中实现负载平衡。通过关于CPU服务时间，网络服务时间，总延迟，最后的能耗的模拟环境，对所提出的方法进行了验证。所获得的结果表明，与现有变体相比，该提议具有优越性。