Cloud computing is characterized by strong computing and storage capabilities, and edge computing has the advantages of low latency and low power consumption. Many service providers have begun to combine the advantages of cloud and edge computing to provide better quality of service, but the heterogeneity of cloud and edge computing represents a challenge for service deployment and resource allocation. This paper proposes a framework for cloud-edge collaboration based on live video webcast services and transforms the resource allocation problem into a constrained integer programming (IP) model. Additionally, we introduce an auction mechanism to solve the problem of resource competition among the anchor users in live services. By solving the IP resource allocation problem and Vickrey–Clarke–Groves mechanism, we obtain the optimal resource allocation mechanism. Based on the dominant resource proportion and matching model, we design a resource allocation mechanism for the online environment. These mechanisms can be used for reservation and live webcast scenarios. Furthermore, we prove that the two mechanisms have individual rationality and truthfulness. Our approach is characterized by high social welfare, high resource utilization and a short execution time.

云计算的特点是计算和存储能力强，边缘计算具有低延迟、低功耗的优势。许多服务提供商已经开始结合云计算和边缘计算的优势来提供更好的服务质量，但是云计算和边缘计算的异构性对服务部署和资源分配提出了挑战。本文提出了一种基于实时视频网络广播服务的云边缘协作框架，并将资源分配问题转化为约束整数规划（IP）模型。此外，我们引入了拍卖机制来解决直播服务中主播用户之间的资源竞争问题。通过解决 IP 资源分配问题和 Vickrey-Clarke-Groves 机制，我们获得了最优的资源分配机制。基于优势资源比例和匹配模型，我们设计了一种在线环境的资源分配机制。这些机制可用于预订和网络直播场景。此外，我们证明了这两种机制具有个体的合理性和真实性。我们的方法具有社会福利高、资源利用率高和执行时间短的特点。