Distributed Placement of Virtual Network Functions in Edge Networks

随着网络的日益复杂，手工部署和管理变得越来越困难。自动化部署和管理是必要的，可以及时响应用户需求。本研究提供了与NFV编排引擎协同工作的资源分配模型，提高了网络性能。研究的重点是对未知故障的容错。提出了静态和动态两种模型。所提模型提供的分配保证了在网络中有k个失效节点的情况下，处理能力仍满足要求。数值结果表明，该模型降低了vnf之间的端到端延迟，增加了可接受需求的数量，保证了功能的弹性。

Manual deployment and management are becoming overwhelming with the increasing complexity of networks. Automated deployment and management are imperative and can respond to user requirements promptly. This research provides resource allocation models that cooperated with the NFV orchestration engine and improve the network performance.The research focuses on fault tolerances for unknown failures. Two models for static and dynamic scenarios are proposed. The allocation provided by the proposed models ensures the processing ability satisfies the requirements even though there are k failed nodes in the network. The numerical results show that the proposed models reduce the E2E latency between the VNFs and increase the number of accepted requirements ensuring the resiliency of the functions.