集中与分布：协同虚拟网络映射

[摘要] 结合传统集中式和分布式两类算法各自的特性，提出了协同虚拟网络映射算法。该算法保留了集中式算法中拥有全局视野的中心控制实体，负责总体控制和关键决策，同时将具体映射方案的计算过程交给有限的底层网络子集实现；唯一的中心控制实体与多个底层节点相互配合协作，共同完成虚拟网络映射的整个过程。该算法继承了集中式和分布式算法各自的优势，有效弥补了二者的缺陷，初步的仿真试验也证明了其可行性和有效性。

[关键词] 网络虚拟化；虚拟网络映射；集中；分布；协同虚拟网络映射

[Abstract] This article proposes a cooperative virtual network embedding algorithm that leverages traditional centralized and distributed algorithms. Cooperative virtual network embedding keeps the central control entity with a global vision in centralized algorithms in charge of general control and vital decisions. At the same time, it trusts a set of limited substrate nodes with detailed embedding solutions. A central control entity and multiple substrate nodes cooperate closely to achieve successful virtual network embedding. The cooperative algorithm incorporates the advantages of both centralized and distributed algorithms and also avoids their shortcomings. Elementary simulations prove the practicability and effectiveness of this algorithm.

[Keywords] network virtualization; virtual network embedding; centralization; distribution; cooperative virtual network embedding