首页|基于延迟和抖动感知的多播服务功能树嵌入算法

基于延迟和抖动感知的多播服务功能树嵌入算法

扫码查看
针对软件定义网络/网络功能虚拟化(SDN/NFV)架构中,多播请求流(MRs)需满足严格时延和抖动约束下遍历由多个虚拟网络功能(VNFs)依序组成的服务功能树(SFT)问题.该文提出一种基于最优链路选择函数进行深度优先搜索构建SFT的路由算法.首先,提出网络资源相对成本函数,以保证网络负载自动均衡.其次,联合考虑网络资源、VNF动态放置及多播流延迟和抖动约束,构建SFT动态嵌入问题的整数线性规划模型(ILP).最后,针对该NP难问题,设计辅助边权图和最优链路选择函数进行路由路径选择,并以最小化资源消耗成本为目标提出具有延迟和抖动感知的SFT嵌入算法(SFT-EA).仿真结果表明,SFT-EA在吞吐量,流接受率和网络负载均衡方面具有更好的性能.
Multicast Service Function Tree Embedding Algorithm Based on Delay and Jitter Awareness
To solve the problem that Multicast Request flows(MRs)need to traverse sequentially a Service Function Tree(SFT)consisting of Virtual Network Functions(VNFs)as well as ensuring stringent delay and jitter constraints of SFT in Network Function Virtualization(NFV)-enabled Software-Defined Networks(SDNs),a routing algorithm for constructing a multicast SFT based on depth-first search with an optimal link selection function is proposed.Firstly,the relative cost functions of network resources are proposed to guarantee the automatic load balancing of the network.Secondly,an Integer Linear Programming model(ILP)for the SFT dynamic embedding is constructed by jointly considering network resources,VNF dynamic placement and delay and jitter constraints of a multicast flow.Finally,for this NP-hard problem,an auxiliary edge-weight graph and optimal link selection function are designed for routing path selection,and a delay and jitter-aware SFT Embedding Algorithm(SFT-EA)is proposed with the objective of minimizing the resource consumption cost.Simulation results demonstrate the SFT-EA has better performance in terms of throughput,traffic acceptance rate,and network load balance.

Network Function Virtualization(NFV)Service Function Tree(SFT)MulticastDelay and Jitter

刘亮、陈翔、桂晓菁、徐勇军、杜娅荣、侯泽天、段洁

展开 >

重庆邮电大学通信与信息工程学院 重庆 400065

重庆邮电大学自动化学院 重庆 400065

网络功能虚拟化 服务功能树 多播 延迟和抖动

国家自然科学基金国家自然科学基金重庆邮电大学博士启动基金

6217107061701058A2023007

2024

电子与信息学报
中国科学院电子学研究所 国家自然科学基金委员会信息科学部

电子与信息学报

CSTPCD北大核心
影响因子:1.302
ISSN:1009-5896
年,卷(期):2024.46(1)
  • 3