Dynamic Network Slice Migration Algorithm Based on Ensemble Deep Neural Network Traffic Prediction
-
摘要: 针对5G网络切片(NS)场景下由于缺乏提前对物理网络资源进行感知而导致切片迁移滞后的问题，该文提出一种基于集成深度神经网络流量预测的动态切片调整和迁移算法(DSAM)。首先建立了基于计算、内存、带宽资源配置的网络总惩罚模型；其次，提出基于集成深度神经网络的流量预测算法预测未来网络流量情况，并根据流量类型的不同将其转换成对未来时刻物理网络的资源占用及切片的资源需求感知；最后，根据感知结果，以尽可能大地降低运营商惩罚为目标，通过动态切片调整和迁移策略将虚拟网络功能(VNF)和虚拟链路迁移到满足资源限制的物理节点和链路上。仿真结果表明，所提算法有效提高了切片迁移的效率和网络资源利用率。Abstract: In order to solve the problem that slice migration lags behind by lacking awareness of physical network resources in 5G Network Slice (NS) scenarios, a Dynamic Slice Adjustment and Migration (DSAM) algorithm based on ensemble deep neural network traffic prediction is proposed. Firstly, a network total penalty model based on computing, memory and bandwidth resource allocation is established. Secondly, in order to predict the future traffic situation, a prediction algorithm based on ensemble deep neural network is proposed. Then the result of prediction are converted to perception of the physical network resource usage and resource requirements of slice in future according to the different types of traffic. Finally, in order to as large as possible to reduce operators punishment according to the result of perception, Virtual Network Functions (VNF) and virtual links are migrated to physical nodes and links that meet resource limits through dynamic slice adjustment and migration policies. The simulation results show that the proposed algorithm improves effectively the efficiency of slice migration and utilization of network resources.
-
算法1 基于流量预测的动态切片调整和迁移算法(DSAM) 输入：服务请求$ R $，底层物理网络$ \bar G $ 输出：每条服务功能链调整迁移后的位置 初始化： (1) for $ {R_k} \in R $ do (2) 根据KSP算法原则计算所有源节点和目的节点间VNF组成
的最短路径$ L $(3) for $ l \in L $ do (4) 选择一条路径$ l $，沿$ l $映射服务请求$ R $的虚拟链路 (5) 选择$ l $中的第1个物理节点，沿$ l $映射服务请求$ R $的虚拟节点 切片调整和迁移： (6) for $ t \in T $ do (7) for $ {R_k} \in R $ do (8) if $ {A_{{R_k}(t + 1)}} > {A_{{R_k}(t)}} $ then (9) 计算$ {R_k}(t + 1) $中$ \{ C_k^m(t + 1),M_k^m(t + 1),B_k^i(t + 1)\} $
的值
(10) $ \{ \hat C_k^m(t + 1),\hat M_k^m(t + 1),\hat B_k^i(t + 1)\} $
$ = \{ C_k^m(t + 1),M_k^m(t + 1),B_k^i(t + 1)\} $(11) if C3～C5成立 then (12) 输出$ \{ \delta _k^{n,m}(t + 1),\eta _k^{f,i}(t + 1)\} $ (13) else (14) if $ l > 1 $ then (15) 计算所有$ l $的$ \{ \overline {{B_l}} (t + 1),\overline {{C_l}} (t + 1),\overline {{M_l}} (t + 1)\} $ (16) 将$ \{ \overline {{B_l}} (t + 1),\overline {{C_l}} (t + 1),\overline {{M_l}} (t + 1)\} $按升序排列 (17) for $ {\text{MAX}}\{ \overline {{B_l}} (t + 1),\overline {{C_l}} (t + 1),\overline {{M_l}} (t + 1)\} $ do (18) 选择$ l $上的第1个物理节点
(19) $\{ \hat C_k^m(t + 1),\hat M_k^m(t + 1),\hat B_k^i(t + 1)\} $
$ = \{ C_k^m(t + 1),M_k^m(t + 1),B_k^i(t + 1)\} $(20) if C3～C5成立 then (21) 输出$ \{ \delta _k^{n,m}(t + 1),\eta _k^{f,i}(t + 1)\} $ (22) else (23) 降级切片，根据式(7)计算路径$ l $上的服务降
级惩罚$ {P_l} $(24) 将$ {P_l} $按升序排列 (25) 切片$ k $迁移到$ {P_{\min }} $上 (26) 输出$ \{ \delta _k^{n,m}(t + 1),\eta _k^{f,i}(t + 1)\} $ (27) end if (28) end for (29) else (30) 降级切片，根据式(7)计算路径$ l $上的服务降级惩罚$ {P_l} $ (31) end if (32) end if (33) 根据$ t + 1 $时刻预测流量值进行资源调整 (34) else (35) 释放相应资源 (36) end if (37) end for (38) end for 
下载: 导出CSV
-
[1] GHONGE M M, MANGRULKAR R, PJAWANDHIYA P M, et al. Future Trends in 5G and 6G: Challenges, Architecture, and Applications[M]. CRC Press, 2021. [2] DEBBABI F, JMAL R, FOURATI L C, et al. Algorithmics and modeling aspects of network slicing in 5G and Beyonds network: Survey[J]. IEEE Access, 2020, 8: 162748–162762. doi: 10.1109/ACCESS.2020.3022162 [3] POZZA M, NICHOLSON P K, LUGONES D F, et al. On reconfiguring 5G network slices[J]. IEEE Journal on Selected Areas in Communications, 2020, 38(7): 1542–1554. doi: 10.1109/JSAC.2020.2986898 [4] LI D F, PHONG P L, XUE K P, et al. Virtual network function placement considering resource optimization and SFC requests in cloud datacenter[J]. IEEE Transactions on Parallel and Distributed Systems, 2018, 29(7): 1664–1677. doi: 10.1109/TPDS.2018.2802518 [5] SCIANCALEPORE V, COSTA-PEREZ X, and BANCHS A. RL-NSB: Reinforcement learning-based 5G network slice broker[J]. IEEE/ACM Transactions on Networking, 2019, 27(4): 1543–1557. doi: 10.1109/TNET.2019.2924471 [6] XIAO Suchao and CHEN Wen. Dynamic allocation of 5G transport network slice bandwidth based on LSTM traffic prediction[C]. 2018 IEEE 9th International Conference on Software Engineering and Service Science (ICSESS), Beijing, China, 2018: 735–739. [7] SONG Chuang, ZHANG Min, HUANG Xuetian, et al. Machine learning enabling traffic-aware dynamic slicing for 5G optical transport networks[C]. 2018 Conference on Lasers and Electro-Optics (CLEO), San Jose, USA, 2018: 1–2. [8] 唐伦, 赵培培, 赵国繁, 等. 基于深度信念网络资源需求预测的虚拟网络功能动态迁移算法[J]. 电子与信息学报, 2019, 41(6): 1397–1404. doi: 10.11999/JEIT180666TANG Lun, ZHAO Peipei, ZHAO Guofan, et al. Virtual network function migration algorithm based on deep belief network prediction of resource requirements[J]. Journal of Electronics &Information Technology, 2019, 41(6): 1397–1404. doi: 10.11999/JEIT180666 [9] LI Taihui, ZHU Xiaorong, and LIU Xu. An end-to-end network slicing algorithm based on deep Q-learning for 5G network[J]. IEEE Access, 2020, 8: 122229–122240. doi: 10.1109/ACCESS.2020.3006502 [10] WEI Fengsheng, FENG Gang, SUN Yao, et al. Proactive network slice reconfiguration by exploiting prediction interval and robust optimization[C]. GLOBECOM 2020 - 2020 IEEE Global Communications Conference, Taipei, China, 2020: 1–6. [11] CHERGUI H and VERIKOUKIS C. Offline SLA-constrained deep learning for 5G networks reliable and dynamic end-to-end slicing[J]. IEEE Journal on Selected Areas in Communications, 2020, 38(2): 350–360. doi: 10.1109/JSAC.2019.2959186 [12] KAO C C, CHANG C W, CHO C P, et al. Deep learning and ensemble learning for traffic load prediction in real network[C]. 2020 IEEE Eurasia Conference on IOT, Communication and Engineering (ECICE), Yunlin, China, 2020: 36–39. [13] YIN Shan, ZHANG Zhan, YANG Chen, et al. Prediction-based end-to-end dynamic network slicing in hybrid elastic fiber-wireless networks[J]. Journal of Lightwave Technology, 2021, 39(7): 1889–1899. doi: 10.1109/JLT.2020.3045600 [14] YU Hao, MUSUMECI F, ZHANG Jiawei, et al. Dynamic 5G RAN slice adjustment and migration based on traffic prediction in WDM metro-aggregation networks[J]. Journal of Optical Communications and Networking, 2020, 12(12): 403–413. doi: 10.1364/JOCN.403829 [15] Small Cell Forum. Small cell virtualization: Functional splits and use cases[R]. Technical Report of Small Cell Forum, 2016. -
图(8) / 表(1)
计量
- 文章访问数: 949
- HTML全文浏览量: 311
- PDF下载量: 141
- 被引次数: 0
下载:
