一種支持硬件加速的虛擬網絡功能部署模型

胡宇翔; 范宏偉; 蘭巨龍; 段通

doi:10.11999/JEIT180861

一種支持硬件加速的虛擬網絡功能部署模型

doi: 10.11999/JEIT180861

國家數(shù)字交換系統(tǒng)工程技術研究中心 ??鄭州 ??450002

基金項目: 國家網絡空間安全專項基金(2017YFB0803204)，國家自然科學基金(61521003)

詳細信息

作者簡介:
胡宇翔：男，1982年生，副教授，研究方向為未來網絡關鍵技術和網絡智慧化

范宏偉：男，1994年生，碩士生，研究方向為網絡功能虛擬化和硬件加速

蘭巨龍：男，1962年生，教授，博士生導師，研究方向為未來通信網絡關鍵理論與技術

段通：男，1992年生，博士生，研究方向為網絡功能虛擬化和可編程硬件

通訊作者:
范宏偉　fhwxd@foxmail.com

中圖分類號: TP393
計量
- 文章訪問數(shù): 2970
- HTML全文瀏覽量: 860
- PDF下載量: 63
- 被引次數(shù): 0
出版歷程
- 收稿日期: 2018-09-04
- 修回日期: 2019-01-04
- 網絡出版日期: 2019-01-16
- 刊出日期: 2019-08-01

A Model for Virtualized Network Function Placement with Hardware Acceleration Support

National Digital Switching System Engineering & Research Center, Zhengzhou 450002, China

Funds: The National Network Security Special Program of China (2017YFB0803204), The National Natural Science Foundation of China (61521003)

摘要

摘要: 為解決以軟件實現(xiàn)的虛擬網絡功能(VNF)性能受限問題，軟件定義網絡和網絡功能虛擬化(SDN/NFV)等新型網絡架構引入了硬件加速資源。硬件加速資源的部署，使得VNF能夠為日益增長的數(shù)據流量提供服務保障。該文針對已有研究未考慮具有高性能數(shù)據處理需求的服務鏈VNF部署問題，提出一種支持硬件加速的VNF部署模型。該模型基于硬件加速資源的承載特性，在保證未加速VNF到商用服務器的優(yōu)化部署下，優(yōu)先實現(xiàn)交換機中加速資源的復用，并根據網絡業(yè)務的性能需求，靈活調整加速資源與VNF的映射約束。仿真實驗表明，與其他典型部署方法相比，在引入相同硬件加速資源的情況下，該模型可以承載更多的業(yè)務流量，滿足服務鏈高性能數(shù)據處理需求，有效提高了部署在網絡中加速硬件的資源利用率。
- 軟件定義網絡 /
- 網絡功能虛擬化 /
- 虛擬網絡功能部署 /
- 硬件加速
Abstract: In order to deal with the limited capacity of Virtualized Network Function (VNF), hardware acceleration resources are adopted in Software-Defined Networking and Network Function Virtualization (SDN/NFV) architecture. The deployment of hardware acceleration resources enables VNF to provide service guarantees for increasing data traffic. To overcome the ignorance of the requirements for VNF with high processing throughput in service chain in existing researches, a model for VNF placement with hardware acceleration support is proposed. Based on the bearing characteristics of hardware acceleration resources, the model prioritizes the reuse of acceleration resources in the switch under the optimal placement of VNF without acceleration to commercial servers. The mapping correlation between hardware acceleration resources and VNF is flexibly adjusted according to the requirements of network services. Simulation results show that the proposed model can bear more service flows and meet the high processing throughput needs of service chains than typical policies in the case of the same amount of resources, which improves effectively the resource utilization of the acceleration hardware deployed in the network.
- Software-Defined Network (SDN) /
- Network Function Virtualization (NFV) /
- Virtualized Network Function (VNF) placement /
- Hardware acceleration

HTML全文

圖 1 硬件加速資源對網絡功能的承載

下載: 全尺寸圖片幻燈片

圖 2 VPHA底層網絡示意

下載: 全尺寸圖片幻燈片

圖 3 算法仿真網絡拓撲

下載: 全尺寸圖片幻燈片

圖 4 算法仿真性能對比

下載: 全尺寸圖片幻燈片

圖 5 模型驗證示意

下載: 全尺寸圖片幻燈片

圖 6 數(shù)據包轉發(fā)性能對比

下載: 全尺寸圖片幻燈片

表 1 HMRP算法

　輸入：服務請求集合D，底層網絡G，VNF參數(shù)集合F

　輸出：VNF部署方案P_d

　(1) 根據服務請求${d_i}$，更新底層網絡$G\ '$

　(2)　for x←1 to |S| do //搜索網絡中轉發(fā)節(jié)點

　(3)　　s_temp←${o_i}$, g*←0, $\left\{ {{s^*}} \right\} \leftarrow \varPhi $; //初始化搜索節(jié)點、鏈路約束和復用節(jié)點集合

　(4)　　find $(s|\alpha \left( {a,s} \right) = 1)$, update g^*; //從${o_i}$出發(fā)搜索含有${A_s}$的交換節(jié)點，并更新鏈路約束

　(5)　　　if $\varepsilon ({f_{i',j'}},s)$==1 && ${f_{i',j'}}{\rm{ = }}{f_{i,j}}$ && ${H_{s,a}} > {c_{n,s}}\left( {{f_{i,j}}} \right)$ && $\left| {{o_i},s} \right| \le j \cdot \theta {g^*}$ then //s 滿足復用約束

　(6)　　　　{s^*}←s ; //將復用s 加入{$s^*$}

　(7)　if $\left\{ {{s^*}} \right\} = \varPhi $ then

　(8)　　reset空白交換節(jié)點$\left(s|\sum {\gamma (f,s) = 0} \right)$ to {$s^*$}; //將搜索到的空白交換節(jié)點置入{$s^*$}

　(9)　deploy (${f_{i,j}} \in {F_{As}}$, {$s^*$}), addroute (${P_i}$, {$s^*$}); //構造初始路徑

　(10)　for v←1 to S^*+1 do // ${P_i}$分為S^*+1段子路徑，${L_i}$分為S^*+1個子集

　(11)　　if $\left\{ {{f_{i,j}} \in {F_{An}}|{f_{i,j}} \in {l_{i,v}}} \right\} \ne \varPhi $ then

　(12)　　　deploy (${f_{i,j}} \in {F_{An}}$, ($n|\alpha \left( {a,n} \right) = 1, \; \beta \left( {n,s} \right) = 1, \; s \in {p_{i,v}}$)); //優(yōu)先部署各子集中加速功能到加速卡

　(13)　　deploy (${q_v}\left| {{\rm{ max}}\left( {{q_j}} \right), \; (n,s} \right|s \in {p_{i,v}},\beta \left( {n,s} \right) = 1 )$); //部署功能最多的方案到子路徑上

　(14)　　if $\left\{ {{f_{i,j}} \in {F_{As}}|{f_{i,j}} \in {l_{i,v}}} \right\} \ne \varPhi $ then //若各子集中仍有未部署的加速表網絡功能

　(15)　　　deploy (${f_{i,j}} \in {F_{As}}, \; (n|\alpha (a,n) = 1, \; \beta \left( {n,s} \right) = 1, \; s \in {p_{i,v}})$); //放寬加速功能映射位置

　(16)　　if $\left\{ {{f_{i,j}} \in {F_{As}} \cup {F_{An}}|{f_{i,j}} \in {l_{i,v}}} \right\} \ne \varPhi $ then //仍有未部署的加速網絡功能

　(17)　　　reject ${d_i}$; //拒絕該服務請求

　(18)　　　break;

　(19)　First-fit_deploy (${f_{i,j}} \in {L_i},n$); //利用First-Fit部署剩余網絡功能

下載: 導出CSV

表 2 VNF資源參數(shù)

VNF類型	1	2	3	4	5	6	7	8	9	10	11
計算存儲資源開銷c_n	3	4	6	5	5	2	7	6	4	7	5
加速表資源開銷${c_{s,a}}$	2	3	–	5	2	3	–	4	2	–	2
加速卡資源開銷${c_{n,a}}$	1	1	33	2	1	2	2	2	1	3	1

下載: 導出CSV

參考文獻(21)

HONG Chiyao, CAESAR M, and GODFREY P B. Finishing flows quickly with preemptive scheduling[C]. ACM SIGCOMM 2012 Conference on Applications, Technologies, Architectures, and Protocols for Computer Communication, Helsinki, Finland, 2012: 127–138.

BOURAS C, KOLLIA A, and PAPAZOIS A. SDN & NFV in 5G: advancements and challenges[C]. The 20th Conference on Innovations in Clouds, Internet and Networks, Paris, France, 2017: 107–111.

HAN Bo, GOPALAKRISHNAN V, JI Lusheng, et al. Network function virtualization: challenges and opportunities for innovations[J]. IEEE Communications Magazine, 2015, 53(2): 90–97. doi: 10.1109/MCOM.2015.7045396

MATIAS J, GARAY J, TOLEDO N, et al. Toward an SDN-enabled NFV architecture[J]. IEEE Communications Magazine, 2015, 53(4): 187–193. doi: 10.1109/MCOM.2015.7081093

BI Jun, ZHU Shuyong, SUN Chen, et al. Supporting virtualized network functions with stateful data plane abstraction[J]. IEEE Network, 2016, 30(3): 40–45. doi: 10.1109/MNET.2016.7474342

YI Xiaodong, DUAN Jingpu, and WU Chuan. GPUNFV: A GPU-accelerated NFV system[C]. The 1st Asia-Pacific Workshop on Networking, Hong Kong, China, 2017: 85–91.

SONG Haoyu. Protocol-oblivious forwarding: unleash the power of SDN through a future-proof forwarding plane[C]. The 2nd ACM SIGCOMM Workshop on Hot Topics in Software Defined Networking, Hong Kong, China, 2013: 127–132.

LI Bojie, TAN Kun, LUO Layong, et al. ClickNP: Highly flexible and high performance network processing with reconfigurable hardware[C]. 2016 ACM SIGCOMM Conference, Florianopolis, Brazil, 2016: 1–14.

LI Xiaoyao, WANG Xiuxiu, LIU Fangming, et al. DHL: Enabling flexible software network functions with FPGA acceleration[C]. The IEEE 38th International Conference on Distributed Computing Systems, Vienna, Austria, 2018: 1–11.

SUN Jian, ZHU Guangyang, SUN Gang, et al. A reliability-aware approach for resource efficient virtual network function deployment[J]. IEEE Access, 2018, 6: 18238–18250. doi: 10.1109/ACCESS.2018.2815614

KUO Tungwei, LIOU Bangheng, LIN K C J, et al. Deploying chains of virtual network functions: on the relation between link and server usage[C]. The 35th Annual IEEE International Conference on Computer Communications, San Francisco, USA, 2016: 1–9. doi: 10.1109/INFOCOM.2016.7524565.

陳卓, 馮鋼, 劉蓓, 等. 運營商網絡中面向資源碎片優(yōu)化的網絡服務鏈構建策略[J]. 電子與信息學報, 2018, 40(4): 763–769. doi: 10.11999/JEIT170641

CHEN Zhuo, FENG Gang, LIU Bei, et al. Construction policy of network service chain oriented to resource fragmentation optimization in operator network[J]. Journal of Electronics &Information Technology, 2018, 40(4): 763–769. doi: 10.11999/JEIT170641

ZENG Chaobing, LIU Fangming, CHEN Shutong, et al. Demystifying the performance interference of co-located virtual network functions[C]. 2018 IEEE Conference on Computer Communications, Honolulu, USA, 2018: 765–773.

湯紅波, 袁泉, 盧干強, 等. 一種支持節(jié)點分割的vEPC虛擬網絡功能部署模型[J]. 電子與信息學報, 2017, 39(3): 546–553. doi: 10.11999/JEIT160507

TANG Hongbo, YUAN Quan, LU Ganqiang, et al. A model for virtualized network function deployment based on node-splitting in vEPC[J]. Journal of Electronics &Information Technology, 2017, 39(3): 546–553. doi: 10.11999/JEIT160507

FENG Hao, LLORCA J, TULINO A M, et al. Approximation algorithms for the NFV service distribution problem[C]. 2017 IEEE Conference on Computer Communications, Atlanta, USA, 2017: 1–9.

ERAMO V, AMMAR M, and LAVACCA F G. Migration energy aware reconfigurations of virtual network function instances in NFV architectures[J]. IEEE Access, 2017, 5: 4927–4938. doi: 10.1109/ACCESS.2017.2685437

SUN Chen, BI Jun, ZHENG Zhilong, et al. NFP: Enabling network function parallelism in NFV[C]. Conference of the ACM Special Interest Group on Data Communication, Los Angeles, USA, 2017: 43–56.

MIJUMBI R, SERRAT J, GORRICHO J L, et al. Design and evaluation of algorithms for mapping and scheduling of virtual network functions[C]. The 2015 1st IEEE Conference on Network Softwarization, London, UK, 2015: 1–9.

ZILBERMAN N, AUDZEVICH Y, KALOGERIDOU G, et al. NetFPGA: rapid prototyping of networking devices in open source[C]. 2015 ACM Conference on Special Interest Group on Data Communication, London, UK, 2015: 363–364.

GE Xiongzi, LIU Yi, DU D H C, et al. OpenANFV: Accelerating network function virtualization with a consolidated framework in openstack[C]. 2014 ACM Conference on SIGCOMM, Chicago, USA, 2014: 353–354.

段通, 蘭巨龍, 胡宇翔, 等. 面向SDN/NFV架構的VNF硬件加速資源編排機制[J]. 通信學報, 2018, 39(6): 98–108. doi: 10.11959/j.issn.1000-436x.2018108

DUAN Tong, LAN Julong, HU Yuxiang, et al. Orchestration mechanism for VNF hardware acceleration resources in SDN/NFV architecture[J]. Journal on Communications, 2018, 39(6): 98–108. doi: 10.11959/j.issn.1000-436x.2018108

施引文獻

資源附件(0)

訪問統(tǒng)計