基于流量工程的軟件定義網(wǎng)絡(luò)控制資源優(yōu)化機(jī)制

胡宇翔; 李子勇; 胡宗魁; 胡濤

doi:10.11999/JEIT190276

基于流量工程的軟件定義網(wǎng)絡(luò)控制資源優(yōu)化機(jī)制

doi: 10.11999/JEIT190276

1.
國(guó)家數(shù)字交換系統(tǒng)工程技術(shù)研究中心 ??鄭州 ??450002
2.
中國(guó)人民解放軍第91445部隊(duì) 大連 116043

基金項(xiàng)目: 國(guó)家自然科學(xué)基金(61521003, 61872382)，國(guó)家重點(diǎn)研發(fā)計(jì)劃(2017YFB0803204)，廣東省重點(diǎn)領(lǐng)域研發(fā)計(jì)劃(2018B010113001)

詳細(xì)信息

作者簡(jiǎn)介:
胡宇翔：男，1982年生，副研究員，研究方向?yàn)閷拵畔⒕W(wǎng)、軟件定義網(wǎng)絡(luò)

李子勇：男，1995年生，碩士生，研究方向?yàn)檐浖x網(wǎng)絡(luò)、計(jì)算機(jī)網(wǎng)絡(luò)

胡宗魁：男，1986年生，助理研究員，研究方向?yàn)榭刂谱詣?dòng)化

胡濤：男，1993年生，博士生，研究方向?yàn)檐浖x網(wǎng)絡(luò)、網(wǎng)絡(luò)空間安全

通訊作者:
李子勇　17629352940@163.com

中圖分類號(hào): TN919.2; TP393.2
計(jì)量
- 文章訪問(wèn)數(shù): 3237
- HTML全文瀏覽量: 1297
- PDF下載量: 125
- 被引次數(shù): 0
出版歷程
- 收稿日期: 2019-04-22
- 修回日期: 2019-06-24
- 網(wǎng)絡(luò)出版日期: 2019-08-02
- 刊出日期: 2020-03-19

Control Resource Optimization Mechanism of SDN Based on Traffic Engineering

1.
National Digital Switching System Engineering & Technology Research Center, Zhengzhou 450002, China
2.
Unit 91445 of PLA, Dalian 116043, China

Funds: The National Natural Science Foundation of China (61521003, 61872382), The National Key Research and Development Program of China (2017YFB0803204), The Research and Development Program in Key Areas of Guangdong Province (2018B010113001)

摘要

摘要:
針對(duì)軟件定義網(wǎng)絡(luò)(SDN)分布式控制平面中由于網(wǎng)絡(luò)分域管理所引發(fā)的控制擴(kuò)張問(wèn)題，該文提出了一種基于流量工程的SDN控制資源優(yōu)化(TERO)機(jī)制。首先基于數(shù)據(jù)流的路徑特征對(duì)流請(qǐng)求的控制資源消耗進(jìn)行分析，指出通過(guò)調(diào)整控制器和交換機(jī)的關(guān)聯(lián)關(guān)系可以降低控制資源消耗。然后將控制器關(guān)聯(lián)過(guò)程分為兩個(gè)階段：先設(shè)計(jì)了最小集合覆蓋算法來(lái)快速求解大規(guī)模網(wǎng)絡(luò)中控制器關(guān)聯(lián)問(wèn)題；在此基礎(chǔ)上，引入聯(lián)合博弈策略來(lái)優(yōu)化控制器和交換機(jī)的關(guān)聯(lián)關(guān)系以減少控制資源消耗和控制流量開(kāi)銷。仿真結(jié)果表明，與現(xiàn)有的控制器和交換機(jī)就近關(guān)聯(lián)機(jī)制相比，該文機(jī)制能在保證較低控制流量開(kāi)銷的前提下，節(jié)省約28%的控制資源消耗。
- 軟件定義網(wǎng)絡(luò) /
- 資源優(yōu)化 /
- 控制器關(guān)聯(lián) /
- 分布式控制平面
Abstract:
In Software-Defined Networking (SDN) with distributed control plane, network expansion problems arise due to network domain management. To address this issue, a Traffic Engineering-based control Resource Optimization (TERO) mechanism of SDN is proposed. It analyzes the control resource consumption of flow requests processing with different path characteristics, and points out that the control resource consumption can be reduced by changing the association relationship between controllers and switches. The controller association mechanism is divided into two phases: firstly, a minimum set cover algorithm is designed to solve the controller association problem efficiently in large-scale network. Then, a coalitional game strategy is introduced to optimize the controller association relationship to reduce both control resource consumption and control traffic overhead. The simulation results demonstrate that while keeping control traffic overhead low, mechanism which in this paper can reduce control resource consumption by about 28% in comparison with the controller proximity mechanism.
- Software-Defined Networking (SDN) /
- Resource optimization /
- Controller association /
- Distributed control plane

HTML全文

圖 1 SDN分布式控制平面對(duì)流請(qǐng)求處理流程

下載: 全尺寸圖片幻燈片

圖 2 通過(guò)交換機(jī)遷移可以降低控制資源消耗

下載: 全尺寸圖片幻燈片

圖 3 控制器-交換機(jī)備選關(guān)系拓?fù)?/p>

下載: 全尺寸圖片幻燈片

圖 4 控制器-交換機(jī)關(guān)聯(lián)后拓?fù)?/p>

下載: 全尺寸圖片幻燈片

圖 5 階段1輸出的控制器-交換機(jī)關(guān)系

下載: 全尺寸圖片幻燈片

圖 6 控制資源消耗對(duì)比

下載: 全尺寸圖片幻燈片

圖 7 控制流量開(kāi)銷對(duì)比

下載: 全尺寸圖片幻燈片

圖 8 MSC算法和TERO控制資源消耗

下載: 全尺寸圖片幻燈片

圖 9 MSC算法和TERO控制流量開(kāi)銷

下載: 全尺寸圖片幻燈片

圖 10 不同網(wǎng)絡(luò)拓?fù)淇刂瀑Y源消耗

下載: 全尺寸圖片幻燈片

圖 11 不同網(wǎng)絡(luò)拓?fù)淇刂破髫?fù)載均衡率

下載: 全尺寸圖片幻燈片

表 1 最小集合覆蓋算法執(zhí)行過(guò)程

算法1 最小集合覆蓋算法(Minimum Set Coverage)
輸入： SDN網(wǎng)絡(luò)拓?fù)溧徑泳仃?{\text{G}} = [{a_{ij}}]$；網(wǎng)絡(luò)中流處理請(qǐng)求矩陣${\text{F}} = [{f_{ij}}]$；控制器所能關(guān)聯(lián)的備選集合：${C_i} = \{ {S_1},{S_2}, ·\!·\!· ,{S_i}\} $；控制器的容量及冗余因子：${\alpha _m}$, $\beta $
輸出：控制器-交換機(jī)之間的映射關(guān)系：${\text{X} } = [{x_{ij} }]$
(1) 初始化：控制器-交換機(jī)關(guān)聯(lián)關(guān)系SC={·}；已關(guān)聯(lián)的交換機(jī)set_switches={·};
(2) 統(tǒng)計(jì)網(wǎng)絡(luò)中端到端流量分布Flow_pair=Flow_sort(F)；
(3)　while I in Flow_pair：遍歷網(wǎng)絡(luò)中流量
(4)　　Path_switch= Dijkstra(G, i)；計(jì)算端到端流量的路徑
(5)　　while Path_switch: 循環(huán)4個(gè)完備策略
(6)　　　if Path_switch $ \subseteq $${C_i}$：若滿足完備策略1, SC[${C_i}$]={Path_switch }；流經(jīng)過(guò)的所有交換機(jī)關(guān)聯(lián)到${C_i}$
(7)　　　if ${S_i}$$ \in $Path_switch AND ${S_i} \in $${C_j}$滿足完備策略2, ${S_i} \to {C_j}$; ${S_i}$關(guān)聯(lián)到${C_j}$
(8)　　　if 存在${C_i} \subseteq {C_j}$：滿足完備策略3，則$ \cup {S_i} \to {C_j}$；交換機(jī)${S_i}$優(yōu)先關(guān)聯(lián)到${C_j}$
(9)　　　if Sn(${S_i}$)$ \subseteq $Sn(${S_j}$)：滿足完備策略4${S_i} \to \cup {\rm{ }}{C_j}$；交換機(jī)${S_i}$優(yōu)先處理
(10)　　　else 如果上述4個(gè)完備策略都不能滿足：實(shí)行貪婪算法switch = max(Path_switch & ${C_i}$)；尋找關(guān)聯(lián)交換機(jī)較多的控制器SC[${C_i}$]= {switch}；將相應(yīng)交換機(jī)關(guān)聯(lián)到控制器${C_i}$上
end if; end while;
(11) end while;
(12) SC={${C_j} = \{ {S_j},{S_{j + 1}}, ·\!·\!· {\rm{,}}{S_n}\} $；輸出控制器-交換機(jī)映射關(guān)系

下載: 導(dǎo)出CSV

表 2 聯(lián)合博弈策略執(zhí)行過(guò)程

算法2 聯(lián)合博弈策略Coalitional Game
輸入：算法1輸出的控制器-交換機(jī)之間的關(guān)聯(lián)關(guān)系${\text{X}} = [{x_{ij} }]$
輸出：控制器-交換機(jī)之間的關(guān)聯(lián)關(guān)系$\text{X}' = [{x'_{ij} }]$
(1) 初始化$\text{X} = [{x_{ij} }]$, ${\alpha _m}$, $\beta $
(2) repeat
(3)　　for each ${s_i}$ in $F$：尋找可能存在的交換機(jī)遷移
(4)　　　Initial migration pair ${s_i}:{\rm{ }}{c_m} \to {C_n}$；找到滿足兩個(gè)條件的交換機(jī)遷移對(duì)
end for
(5)　　for each $ {c_m}$：對(duì)于每一個(gè)控制器
(6)　　　${L_{ij}}$=${f_{ij}}{d_{kl}}{x_{ik}}{x_{jl}} + \delta {c_{ik}}$;計(jì)算每條數(shù)據(jù)流的資源消耗
(7)　　　if ${s_i}:{\rm{ }}{c_m} \to {C_n}$ and ${\theta _n} \le \alpha \times \beta $；保證控制器不過(guò)載，尋找可能的交換機(jī)遷移
(8)　　　　${L'_{ij}}$=${f_{ij}}{d_{kl}}{x'_{ik}}{x'_{jl}} + \delta {c'_{ik}}$；假設(shè)遷移，計(jì)算新的資源消耗
(9)　　　　if ${L'_{ij}} \le {L_{ij}}$: 若交換機(jī)遷移前后，資源消耗減少了，則接受遷移
(10)　　　　　${s_i} \to {c_j}$；實(shí)施交換機(jī)遷移${L_{ij}} = {L'_{ij}}$；更新的資源消耗
(11)　　　end if; end for；
(12) 直到系統(tǒng)沒(méi)有任何交換機(jī)要求遷移，則算法收斂

下載: 導(dǎo)出CSV

表 3 實(shí)驗(yàn)拓?fù)鋽?shù)據(jù)

網(wǎng)絡(luò)拓?fù)?/td>	節(jié)點(diǎn)數(shù)	鏈路數(shù)	控制器數(shù)	距離閾值
ARNES	34	47	4	3
ChinaNet	42	66	5	4
Interllifiber	73	93	6	5
Interoute	110	149	7	6

下載: 導(dǎo)出CSV

參考文獻(xiàn)(16)

ZHANG Yuan, CUI Lin, WANG Wei, et al. A survey on software defined networking with multiple controllers[J]. Journal of Network and Computer Applications, 2018, 103: 101–118. doi: 10.1016/j.jnca.2017.11.015

KARAKUS M and DURRESI A. A survey: Control plane scalability issues and approaches in Software-Defined Networking (SDN)[J]. Computer Networks, 2016, 112: 279–293. doi: 10.1016/j.comnet.2016.11.017

XU Yang, CELLO M, WANG I C, et al. Dynamic switch migration in distributed software-defined networks to achieve controller load balance[J]. IEEE Journal on Selected Areas in Communications, 2019, 37(3): 515–529. doi: 10.1109/JSAC.2019.2894237

MUQADDAS A S, GIACCONE P, BIANCO A, et al. Inter-controller traffic to support consistency in ONOS clusters[J]. IEEE Transactions on Network and Service Management, 2017, 14(4): 1018–1031. doi: 10.1109/TNSM.2017.2723477

BENSON T, AKELLA A, and MALTZ D A. Network traffic characteristics of data centers in the wild[C]. The 10th ACM SIGCOMM Conference on Internet Measurement, Melbourne, Australia, 2010: 267–280.

WANG Tao, LIU Fangming, GUO Jian, et al. Dynamic SDN controller assignment in data center networks: Stable matching with transfers[C]. The 35th Annual IEEE International Conference on Computer Communications, San Francisco, USA, 2016: 1–9.

胡濤, 張建輝, 鄔江, 等. SDN中基于分布式?jīng)Q策的控制器負(fù)載均衡機(jī)制[J]. 電子學(xué)報(bào), 2018, 46(10): 2316–2324. doi: 10.3969/j.issn.0372-2112.2018.10.002

CUI Jie, LU Qianzhe, ZHONG Hong, et al. A load-balancing mechnism for distributed SDN control using response time[J]. Thansactions on Network and Service Management, 2018, 15(4): 1197–1206. doi: 10.1109/TNSM.2018.2876369

伊鵬, 劉邦舟, 王文博, 等. 一種考慮軟件定義網(wǎng)絡(luò)控制節(jié)點(diǎn)故障的控制器部署和交換機(jī)遷移方法[J]. 電子與信息學(xué)報(bào), 2017, 39(8): 1972–1978. doi: 10.11999/JEIT161216

YI Peng, LIU Bangzhou, WANG Wenbo, et al. Controller placement and switch immigration strategy for SDN controller failure[J]. Journal of Electronics &Information Technology, 2017, 39(8): 1972–1978. doi: 10.11999/JEIT161216

ZHOU Yang, ZHENG Kangfeng, NI Wei, et al. Elastic switch migration for control plane load balancing in SDN[J]. IEEE Access, 2018(6): 3909–3919. doi: 10.1109/ACCESS.2018.2795576

張少軍, 蘭巨龍, 江逸茗, 等. 流特征感知的軟件定義網(wǎng)絡(luò)控制器動(dòng)態(tài)關(guān)聯(lián)機(jī)制[J]. 電子與信息學(xué)報(bào), 2018, 40(9): 2050–2056. doi: 10.11999/JEIT171149

ZHANG Shaojun, LAN Julong, JIANG Yiming, et al. Flow characteristics aware dynamic controller assignment in software-defined networking[J]. Journal of Electronics &Information Technology, 2018, 40(9): 2050–2056. doi: 10.11999/JEIT171149

SALMAN O, ELHAJJ I H, KAYSSI A, et al. SDN controllers: A comparative study[C]. The 18th Mediterranean Electrotechnical Conference, Lemesos, Cyprus, 2016: 1–6.

PAL C, VEENA S, RUSTAGI R P, et al. Implementation of simplified custom topology framework in Mininet[C]. 2014 Asia-Pacific Conference on Computer Aided System Engineering, South Kuta, Indonesia, 2014: 48–53. doi: 10.1109/APCASE.2014.6924470.

KNIGHT S, NGUYEN H X, FALKNER N, et al. The internet topology zoo[J]. IEEE Journal on Selected Areas in Communications, 2011, 29(9): 1765–1775. doi: 10.1109/jsac.2011.111002

LIAO Jianxin, SUN Haifeng, WANG Jingyu, et al. Density cluster based approach for controller placement problem in large-scale software defined networkings[J]. Computer Networks, 2017, 112: 24–35. doi: 10.1016/j.comnet.2016.10.014

YAO Guang, BI Jun, LI Yuliang, et al. On the capacitated controller placement problem in software defined networks[J]. IEEE Communications Letters, 2014, 18(8): 1339–1342. doi: 10.1109/LCOMM.2014.2332341

相關(guān)文章

施引文獻(xiàn)

資源附件(0)

訪問(wèn)統(tǒng)計(jì)