ProgLIMI: Programmable LInk Metric Identification in Software-Defined Networks

ProgLIMI: Programmable LInk Metric Identification in Software-Defined Networks
复制标题

ProgLIMI:软件定义网络中的可编程链接度量标识

DOI:
10.1109/tnet.2018.2865892
复制
发表时间:
2018-09
期刊:
IEEE/ACM Transactions on Networking
影响因子:
--
通讯作者:
Chen Nee Chuah
Chen Nee Chuah
中科院分区:
其他
文献类型:
--
作者:
Xiong Wang;Mehdi Malboubi;Zhihao Pan;Jing Ren;Sheng Wang;Shizhong Xu;Chen Nee Chuah

文献摘要

参考文献

被引文献

相似文献

在本文中,我们提出了用于软件定义网络(SDN)网络的可编程链路度量识别(ProgLIMI)基础结构。ProgLIMI通过利用SDN网络灵活的路由控制能力,从选定测量路径的累积端到端度量中识别往返链路度量(rtlm)。ProgLIMI主要解决三个子问题:1)监视器放置;2)线性独立的测量路径构建;3)流程规则设计。为了降低测量成本,ProgLIMI尽量减少所需监视器和流程规则的数量。在本文中,我们解决了全SDN和混合SDN网络的三个子问题。对于全SDN网络,ProgLIMI只需要在每个SDN交换机中使用一个监视器和两条流规则就可以实现全RTLM识别。相比之下,由于混合SDN网络的路由约束,混合SDN网络中的RTLM识别更为复杂。我们首先证明了在混合SDN网络中监视器的放置问题是np困难的。在此基础上,提出了混合SDN网络中监视器放置和测量路径选择问题,并提出了一种贪婪启发式算法来有效地解决该问题。我们在物理测试平台和仿真平台上的评估表明,ProgLIMI可以准确识别rtlm(延迟和损失率)。此外,ProgLIMI还具有资源效率,即在每个SDN交换机中只需要两条流规则和少量监视器,并且ProgLIMI带来的额外探测流量负载也很低。
In this paper, we propose the Programmable LInk Metric Identification (ProgLIMI) infrastructure for software-defined networking (SDN) networks. ProgLIMI identifies round-trip link metrics (RTLMs) from accumulated end-to-end metrics of selected measurement paths by leveraging the flexible routing control capability of SDN networks. ProgLIMI mainly solves three sub-problems: 1) monitor placement; 2) linearly independent measurement path construction; and 3) flow rule design. To reduce measurement cost, ProgLIMI tries to minimize the number of required monitors and flow rules. In this paper, we address the three sub-problems for both full and hybrid SDN networks. For full SDN networks, ProgLIMI can achieve full RTLM identification using only one monitor and two flow rules in each SDN switch. In contrast, the RTLM identification in hybrid SDN networks is more complicated due to the routing constraint of hybrid SDN networks. We first prove that the monitor placement problem in hybrid SDN networks is NP-hard. We then formulate the monitor placement and measurement path selection problem in hybrid SDN networks and propose a greedy heuristic algorithm to solve the problem efficiently. Our evaluations on both physical testbed and simulation platform reveal that ProgLIMI can accurately identify the RTLMs (delay and loss rate). Besides, ProgLIMI is also resource efficient, i.e., it only requires two flow rules in each SDN switch and a small number of monitors, and the extra probing traffic load incurred by ProgLIMI is also low.
DOI: 10.1145/2602204.2602216
发表时间: 2014-04-01
影响因子: 2.8
作者:
Vissicchio, Stefano;Vanbever, Laurent;Bonaventure, Olivier
通讯作者: Bonaventure, Olivier
DOI: 10.1145/1254882.1254891
发表时间: 2007-06
期刊: --
影响因子: --
作者:
Himabindu Pucha;Ying Zhang;Z. Morley Mao;Y. C. Hu
通讯作者: Himabindu Pucha;Ying Zhang;Z. Morley Mao;Y. C. Hu
DOI: 10.1145/1217722
发表时间: 2003-07
期刊: IEEE/ACM Transactions on Networking
影响因子: --
作者:
Yigal Bejerano;R. Rastogi
通讯作者: Yigal Bejerano;R. Rastogi
DOI: 10.1145/1111572.1111578
发表时间: 2005-12
期刊: SIGMETRICS Perform. Evaluation Rev.
影响因子: --
作者:
H. Burch;C. Chase
通讯作者: H. Burch;C. Chase
DOI: 10.1109/icassp.2001.940576
发表时间: 2001-05
期刊: 2001 IEEE International Conference on Acoustics, Speech, and Signal Processing. Proceedings (Cat. No.01CH37221)
影响因子: --
作者:
Meng-Fu Shih;A. Hero
通讯作者: Meng-Fu Shih;A. Hero