軟體定義網路（Software Defined Network, SDN）透過中央控制器來管理交換器，藉以提供靈活的網路資源分配。然而，控制器的能力受到繼農曆以及網路頻寬的限制。為了獲得更好的可擴展性，多控制器的架構在大型網路中顯得更重要。而控制器和網路交換器之間傳輸的延遲對網路性能的影響更是至關重要。

在這種狀況下，選擇控制器在網路拓墣中放置的節點以及如何分配各個控制器管理的交換器是很重要的問題。在過去關於控制器佈局的研究當中，大多數都只考慮到靜態的狀況，而忽略了流量是動態的事實。動態流量可能導致控制器過載，從而降低網路性能。

在本篇論文中，我們提出了一個兩階段的方法來處理控制器佈局的問題，其中包含了原形架構和演算法。在第一階段，我們先計算控制器的初始位置，並根據每個交換器可能對控制器產生負載的機率來分配各個控制器負責的交換器。接著在第二階段，交換器與控制器之間的配置會動態的變化，以避免控制器過載。

最後，我們通過使用一些真實的網路拓墣資料來實驗我們的方法。實驗結果顯示我們的方法可以提供更小的延遲，同時也保持控制器的負載低於一定的值。

Software-Defined Networking (SDN) provides flexible network resource allocations for traffic engineering by relying on a central controller to manage switches. However, the capacity of the central controller is limited by computational power and network bandwidth. For better scalability, multiple controllers are deployed in large networks. The latency between the controllers and switches is critical for network performance.

Under these circumstances, choosing the nodes on which to place the controllers and assign switches to the controllers are significant problems. Most of the previous research about controller placement only considers the static situation and ignore the fact that flows are dynamic. The dynamic flow may lead to controller overloading thus degrading network performance.

In this paper, we propose a two-stage method to deal with the controller placement problem, including an architecture prototype and algorithms. In the first stage, we compute an initial placement for the controllers and assign switches to them based on the probability that a demand will load on a given node to keep the average demand latency minimal. Then, in the second stage, switches are dynamically migrated to different controllers to avoid controller overloading.

Finally, we evaluate our method by using some real networks. The results show that our method can provide smaller latency while also keeping the controller loads below a given threshold.

[1] N. McKeown, T. Anderson, H. Balakrishnan, G. Parulkar, L. Peterson, J. Rexford, S. Shenker, and
J. Turner, “Openflow: Enabling innovation in campus networks,” SIGCOMM Comput. Commun. Rev.,
vol. 38, pp. 69–74, Mar. 2008.
[2] A. Dixit, F. Hao, S. Mukherjee, T. Lakshman, and R. Kompella, “Towards an elastic distributed sdn
controller,” in HotSDN 2013, (New York, NY, USA), pp. 7–12, ACM, 2013.
[3] “Open networking foundation.” https://www.opennetworking.org/.
[4] “Openstack, open source software for creating private and public clouds..” https://www.
openstack.org/.
[5] A. Tootoonchian, S. Gorbunov, Y. Ganjali, M. Casado, and R. Sherwood, “On controller performance
in software-defined networks,” in Hot-ICE 2012, (San Jose, CA), USENIX, 2012.
[6] D. Hock, S. Gebert, M. Hartmann, T. Zinner, and P. Tran-Gia, “Poco-framework for pareto-optimal
resilient controller placement in sdn-based core networks,” in NOMS 2014, pp. 1–2, 2014.
[7] A. Panda, W. Zheng, X. Hu, A. Krishnamurthy, and S. Shenker, “SCL: Simplifying distributed SDN
control planes,” in NSDI 2017, (Boston, MA), pp. 329–345, USENIX Association, 2017.
[8] D. M. F. Mattos, O. C. M. B. Duarte, and G. Pujolle, “A resilient distributed controller for software
defined networking,” in ICC 2016, pp. 1–6, 2016.
[9] L. Schiff, S. Schmid, and P. Kuznetsov, “In-band synchronization for distributed sdn control planes,”
SIGCOMM Comput. Commun. Rev., vol. 46, no. 1, pp. 37–43, 2016.
[10] M. Canini, P. Kuznetsov, D. Levin, and S. Schmid, “A distributed and robust SDN control plane for
transactional network updates,” in INFOCOM 2015, pp. 190–198, 2015.
[11] A. R. Curtis, J. C. Mogul, J. Tourrilhes, P. Yalagandula, P. Sharma, and S. Banerjee, “Devoflow: Scaling
flow management for high-performance networks,” SIGCOMM Comput. Commun. Rev., vol. 41,
pp. 254–265, Aug. 2011.
[12] T. Koponen, M. Casado, N. Gude, J. Stribling, L. Poutievski, M. Zhu, R. Ramanathan, Y. Iwata,
H. Inoue, T. Hama, and S. Shenker, “Onix: A distributed control platform for large-scale production
networks,” in OSDI 2010, (Berkeley, CA, USA), pp. 351–364, USENIX Association, 2010.
[13] “OpenDaylight Controller.” https://www.opendaylight.org/.
[14] A. Tootoonchian and Y. Ganjali, “Hyperflow: A distributed control plane for openflow,” in INM/WREN
2010, (Berkeley, CA, USA), pp. 3–3, USENIX Association, 2010.
[15] B. Heller, R. Sherwood, and N. McKeown, “The controller placement problem,” in HotSDN 2012,
(New York, NY, USA), pp. 7–12, ACM, 2012.
[16] G. Wang, Y. Zhao, J. Huang, Q. Duan, and J. Li, “A k-means-based network partition algorithm for
controller placement in software defined network,” in ICC 2016, pp. 1–6, 2016.
[17] L. Yao, P. Hong, W. Zhang, J. Li, and D. Ni, “Controller placement and flow based dynamic management
problem towards sdn,” in ICCW 2015, pp. 363–368, 2015.
[18] L. F. Mller, R. R. Oliveira, M. C. Luizelli, L. P. Gaspary, and M. P. Barcellos, “Survivor: An enhanced
controller placement strategy for improving sdn survivability,” in GLOBECOM 2014, pp. 1909–1915,
2014.
[19] Y. Hu, T. Luo, N. C. Beaulieu, and C. Deng, “The energy-aware controller placement problem in
software defined networks,” IEEE Communications Letters, vol. 21, pp. 741–744, April 2017.
[20] Y. nan HU, W. dong WANG, X. yang GONG, X. rong QUE, and S. duan CHENG, “On the placement
of controllers in software-defined networks,” The Journal of China Universities of Posts and
Telecommunications, vol. 19, no. Supplement 2, pp. 92 – 171, 2012.
[21] J. Xia, D. Pang, Z. Cai, M. Xu, and G. Hu, “Reasonably migrating virtual machine in nfv-featured
networks,” in CIT 2016, pp. 361–366, 2016.
[22] Y. Fu, J. Bi, K. Gao, Z. Chen, J. Wu, and B. Hao, “Orion: A hybrid hierarchical control plane of
software-defined networking for large-scale networks,” in ICNP 2014, pp. 569–576, 2014.
[23] H. Bo, W. Youke, W. Chuan’an, and W. Ying, “The controller placement problem for software-defined
networks,” in ICCC 2016, pp. 2435–2439, 2016.
[24] R. Cohen, L. Katzir, and D. Raz, “An efficient approximation for the generalized assignment problem,”
Information Processing Letters, vol. 100, no. 4, pp. 162 – 166, 2006.
[25] H. W. Kuhn, “The hungarian method for the assignment problem,” Naval Research Logistics Quarterly,
vol. 2, no. 1-2, pp. 83–97, 1955.
[26] “Rocketfuel:an isp topology mapping engine..” http://www.cs.washington.edu/
research/networking/rocketfuel/.
[27] I. Gurobi Optimization, “Gurobi optimizer reference manual,” 2016.