在此論文中，我們探討在軟體定義網路 (SDNs)下網路流量工程的
優化。我們將原本網路虛擬化成眾多虛擬管道，每條管道都會由一個
標籤代表，而所有的路由都會由標籤進行優化。為了達到更適應性的
路由，我們將原先交換機中的路油表(Flow Table)區分為: 1) 儲存管道
路由資訊的管道表 2) 儲存流量與管道之間對應關係的流量表，而我
們提出四種演算法優化儲存的路由資訊與對應關係，並且施行流量
工程以優化系統。我們實作真實測試環境來驗證我們的方法，並且
除了實作我們提出的系統架構外，我們另外也實作分段路由(Segment
Routing, SR)，也利用Ryu提供的簡單路由(Ryu Simple Routing, RSR)。
實驗結果顯示出我們的系統在最高鏈接使用量和初始延遲的兩項指標
中都超越SR跟RSR: 1) 與SR相比，我們的系統有效降低20%到46%的最
高鏈接使用量；跟RSR相比，我們更進一步降低30%到50%的使用量。
2) 跟SR相比，與我們的系統降低了39%的初始延遲；以及與RSR相
比，93.2%的初始化延遲被降低。除此之外，我們的方法更為系統帶來
了負載平衡和快速錯誤回復的能力。

We study the problem of Traffic Engineering in multi-site Software Defined Network (SDN). In our system, we virtualize the physical links with virtual tunnels and each tunnel will be represented by a label. We propose to decouple the flow table into two tables: the more static tunnel table, and the more dynamic path table, for higher flexibility. The tunnel table contains a set of pre-built tunnels, while the path table matches each traffic flow to the labels. We formulate two optimization problems for tunnel construction and path assigner. Four algorithms are proposed to solve the Traffic Engineering properly, and a real testbed is implemented to prove the concept. Furthermore, we also compare the proposed solution with Segment Routing~(SR) and Ryu Simple Routing~(RSR), the experimental results show that our solution outperforms SR and RSR: 1) maximal link utilization reduction between 20\% and 46\% compared to SR, 30\% and 50\% compared to RSR. 2) Initial delay reduction by 39\% compared to SR, and 93.2\% compared to RSR. In addition, our proposed solution also brings load balancing and error resilience features to the network system.

Acknowledgments i
致 謝 ii
Abstract iii
中 文 摘 要 iv
1 Introduction 1
1.1 Contribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2 Thesis Organization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2 background 4
2.1 Software Defined Network . . . . . . . . . . . . . . . . . . . . . . . . . 4
2.2 Tunneling in the Internet . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.3 Segment Routing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.4 Multi-Protocol Label Switching . . . . . . . . . . . . . . . . . . . . . . 5
2.5 Traffic Engineering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
3 Usage Scenario 7
3.1 Network Topology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
3.2 Label Switching Operation . . . . . . . . . . . . . . . . . . . . . . . . . 7
4 Problem Statement 10
4.1 Initialization Delay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
4.2 Load Balancing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
4.3 Error Resilience . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
5 System Architecture 13
5.1 Controller . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
5.1.1 Component Diagrams . . . . . . . . . . . . . . . . . . . . . . . 13
5.1.2 Tunnel Constructor . . . . . . . . . . . . . . . . . . . . . . . . . 14
5.1.3 Admission Controller . . . . . . . . . . . . . . . . . . . . . . . . 14
5.2 Flow Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
5.2.1 Packet Forwarding . . . . . . . . . . . . . . . . . . . . . . . . . 15
5.2.2 Dynamic Adjustment . . . . . . . . . . . . . . . . . . . . . . . . 17
v
6 Tunnel Table Construction 18
6.1 Problem Formulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
6.2 Static Tunnel Finder algorithm . . . . . . . . . . . . . . . . . . . . . . . 19
6.3 Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
7 Path Table Construction 22
7.1 Problem Formulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
7.2 Static Path Assigner Algorithm . . . . . . . . . . . . . . . . . . . . . . . 24
7.3 Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
8 Error Resilience and Network Dynamics 27
8.1 Switch Dynamic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
8.2 Dynamic Tunnel Finder Algorithm . . . . . . . . . . . . . . . . . . . . . 28
8.3 Path Dynamic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
8.4 Dynamic Path Assigner algorithm . . . . . . . . . . . . . . . . . . . . . 29
9 Implementation 31
9.1 RYU Controller . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
9.2 Mininet Emulator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
9.3 Algorithms and Utilizations . . . . . . . . . . . . . . . . . . . . . . . . . 32
9.3.1 Traffic Engineering with Label switching . . . . . . . . . . . . . 32
9.3.2 Segment Routing Control Logic . . . . . . . . . . . . . . . . . . 33
9.3.3 Ryu Default Control Logic . . . . . . . . . . . . . . . . . . . . . 33
9.3.4 Network Monitor . . . . . . . . . . . . . . . . . . . . . . . . . . 33
10 Evaluations 34
10.1 Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
10.1.1 Topology Generator . . . . . . . . . . . . . . . . . . . . . . . . 34
10.1.2 Traffic Generator . . . . . . . . . . . . . . . . . . . . . . . . . . 36
10.2 Scenario and Metrics . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
10.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
10.3.1 Link Utilization . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
10.3.2 Delay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
10.3.3 Running Time and Flow Entries . . . . . . . . . . . . . . . . . . 43
10.3.4 Error Resilience . . . . . . . . . . . . . . . . . . . . . . . . . . 50
11 Related Work 53
11.1 Proactive Flow Set-up . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
11.2 Traffic Engineering using MPLS in legacy IP Networks . . . . . . . . . . 53
11.3 Traffic Engineering in SDNs . . . . . . . . . . . . . . . . . . . . . . . . 54
11.4 Label Switching . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
12 Conclusion 56
12.1 Migrating Proposed System on Segment Routing . . . . . . . . . . . . . 56
12.2 Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
12.3 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
Bibliography 58

[1] Segment Routing. http://www.segment-routing.net/.
[2] Transition scenarios for 3gpp networks. RFC 3574, 2003. https://tools.
ietf.org/html/rfc3574.
[3] D. Awduche. MPLS and traffic engineering in IP networks. IEEE Communications
Magazine, 37(12):42–47, 1999.
[4] J. BELLESSA. Implementing MPLS with label switching in software-defined net-
works. Master’s thesis, University of Illinois at Urbana-Champaign, 2015.
[5] T. Benson, A. Anand, A. Akella, and M. Zhang. MicroTE: fine grained traffic engi-
neering for data centers. In Proceedings of ACM Conference on emerging Network-
ing EXperiments and Technologies (CoNEXT’11). ACM, 2011.
[6] S. Bidkar, A. Gumaste, and A. Somani. A scalable framework for segment routing
in service provider networks: The omnipresent Ethernet approach. In In Proc. of
IEEE 15th International Conference on High Performance Switching and Routing
(HPSR), pages 76–83, 2014.
[7] M. Casado, T. Koponen, S. Shenker, and A. Tootoonchian. Fabric: a retrospective
on evolving SDN. In Proceedings of the first workshop on Hot topics in software
defined networks, pages 85–90. ACM, 2012.
[8] L. Cowen. Compact routing with minimum stretch. Journal of Algorithms,
38(1):170–183, 2001.
[9] IBM ILOG CPLEX optimizer. http://www-01.ibm.com/software/
integration/optimization/cplex-optimizer/.
[10] A. Curtis, J. Mogul, J. Tourrilhes, P. Yalagandula, P. Sharma, and S. Banerjee. De-
voFlow: scaling flow management for high-performance networks. In Proc. of ACM
Conference on Applications, technologies, architectures, and protocols for computer
communications (SIGCOMM’11), 2011.
58
[11] E.Danna, S.Mandal, andA.Singh. Apracticalalgorithmforbalancingthemax-min
fairness and throughput objectives in traffic engineering. In Proc of IEEE Interna-
tional Conference on Computer Communications (INFOCOM’12), 2012.
[12] L. Davoli, L. Veltri, P. Ventre, G. Siracusano, and S. Salsano. Traffic engineering
with segment routing: Sdn-based architectural design and open source implemen-
tation. In In Proc. of Fourth European Workshop on Software Defined Networks,
pages 111–112, 2015.
[13] S. Even, A. Itai, and A. Shamir. On the complexity of time table and multi-
commodity flow problems. In Foundations of Computer Science, 1975., 16th Annual
Symposium on, pages 184–193. IEEE, 1975.
[14] C. Filsfils, N. Nainar, C. Pignataro, J. Cardona, and P. Francois. The segment rout-
ing architecture. In in prec. of 2015 IEEE Global Communications Conference
(GLOBECOM), pages 1–6. IEEE, 2015.
[15] C. Guo, G. Lu, H. Wang, S. Yang, C. Kong, P. Sun, W. Wu, and Y. Zhang. Second-
net: a data center network virtualization architecture with bandwidth guarantees. In
Proceedings of the 6th International COnference, page 15. ACM, 2010.
[16] S. Hanks, T. Li, D. Farinacci, and P. Traina. Generic routing encapsulation (GRE).
RFC 1701, 1994.
[17] C. Hong, S. Kandula, R. Mahajan, M. Zhang, V. Gill, M. Nanduri, and R. Wat-
tenhofer. Achieving high utilization with software-driven WAN. In Proc. of ACM
Conference on Applications, technologies, architectures, and protocols for computer
communications (SIGCOMM’13). ACM, 2013.
[18] Y. Huang, M. Lee, T. Fan-Chiang, X. Huang, and C. Hsu. Minimizing flow initial-
ization latency in software defined networks. In in Proc. of Network Operations and
Management Symposium (APNOMS), 2015, pages 303–308. IEEE, 2015.
[19] Iperf home page. https://iperf.fr/.
[20] S. Jain, A. Kumar, S. Mandal, J. Ong, L. Poutievski, A. Singh, S. Venkata, J. Wan-
derer, J. Zhou, M. Zhu, et al. B4: Experience with a globally-deployed software de-
fined wan. ACM SIGCOMM Computer Communication Review, 43(4):3–14, 2013.
[21] S. Jain, A. Kumar, S. Mandal, J. Ong, L. Poutievski, A. Singh, S. Venkata, J. Wan-
derer, J. Zhou, M. Zhu, J. Zolla, U. Hölzle, S. Stuart, and A. Vahdat. B4: Experience
with a globally-deployed software defined WAN. In Proc. of ACM Conference on
59
Applications, technologies, architectures, and protocols for computer communica-
tions (SIGCOMM’13). ACM, 2013.
[22] S. Kandula, D. Katabi, B. Davie, and A. Charny. Walking the tightrope: re-
sponsive yet stable traffic engineering. In Proc. of ACM Conference on Appli-
cations, technologies, architectures, and protocols for computer communications
(SIGCOMM’05), 2005.
[23] S. Kandula, I. Menache, R. Schwartz, and S. Babbula. Calendaring for wide area
networks. In Proc. of ACM Conference on Applications, technologies, architectures,
and protocols for computer communications (SIGCOMM’14), 2014.
[24] C. Kozierok. The TCP/IP Guide: A Comprehensive, Illustrated Internet Protocols
Reference. No Starch Press, 1th edition, 2005.
[25] M. Lee and J. Sheu. An efficient routing algorithm based on segment routing in
software-defined networking. Computer Networks, 103:44–55, 2016.
[26] P. Lin, J. Hart, U. Krishnaswamy, T. Murakami, M. Kobayashi, A. Al-Shabibi,
K. Wang, and J. Bi. Seamless interworking of SDN and IP. In Proc. of ACM
SIGCOMM’13, 2013.
[27] P. Lin, J. Hart, U. Krishnaswamy, T. Murakami, M. Kobayashi, A. Al-Shabibi,
K. Wang, and J. Bi. Seamless interworking of sdn and ip. In ACM SIGCOMM
computer communication review, volume 43, pages 475–476. ACM, 2013.
[28] N. McKeown, T. Anderson, H. Balakrishnan, G. Parulkar, L. Peterson, J. Rexford,
S. Shenker, and J. Turner. OpenFlow: enabling innovation in campus networks. In
Proc. of ACM SIGCOMM’08, 2008.
[29] A.Medina, A.Lakhina, I.Matta, andJ.Byers. Brite: Anapproachtouniversaltopol-
ogy generation. In Modeling, Analysis and Simulation of Computer and Telecom-
munication Systems, 2001. Proceedings. Ninth International Symposium on, pages
346–353. IEEE, 2001.
[30] Mininet home page. http://mininet.org/.
[31] M. Moshref, A. Bhargava, A. Gupta, M. Yu, and R. Govindan. Flow-level state
transition as a new switch primitive for SDN. In Proc. of ACM Workshop on Hot
Topics in Software Defined Networking (HotSDN’14), 2014.
[32] RFC3031: Multiprotocol Label Switching Architecture. https://tools.
ietf.org/html/rfc3031#section-3.1.
60
[33] Open Networking Foundation. https://www.opennetworking.org/.
[34] OpenDaylight home page. http://http://www.opendaylight.org/.
[35] Open vSwitch home page. http://openvswitch.org/.
[36] OpenDaylight Pathman SR App from cisco. https://github.com/
CiscoDevNet/pathman-sr.
[37] C. Perkins. IP encapsulation within IP. RFC 2003, 1996.
[38] P. Pham and S. Perreau. Increasing the network performance using multi-path rout-
ing mechanism with load balance. Ad Hoc Networks, 2(4):433–459, 2004.
[39] R. Prabagaran and J. Evans. Experiences with class of service (CoS) translations
in IP/MPLS networks. In Proc. of IEEE Conference on Local Computer Networks
(LCN’01), 2001.
[40] V. Rawat, R. Tio, S. Nanji, and R. Verma. Layer two tunneling protocol (L2TP) over
frame relay. RFC 3070, 2001.
[41] Ryu SDN Framework Home Page. http://osrg.github.io/ryu/.
[42] C. Scoglio, T. Anjali, J. C. de Oliveira, I. Akyildiz, and G. Uhl. TEAM: A traffic
engineering automated manager for DiffServ-based MPLS networks. IEEE Com-
munications Magazine, 42(10):134–145, 2004.
[43] A. Sgambelluri, A. Giorgetti, F. Cugini, G. Bruno, F. Lazzeri, and P. Castoldi.
First demonstration of sdn-based segment routing in multi-layer networks. In Opti-
cal Fiber Communication Conference, pages Th1A–5. Optical Society of America,
2015.
[44] M. Soliman, B. Nandy, T. Lambadaris, and P. Ashwood-Smith. Source routed for-
warding with software defined control, considerations and implications. In in Pro-
ceedings of the 2012 ACM conference on CoNEXT student workshop, pages 43–44.
ACM, 2012.
[45] Source packet routing in networking (SPRING), 2016. https://tools.ietf.
org/html/rfc7855.
[46] G. Swallow. MPLS advantages for traffic engineering. IEEE Communications Mag-
azine, 37(12):54–57, 1999.
61
[47] G. Swallow. MPLS advantages for traffic engineering. IEEE communications mag-
azine, 37(12):54–57, 1999.
[48] R. R. T. Cormen, C. Leiserson and C. Stein. Introduction to Algorithms. The MIT
Press, 3th edition, 2009.
[49] RFC2702: Requirements for Traffic Engineering Over MPLS. https://tools.
ietf.org/html/rfc2702#section-2.0.
[50] D. Torrieri. Algorithms for finding an optimal set of short disjoint paths in a com-
munication network. IEEE Transactions on Communications, 40(11):1698–1702,
1992.
[51] H. Wang, H. Xe, L. Qiu, Y. Yang, Y. Zhang, and A. Greenberg. Cope: Traf-
fic engineering in dynamic networks. In Proc. of ACM Conference on Appli-
cations, technologies, architectures, and protocols for computer communications
(SIGCOMM’06), 2006.
[52] X. Xiao, A. Hannan, B. Bailey, and L. Ni. Traffic engineering with MPLS in the
internet. IEEE Network Magazine, 14(2):28–33, 2000.
[53] X. Xiao and L. Ni. Internet QoS: a big picture. IEEE Network Magazine, 13(2):8–
18, 1999.
[54] S. H. Yeganeh, A. Tootoonchian, and Y. Ganjali. On scalability of Software-Defined
Networking. IEEE Communications Magazine, 51(2):136–141, 2013.
[55] M. Zhu, J. Li, Y. Liu, D. Li, and J. Wu. TED: Inter-domain traffic engineering
via deflection. In Proc of IEEE International Symposium on Quality of Service
(IWQoS’14), 2014.