As a new network architecture, software defined networking (SDN) separates the control and data forwarding planes and implements programmable control. SDN provides a new way to improve the overall performance of the Internet. Although SDN has global view capacity, it also suffers from performance bottlenecks when handling massive Internet data, i.e., frequent interlayer communication between the control and data forwarding planes of SDN reduces the computational efficiency of the controller, and massive flow table entry data incur very high storage costs on switches. To further improve SDN performance and make it adaptable to the massive traffic processing of the Internet, this paper proposes a traffic multi-granularity processing (MGP) mechanism for Internet-oriented SDN. The proposed MGP applies the SDN architecture to traffic processing of the Internet backbone network and implements a multi-granularity traffic processing mechanism in consideration of routing and scheduling. Simulation results demonstrate that the proposed MGP can reduce the amount of interlayer communication and flow table entries, maintain network load balancing, improve the correctness and effectiveness of route selection, improve SDN performance, and improve the processing capacity of mass data on the Internet.
国家重点研发计划(2019YFB1802800)
国家自然科学基金(61872073,61572123)
[1] Kreutz D, Ramos F M V, Esteves Verissimo P. Software-Defined Networking: A Comprehensive Survey. Proc IEEE, 2015, 103: 14-76 CrossRef Google Scholar
[2] Yan C J, Wu D J, Xiong Y. SDN principle analysis: transfer control separation SDN architecture. Beijing: People's Posts and Telecommunications Press, 2016. Google Scholar
[3] Meyer D. The Software-Defined-Networking Research Group. IEEE Internet Comput, 2013, 17: 84-87 CrossRef Google Scholar
[4] Singh S, Jha R K. A Survey on Software Defined Networking: Architecture for Next Generation Network. J Netw Syst Manage, 2017, 25: 321-374 CrossRef Google Scholar
[5] Xu S, Wang X, Gao B, et al. Controller placement in software-defined satellite networks. In: Proceeedings of 2018 14th International Conference on Mobile Ad-Hoc and Sensor Networks (MSN), 2018. 146--151. Google Scholar
[6] Zhang B, Wang X, Huang M. Dynamic controller assignment problem in software-defined networks. Trans Emerging Tel Tech, 2018, 29: e3460 CrossRef Google Scholar
[7] Zhang B, Wang X, Huang M. Multi-objective optimization controller placement problem in internet-oriented software defined network. Comput Commun, 2018, 123: 24-35 CrossRef Google Scholar
[8] Guo Z, Xu Y, Liu R. Balancing flow table occupancy and link utilization in software-defined networks. Future Generation Comput Syst, 2018, 89: 213-223 CrossRef Google Scholar
[9] Katta N, Alipourfard O, Rexford J, et al. Infinite cacheflow in software-defined networks. In: Proceedings of the 3rd Workshop on Hot Topics in Software Defined Networking, Chicago, 2014. 175--180. Google Scholar
[10] Shirali-Shahreza S, Ganjali Y. ReWiFlow. SIGCOMM Comput Commun Rev, 2015, 45: 29-35 CrossRef Google Scholar
[11] Duliński Z, Rzym G, Cho$\nshortmid~$P. MPLS-based reduction of flow table entries in SDN switches supporting multipath transmission. arXiv preprint,. arXiv Google Scholar
[12] Jia X, Jiang Y, Guo Z, et al. Reducing and balancing flow table entries in software-defined networks. In: Proceedings of IEEE 41st Conference on Local Computer Networks (LCN), Dubai, 2016. 575--578. Google Scholar
[13] Zhang Q Y, Wang X W, Huang M. Software Defined Networking Meets Information Centric Networking: A Survey. IEEE Access, 2018, 6: 39547-39563 CrossRef Google Scholar
[14] Xu S, Wang X, Huang M. Modular and deep QoE/QoS mapping for multimedia services over satellite networks. Int J Commun Syst, 2018, 31: e3793 CrossRef Google Scholar
[15] Zhang B, Wang X, Huang M. Adaptive Consistency Strategy of Multiple Controllers in SDN. IEEE Access, 2018, 6: 78640-78649 CrossRef Google Scholar
[16] Dijkstra E W. A note on two problems in connexion with graphs. Numer Math, 1959, 1: 269-271 CrossRef Google Scholar
[17] Kodialam M, Lakshman T V. Minimum interference routing with applications to MPLS traffic engineering. In: Proceedings IEEE INFOCOM 2000, 2000. 2: 884--893. Google Scholar
[18] Riggio R, Rasheed T, Granelli F. EmPOWER: a testbed for network function virtualization research and experimentation. In: Proceedings of IEEE SDN for Future Networks and Services (SDN4FNS), Trento, 2013. 1--5. Google Scholar
[19] Dixit A, Prakash P, Hu Y C, et al. On the impact of packet spraying in data center networks. In: Proceedings of IEEE INFOCOM, Turin, 2013. 2130--2138. Google Scholar
[20] Al-Fares M, Radhakrishnan S, Raghavan B, et al. Hedera: dynamic flow scheduling for data center networks. In: Proceedings of Usenix Symposium on Networked Systems Design and Implementation, San Jose, 2010. 281--296. Google Scholar
[21] Long H, Shen Y, Guo M, et al. LABERIO: dynamic load-balanced routing in openflow-enabled networks. In: Proceedings of Advanced Information Networking and Applications, Barcelona, 2013. 290--297. Google Scholar
[22] Bannour F, Souihi S, Mellouk A. Distributed SDN Control: Survey, Taxonomy, and Challenges. IEEE Commun Surv Tutorials, 2018, 20: 333-354 CrossRef Google Scholar
[23] Hu T, Yi P, Guo Z. Dynamic slave controller assignment for enhancing control plane robustness in software-defined networks. Future Generation Comput Syst, 2019, 95: 681-693 CrossRef Google Scholar
Figure 1
The system framework
Figure 2
The network model
Figure 3
The flow chart of SDN traffic multi-granularity routing mechanism
| sip_seg | dip_seg | type |
| Source IP address segment | Destination IP address segment | Traffic type |
| |
|---|
| |
| Parameter | Parameter description |
| $A$ | Aggregations to be scheduled, a aggregation of a collection is represented by $a$. $A_l$ represents the aggregations of $A$ flowing through link $l$. |
| BR | The BR aggregations to be scheduled. |
| ${\rm~bw}_a$, ${\rm~delay}_a$, $P_a$ | The bandwidth requirement, the upper limit of the delay requirement, the available paths of the $a$-th aggregation to be scheduled. $P_{ai}$ represents the $i$-th alternative path available for aggregation $a$. |
| $L$ | Set of links included in the collection $P$, where a single link is represented by $l$, $l\in~L$. $l_{aj}$ denotes the $j$-th link in the link set of the $a$-th aggregation alternative path. |
| LC | Core links contained in collection $P$, where a single link is represented by $l_c$, $l_c\in$ LC. LC$_a$ represents the core link set of the $a$-th aggregation alternative path. |
| $n_{lc}$ | The number of links included in the set LC. |
| $C_{lp_{ai}}$ | Whether the link $P_{ai}$ contains link $l$. If it contains, the value is 1, otherwise the value is 0. |
| $T$ | The total amount of bandwidth of the link, and $T_l$ represents the total amount of bandwidth of link 1. |
| $U$ | The link has used bandwidth, and Ul indicates the used bandwidth of the link $l$. |
| $D$ | Link transmission delay, Dl is the link $l$ transmission delay. |
| $\sigma$ | The parameter that determines the congestion of the link, the upper limit of the required bandwidth of the link must not exceed the total bandwidth of the bandwidth. The range of value is (0, 1). |
| $X_{P_{ai}}$ | The result of the aggregation rerouting. If the aggregation $a$ is assigned to the $i$-th path in the alternative path set, the value is $l$, otherwise the value is 0. |
| $Ip_{ai}$ | Whether the $i$-th feasible path of the aggregation $a$ is its original path, and if so, the value is 1, otherwise the value is 0. |
| $N_l$, ${\rm~UR}_{l_c}$ | The bandwidth load of link $l$ and the bandwidth resource utilization of core link $l_c$ after aggregation scheduling. |
Download PDF
