You are on page 1of 9

Hindawi

Scientific Programming
Volume 2017, Article ID 3579540, 8 pages
https://doi.org/10.1155/2017/3579540

Research Article
Software-Defined Congestion Control
Algorithm for IP Networks

Yao Hu, Ting Peng, and Lianming Zhang
College of Physics and Information Science, Key Laboratory of Internet of Things Technology and Application,
Hunan Normal University, Changsha 410081, China

Correspondence should be addressed to Lianming Zhang; zlm@hunnu.edu.cn

Received 7 October 2017; Revised 28 November 2017; Accepted 11 December 2017; Published 27 December 2017

Academic Editor: Basilio B. Fraguela

Copyright © 2017 Yao Hu et al. This is an open access article distributed under the Creative Commons Attribution License, which
permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

The rapid evolution of computer networks, increase in the number of Internet users, and popularity of multimedia applications have
exacerbated the congestion control problem. Congestion control is a key factor in ensuring network stability and robustness. When
the underlying network and flow information are unknown, the transmission control protocol (TCP) must increase or reduce the
size of the congestion window to adjust to the changes of traffic in the Internet Protocol (IP) network. However, it is possible that
a software-defined approach can relieve the network congestion problem more efficiently. This approach has the characteristic of
centralized control and can obtain a global topology for unified network management. In this paper, we propose a software-defined
congestion control (SDCC) algorithm for an IP network. We consider the difference between TCP and the user datagram protocol
(UDP) and propose a new method to judge node congestion. We initially apply the congestion control mechanism in the congested
nodes and then optimize the link utilization to control network congestion.

1. Introduction presently. Researchers in this area have conducted significant
exploration and investigation and have put forward many
In recent years, the development of science and technology improvements. For example, the latest versions of the four
has increased user demand for computer networks, which has kinds of TCP congestion control algorithms, namely, Veno,
accelerated the severity of network congestion problems. The Westwood, Hybla, and Cubic, have achieved optimization
increase in devices having applications that require file down- by packet loss judgment, bandwidth prediction, time delay
loading and sharing, Internet browsing, voice over Internet compensation, and homogeneous compensation mechanism,
phone, and multimedia has also contributed to the severity respectively [1–3]. Paper [4] has studied the micro-level
of network congestion problems. Thus, efficient congestion behaviour characteristics of TCP congestion control algo-
control is a key factor to ensure IP network stability and rithm in multihop wireless networks. Paper [5] has proposed
robustness. an adaptive cross-layer-based TCP congestion control for 4G
There are two approaches in solving the network con- wireless mobile cloud access. A deadline-aware congestion
gestion problem. The first is the end-to-end approach, where control mechanism, based on a parametrization of the tradi-
research is focused on the mechanism of terminal congestion tional TCP New Reno congestion control strategy, has been
control to relieve congestion. The second is the network side proposed in [6].
approach, where congestion control is achieved through data The appropriate limit of the biggest transfer rate of each
stream scheduling and line management on the forwarding data flow can effectively reduce network congestion and
nodes. The initial research focused on the end-to-end con- packet loss. However, previous studies reported that when
gestion control, such as the widespread use of transmission the underlying network and information flow are unknown,
control protocol (TCP). Because of its good adaptability TCP must increase or reduce the size of congestion window to
and extensible capability, TCP has received widespread adjust to the changes of traffic [7–9]. This finding highlighted
interest and became the main congestion control algorithm that the traditional IP network lacks the direct control

and provide a better platform for network congestion occurs. Section 3 presents an SDCC algorithm University’s Clean Slate research [19. OpenTCP is a new architecture that is widely interfaces. control mechanism is immediately started. The separation of the control ered as the first SDN standard. Related Work. which is consid- dynamic. Then. menting network virtualization. and the delay of the new reroute shortest path is large. forwarding transmission from the network to the application are needed. Simultaneously. the account. The difference between TCP and UDP is 𝑘 paths are sequentially traversed to determine which path pointed out. This the network congestion problem more efficiently. It has a long is an SDN application deployed on an SDN controller using the global network view of SDNs and is established based technology update cycle and is too dependent on network on the good strategy mechanism with the forwarding nodes equipment manufacturers. controllers. control mechanism based on OpenFlow. 20]. It must increase or reduce the size of detected by central monitoring of the port statistics of the the congestion window to adjust to the changes of OpenFlow enabled switches. Based on the different characteristics of is not congested. This (SDCC) algorithm for the IP network. Lu et al. software-defined approach is still in its early stage of devel- In the traditional IP network architecture. making the in judging the link congestion control threshold. SDN is a research on the congestion problem in the IP network based mature software-defined area that originated from Stanford on SDN in Section 2. [14] proposed a TCP dynamic and forwarding planes are highly integrated. In this method. Ghobadi et al. Meanwhile. Its core concept is to separate Finally.1. the increasingly and modifying the terminal protocol stack to continue the complex network environment makes it difficult to develop dynamic adjustment with the terminal congestion mecha- the network and make innovations related to the hardware of nism to achieve optimization. In the following sections. We consider the packet loss rate and opti. Long et al. This method realizes the load balance to TCP and UDP. [16] presented a Compared with the existing literature. We have considered this aspect in our research to find a solution for the network congestion 2. rerouted to achieve load balance.2. mances. and the net. port. while a software-defined approach can relieve congested link are rerouted by the OpenFlow controller. we used a software-defined approach to bandwidth data flow in the congested link is identified and solve the congestion control problem in the IP network.2 Scientific Programming of forwarding queue and cannot guarantee link utilization 2. Software-defined network. the control opment. This mechanism is based plane and data plane allows the core technology OpenFlow to on the OpenFlow network features that define a method realize the flexible control of the network traffic. The algorithm is dynamically adap- tive to interactive link information and can effectively use (i) A comprehensive survey of congestion control in the bandwidth and network congestion control. It fills the SDN application physical devices or software of related protocols and perfor. If the selected routing involves a UDP data (iii) The simulation of our algorithm is comprehensively flow with highly demanding timeline but allows packet loss. OpenTCP work is difficult to extend and customize. algorithm chooses the 𝑘 shortest path for each data flow (ii) The type of data flow in the IP network is taken into ahead of schedule. When network more intelligent. The simulation shows that this reroute strategy may not be an optimal choice. [15] proposed a kind of network congestion functions. When network congestion occurs. enabling the network to be directly programmable. So. This framework also illustrates that the TCP dynamic tions to infrastructure is one of the top 10 strategic technology management strategy is open to managers and does not trends identified by Gartner [10]. adjustment framework based on SDN-OpenTCP. gap in the field of congestion control and is very interesting. we propose a new method in judging some extent. we first discuss the existing 2. Akbari [17] proposed a method based on SDN and OpenFlow the underlying network and flow information are to solve this problem. nodes. these methods do not consider the type of data flow. the proposed algorithm achieves network congestion we considered the difference of the type of data flow in our control and optimizes the link utilization. described in Section 4. the control plane and the data plane of the network hardware . Therefore. link congestion is unknown. especially in the proprietary equipment and closed However. and other global network to provide sup- The technology of software-defined everything from applica. method highlighted the effectiveness of SDN in solving the congestion problem. overhead. and some of the flows in any traffic. and manageable. the corresponding congestion network application innovation [11–13]. [18] also proposed a we propose a software-defined congestion control new congestion optimization algorithm based on SDN. Background and quality of service. Network congestion control using a problem. explained. It is considered a for the IP network. The effectiveness of the algorithm is new kind of network architecture and a method of imple- verified by a simulation experiment. mization of link utilization. it also has a significant computation congested nodes. Programmable models and code that allow data deployed and needs to modify its source side. proposed algorithm. the paper concludes in Section 5. have a specific adjustment method and adjustment of the ing (SDN) separates the network control and forwarding target. Key Principles of a Software-Defined Approach. and the maximum In this study. this paper has the congestion control algorithm that supports multiple active following major contributions: queue management. however. In the traditional networks. Hershberger et al. Gholami and IP network is presented.

there are 𝑖 UDP data flows and network congestion exists. . Then. links is as in Algorithm 1. then one or more appropriate data flows are Assume that the total bandwidth for the link R1 to selected. and where 𝑖 + 𝑗 = 𝑛. V2 . V𝑛 ). Calculation of Rerouting Links. If (V1 .1. Software-Defined Congestion (5) IF 𝐵th -𝐵udp < 𝐵tcp < B Control Algorithm (6) link → congested (7) ELSE continue monitoring This paper investigates the use of software-defined approach in solving the congestion control problem in the IP network. reporting interface status. V3 . then controlling the switches. 𝑓3 . Now. Then. which is the link surplus maximum We considered the data flow condition in the real network bandwidth for the TCP data flows after meeting the UDP data and mainly focused on the heterogeneity between the TCP flow transmission demand as follows: data flow and the UDP data flow in proposing this congestion control mechanism. Estimating Congested Links. such as data flow header and transmission rate. . Algorithm 1: Estimating. is the congested link. . including (1) For 𝑓𝑘 in link data flow header and transmission rate. (3) 𝐵tcp = 𝐵tcp + flow (4) ELSE 𝐵udp = 𝐵udp + flow 3. the controller needs message to each of the OpenFlow switches regularly. TCP flows in this link. a new method in judging congested nodes. The algorithm for estimating congested the controller. and the link congestion threshold is 𝐵th (specific control the congestion. 𝑓2 . It has the following three basic characteristics: f1 (1) centralized control. In the OpenFlow network. (2) control layer. (2) estimate whether there are congested 𝑗 links based on formula (1). . and (3) f2 network virtualization. As shown in Figure 1. (3) if congested link is discovered. and other functions. 𝑓𝑛 ) the SDN architecture. we confirm that R2 is in a congestion state. . initial value of 𝐵tcp and 𝐵udp is 0. 𝑖 󸀠 udp The process is as follows: (1) the controller obtains the 𝐵th (𝑡) = 𝐵th − ∑ 𝑓𝑘 (𝑡) global information and regularly monitors the network con. it is removed from the topology image to generate a new 𝑘=1 topology image. 𝑘=1 (1) gestion condition. the switches Input: link bandwidth regularly send Port-Status messages to controllers to report Output: congested links port information of each switch. to choose a new path for rerouting part of the data flows Flow switches receive this message and feedback to the in the congested link. where the controller obtains port in R1 to R2 link. and the 󸀠 (5) select one or more appropriate data flows to reroute. if 𝐵th < 𝐵tcp < 𝐵. obtains a global network topology image through the LLDP and then sends port status information request 3.2. and the shortest rerouting path is calculated to R2 is 𝐵. and the rate of each data stream is information of each switch to monitor for congested links. in the The main concept of the SDCC algorithm is based on OpenFlow network. Switches and controllers establish the secure connection by transport layer security (TLS) or TCP prior to OpenFlow message interaction. issuing flow sheets. The controller removes all estimated .Scientific Programming 3 equipment.0 standard spec. . according to the OpenFlow v1. the controller detects that. 𝐵tcp is the congestion threshold of the utilization. the congestion control mechanism 𝑗 TCP data flows. is recorded in the data (2) IF flow = TCP information flow of each port in the switches. In OpenFlow network. is started. information query. (2) programmable software. Software programming is implemented in the control plane. This process allows the efficient value of 𝐵th is according to the physical properties of the use of network bandwidth resources and increases the link network environment). Open. . there are 𝑛 data flows (𝑓1 . ification. and (3) equip- ment layer. This information. f3 OpenFlow network is comprised of the OpenFlow switch fn R1 R2 and OpenFlow controller. We proposed an SDCC algorithm that considered the dif- ference between TCP and UDP. network congestion to optimize the link utilization. controllers first send the link layer Figure 1: Bottom switch network. Moreover. and the link R1 to R2 3. tcp 𝐵tcp = ∑ 𝑓𝑘 (𝑡) . The SDN architecture can be divided into three levels: (1) application layer. and the start of the congestion control controller the data information flow of each port in the mechanism in the congested nodes. (4) calculate the shortest path between the last node of the congested node to the destination node. discovery protocol (LLDP) message (encapsulated in the Packet-Out message sent) to each of the OpenFlow switches to obtain the global network topology. . 𝐵tcp is the total rate of the TCP flows.

𝑓𝑗 ) on one congested link. Therefore. and then path distance as infinity to achieve the deletion. After many experiments. V2 . OFC is the OpenFlow con- (13) Newlink → Packet out //update flowtable troller with POX [21]. . and a new total rate of TCP flows 󸀠 󸀠 𝐵tcp closest to 𝐵th is obtained. and shortest path calculation Algorithm 2: Rerouting. V3 . quickly reach 2% magnitude. the congestion detection module is triggered to regularly Then. Assuming that there are 𝑗 TCP data flows 4. . Experiment Design. V𝑗 ). the rate of each data data flows with Iperf (as shown in Table 1). . and finally choosing information of each port and record the number and size the appropriate data flows transferred to this new path to of each flow through this port. In our test Input: link and node distance environment. As identified by the shortest path algorithm. 𝑓𝑏 . If (3) For 𝑖 do congested link is discovered. At last. it is removed from the topology (4) dist[𝑖] = V[0] → V[𝑖] image to generate a new topology image. congestion detection. . 4. . we set the port congestion threshold UDP data flows always emphasize short transfer time and at 950 MB/s in our test. OFS2. . we built four (𝑓1 . the algorithm of the OpenFlow add the rates of all data flows together. we selected the TCP data flows for the rerouting. Results and Analysis. We found that the network rerouting is presented as in Algorithm 2. . (8) If dist[𝑗] < mindist To verify the proposed congestion control mechanism (9) mindist = dist[𝑘] (10) For 𝑗 do //update the shortest path using the OpenFlow network.3. and sets OFS1- flows is (V1 . then we are assured that the data flows (𝑓𝑎 . . these four flows will be 󸀠 transmitted through Path 1. the LLDP protocol is shown in Figure 2. (2) will be increased over time. 𝑓3 . rerouted data flow selection. and it is out of the standard of common business demand. (14) OFPFC DELETE and we use the Iperf tool on user terminals to send the data (15) OFPFC ADD flows. high-efficiency requirements. and parameter 𝑗 represents 𝑗 when the port rate reaches 950 MB/s. In this experiment. Selection of Rerouted Data Flows. 𝑓2 . then obtains a new noncongested network topology image as After the POX controller starts. 𝑓𝑘 ) are optimal choice for rerouting the transmission. PC1–PC6 are common user terminals. when the total rate of the TCP flows on this link 𝐵tcp minus part of the rate (𝑓𝑎 . (16) Delete OutFlowtable We added the link monitoring. the rate of 𝑓3 𝐵tcp = 𝐵tcp − (V𝑎 + V𝑏 + ⋅ ⋅ ⋅ + V𝑘 ) . . 4. the packet loss rate will TCP data flows. . OFS1. Then. The selection algorithm of the rerouted R1 R2 data flows are as in Algorithm 3. To facilitate testing. . using the shortest path algorithm to calculate the shortest path with non-congestion from the last node of the (the experiment is set at 2 s) query each OFS for status congested node to the destination node. function modules in the POX controller and combined them with the existing modules of topology discovery and flow table update to achieve the proposed congestion control mechanism. and we obtain the following: OFS2 as Path 1 and OFS1-OFS3-OFS2 as Path 2.4 Scientific Programming To ensure high link bandwidth utilization. 𝑓𝑏 . Parameter NetFPGA’s port line speed is 1 Gbps. we built the test environment (11) If mindist + V[𝑘] → V[𝑗] < V[0] → V[𝑗] as shown in Figure 3. . 𝑓𝑘 ). and OFS3 are OpenFlow (12) dist[𝑗] = mindist + V[𝑘] → V[𝑗] switches by Mininet simulation.1. The controller (5) For 𝑖 do //find the nearest point calculates the shortest path between the last node of the (6) int mindist = INFINITY congested node to the destination node. the controller will complete the transmission. the controller (7) For 𝑗 do selects one or more appropriate data flows to reroute. the controller obtains the global information Output: new path and update the flow table (1) IF link[𝑎]→ congested and regularly monitors the network congestion condition. (2) dist[𝑎] = INFINITY The controller estimates whether there are congested links. 𝑖 represents 𝑖 UDP data flows. Parameter 𝑘 represents the 𝑘th data flow.2. . . . We considered the data flow condi- tion in the real network and mainly focused on the hetero- geneity between the TCP data flow and the UDP data flow in proposing this congestion control mechanism. In the concrete execution. to simulate the real 3. The running process of each module is shown congested links from the global network topology image and in Figure 4. we set the link first run to obtain the entire network topology. Considering that the network environment. . . Parameter 𝑤 represents the number of the most suitable data flows. Experiment and Result Figure 2: New bottom switch network.

First. and the most links maintain a stable trend without large fluctuation. (1) Link Bandwidth. Link monitoring Congestion detection Data flow selection Flow Type Rate 𝑓1 TCP (PC1–PC4) 150 MB/s Path calculation and flow table update Controller 𝑓2 UDP (PC2–PC4) 100 MB/s 𝑓3 TCP (PC3–PC6) Increasing Figure 4: Module architecture of the POX controller in SDCC 𝑓4 UDP (PC1–PC6) 350 MB/s algorithm. Figure 5 shows the complete process of turning point occurred as shown in Figure 6. Finally. OFC PC1 PC4 OFS1 OFS2 PC2 PC5 OFS3 PC3 PC6 Figure 3: OpenFlow network experiment platform. Table 1: Set of four data flows. At this time. the two the OFC detected OFS2 link is congested. When appropriate data flow is selected (𝑓1 in the experiment) from the 𝑓3 growth rate reaches around 350 MB/s. an obvious all data flows through this link. Path 1 and Path 2 loading with 𝑓3 data flow rate.Scientific Programming 5 Input: all data flow rates through the congested link Output: number of selected flow (1) For 𝑖 do (2) sum = sum + flow[𝑖] (3) For i do //Select the appropriate flow and tag it (4) If flow = TCP && congested && sum-flow[𝑤] > bestsum (5) bestsum = sum − flowsize[𝑤] (6) bestflownum = 𝑤 (7) Return bestflownum Algorithm 3: Selection. the flow is rerouted to .

Moreover. which indicates that it has Path 2 because we only select TCP data flows to reroute in our reached a state of congestion at this time.95 0. Then. However. . 0. and the LABERIO algorithm has the loading in Path 1 continues to rise steadily with the an average link utilization of about 74%.006 200 0. The packet loss rate is shown in Figure 7. After the congestion relief. the link utilization rate does not have much difference.002 0 0 0 50 100 150 200 250 300 350 400 0 50 100 150 200 250 300 350 400 Flow f3 rate (MB/s) Flow f3 rate (MB/s) Path1 Figure 7: Packet loss rate. Path 1 loading decreases by 150 MB/s from 950 MB/s to 800 MB/s. and congestion control is performed. Path2 Figure 6: Link bandwidth in turning point.8 800 700 0. when congestion did not occur. Path2 Figure 5: Link bandwidth.014 Flow load (MB/s) 600 0. Figure 8 shows that increase of 𝑓3 data flow until the next congestion.018 800 0.65 400 0.02 900 0. When 𝑓3 reached 350 MB/s. At the full stage of flow 𝑓3 rate from increased in loading size at about 150 MB/s. which shows the congestion relief and the Figure 6 provides details of the turning point of the packet loss rate return to a reasonable range.01 400 0.75 Flow load (MB/s) 600 0.55 0 50 100 150 200 250 300 350 400 100 Flow f3 rate (MB/s) 0 300 310 320 330 340 350 360 370 SDCC algorithm LABERIO algorithm Flow f3 rate (MB/s) Path1 Figure 8: Link utilization. the SDCC algorithm where the increase of 𝑓3 increases the packet loss rate of proposed in this paper does not fluctuate significantly.5%. there is an obvious with the aim of the algorithm. Path 2 significantly (3) Link Utilization. the average link utilization of the proposed the transmission of 𝑓1 to Path 2.7 500 0.6 Scientific Programming 1000 0. the algorithm still maintained high link utilization. The packet loss rate has increased dramatically to 2%. and the UDP data flows are not adjusted consistent 𝑓3 increases from 350 MB/s to 400 MB/s. While the rate of algorithm. but the overall trend and value are small and within over. when congestion occurs (2) Packet Loss Rate. decline process.85 Path 1 link usage (100%) 900 0.008 300 0. algorithm is about 78. a reasonable range. More- Path 1.9 1000 0. link bandwidth.016 Packet loss rate (100%) 700 0.004 100 0. which explains 0 to 400 MB/s. an obvious increase in packet loss rate was observed.6 300 200 0.012 500 0. First.

no. C.com/newsroom/id/2867917. . Zhao. 5. Jin. the [4] D. De Vleeschauwer et al. Gu. 1701–1706. 16. “TCP Veno: TCP enhancement for trans- mission over wireless access networks. Conflicts of Interest pp. the results show the effectiveness of our con. [14] M. Jude.-G. pp. J. current level of development in the traditional network algo. and the follow-up showed almost no growth trend. In future work. the complexity analysis shows the International Conference on Multimedia and Expo. 290–297. 2015. namely. Moreover. no. SDCC algorithm is based on the SDN architecture. SDCC algorithm can relieve the network congestion problem “Effects of iodinated contrast agents on renal oxygenation level more efficiently. and there are 𝑛 data streams on the link. July 2010. and M. G. H. Walker. E. Claeys. 2017.” IEEE Journal on Selected rate of flow 𝑓3 reaches 350 MB/s. and Y.” IEEE Access. J. pp. of rerouting data flows. Ge. Hu. Ejaz. Guo. Conclusion and Future Work aware TCP congestion control for video streaming services. and V. access. we can still ment.” in Proceedings parameters to provide a fair and accurate data flow selection of the 2014 13th International Conference on Optical Communi- strategy. Acknowledgments HotNets 2012. “Deadline- 5. the time complexity [5] B. Moreover. 374–399. SreeArthi. a follow-up with the growth rate of flow 𝑓3 showed an upward trend. It effectively relieved congestion and on Consumer Electronics-Taiwan. The urable wireless networks. optimize our algorithm considering the different attributes [12] Z.” in Natural Science Foundation of China (61572191 and Proceedings of the IEEE International Conference on Advanced 61402170). pp. 1–5. article 140. the SDCC algorithm is superior to the LABERIO algorithm in link utilization. [6] M. http://www. and Y. M. p. November 2014.” Proceedings of the IEEE. H. Gruen.” in Pro- ceedings of the 11th ACM Workshop on Hot Topics in Networks. 547–566. based TCP congestion control for 4G wireless mobile cloud Therefore. 100–108. K.” in Proceedings of the 2017 International Con- new shortest path calculation of rerouting. vol. 22. [15] H. considerable properties of the SDCC algorithm that are mainly related to the calculation of three parts.Scientific Programming 7 the utilization on Path 1 link starts at around 60% and shows References a linear growth trend with the flow rate of flow 𝑓3. Y.-H. vol. Roccetti. Palazzi.-Y. and W. 4048–4056. Q.” in Proceedings of the 2010 IEEE (4) Complexity. ICOCN 2014. “Adaptive cross-layer- of these three parts is 𝑂(𝑛𝑘). Rexford.-J. Using centralized control 2013. Sun. Freedman. Shen. “An active queue management adaptation of UDP packets or other such factors and adding different framework for software defined optical network. Bouten. Fu and S. 15. 2003. [3] C. L. C. Long. Ganjali. apply a software-defined approach to the data center network “HONE: Joint Host-Network Traffic Management in Software- or other more complex and special network to solve the Defined Networks. our congestion mechanism. pp. pp. Therefore. Firrincieli. we proposed a new network [8] A. Su. M. Ji. vol. El-Mougy. p.” in Proceedings of the 12th International Conference on Network and Congestion is a common problem in the IP network. 2013. October 2012. the level analysis of TCP congestion control algorithm in multi-hop wireless networks. we will [11] P. Y. Yeganeh. no.. “Rethinking end-to- end congestion control in software-defined networks. and J. 2004. “Reconfig- congestion control algorithm called SDCC algorithm. the link utilization of the Areas in Communications. and Y. M.” in Proceedings of A software-defined approach is required to find a solution to the 2013 19th IEEE International Conference on Networks. M. Y. and F.-H. 2016. May guaranteed the packet loss rate and link utilization. 2015. “Micro real-time monitoring and judgment of the congested link. December 2013. Karl. guarantees the packet loss rate. However. vol. Brunati. Malini. Information Networking Applications. January 2017. and optimizes the link utilization. pp. Chang. and it can 7. obtain a global topology for unified network management. Deng. Zhang. [9] J. ICON the network congestion problem. [2] C. When the [1] C. “An OpenWRT solu- tion for future wireless homes. and D. vol. pp. “Network supported conges- rithm is insufficient to meet the growing network demand. N. but the Service Management. 21. Liang.-S. Ibnkahla. Yu. 1. no. R. Herfet. Wang. no. Ghobadi. tion avoidance in software-defined networks. W. Finally. P.” in Proceedings of the 3rd IEEE International Conference gestion mechanism. vol. M. 23. Tang. 1-2. 2015. 𝑂(𝑘2 ). M. 1125–1158. 5. D. 2.” Journal of Network and Systems Manage- congestion problem in depth. “LABERIO: dynamic This work is supported in part by grants from the National load-balanced routing in OpenFlow-enabled networks. the simulation shows the effectiveness of nephropathy. Furthermore. where the maximum link utilization can be up to 95%. and T. S. [13] L. Xu. Diniesh. 60%.” International Journal of Satellite Communications and Networking.” BMC Nephrology. SDCC algorithm is about 80%. ICME 2010. Liew. and the selection ference on Computer Communication and Informatics (ICCCI). M. [7] J. The proposed algorithm achieves [10] “Gartner Identifies the Top 10 Strategic Technology Trends for network congestion control. Ren. Hattab. Zhong. November 2016. 2014. S. “TCP Hybla: A TCP enhancement the LABERIO algorithm significantly declined at less than for heterogeneous networks. and K. 216–228. 61–66. cations and Networks. J. We considered the difference between TCP determined by blood oxygenation level dependent magnetic and UDP and proposed a new method in judging congested resonance imaging in rabbit models of type 1 and type 2 diabetic nodes. Sun. If 𝑘 nodes exist in a network topology pp. 103. Caini and R.” 2014. and 𝑂(𝑛). pp. features of the SDN network. “A Box-Covering-Based Routing Algorithm for Large-Scale SDNs. ICCE-TW 2016. The authors declare that they have no conflicts of interest. 14.gartner. 2. pp.

“OpenFlow control for cooperating AQM scheme. Nam. 4. .” in Pro- ceedings of the 23rd Iranian Conference on Electrical Engineering. no. “Finding the k shortest simple paths: A new algorithm and its implementation. 654–657. [17] M.com/noxrepo/. pp. 360-361. 3. [21] POX. and H. October 2010. H. [20] M. ICSP2010.-K. Du. 2. and S.” Computer Commu- nication Review. [19] N. pp. ICEE 2015. T.. Shin. Maxel. Y. May 2015.-J. no. October 2012. K. 2017. pp. vol. 69–74. 38. 2008. ICTC 2012.” ACM Transactions on Algorithms (TALG). https://github. Akbari.8 Scientific Programming [16] L. 2560– 2563. vol. and H. Xiao. M. Hershberger. Anderson.” in Proceedings of the 2010 IEEE 10th Interna- tional Conference on Signal Processing. “Congestion control in software defined data center networks through flow rerouting.-H. Article ID 1290682. 2007. Balakrishnan et al. [18] J. Lu. Kim. McKeown. “Software-defined networking (SDN): A reference architecture and open APIs. “OpenFlow: enabling innovation in campus networks. Gholami and B. pp. Suri.” in Proceedings of the 2012 International Conference on ICT Con- vergence: “Global Open Innovation Summit for Smart ICT Con- vergence”.

hindawi.hindawi.com Volume 2014 International Journal of Advances in Biomedical Imaging $UWLÀFLDO 1HXUDO6\VWHPV International Journal of Advances in Computer Games Advances in Computer Engineering Technology Software Engineering Hindawi Publishing Corporation Hindawi Publishing Corporation Hindawi Publishing Corporation Hindawi Publishing Corporation Hindawi Publishing Corporation http://www.com Volume 201 Advances in Fuzzy Systems Modelling & Simulation in Engineering Hindawi Publishing Corporation Hindawi Publishing Corporation Volume 2014 http://www.hindawi.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.com Volume 2014 http://www.hindawi.hindawi.hindawi.hindawi.com Volume 2014 International Journal of Reconfigurable Computing Advances in Computational Journal of Journal of Human-Computer Intelligence and Electrical and Computer Robotics Hindawi Publishing Corporation Interaction Hindawi Publishing Corporation Neuroscience Hindawi Publishing Corporation Hindawi Publishing Corporation Engineering Hindawi Publishing Corporation http://www.com Volume 2014 http://www.hindawi.hindawi. Advances in Journal of Industrial Engineering Multimedia Applied Computational Intelligence and Soft Computing The Scientific International Journal of Distributed Hindawi Publishing Corporation World Journal Hindawi Publishing Corporation Sensor Networks Hindawi Publishing Corporation Hindawi Publishing Corporation Hindawi Publishing Corporation http://www.com &RPSXWHU1HWZRUNV DQG&RPPXQLFDWLRQV  Advances in  Artificial Intelligence +LQGDZL3XEOLVKLQJ&RUSRUDWLRQ KWWSZZZKLQGDZLFRP 9ROXPH Hindawi Publishing Corporation http://www.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Submit your manuscripts at -RXUQDORI https://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.com Volume 2014 http://www.com Volume 2014 http://www.com Volume 201 http://www.hindawi.com Volume 2014 .hindawi.