Limits...
Distributed controller clustering in software defined networks

View Article: PubMed Central - PubMed

ABSTRACT

Software Defined Networking (SDN) is an emerging promising paradigm for network management because of its centralized network intelligence. However, the centralized control architecture of the software-defined networks (SDNs) brings novel challenges of reliability, scalability, fault tolerance and interoperability. In this paper, we proposed a novel clustered distributed controller architecture in the real setting of SDNs. The distributed cluster implementation comprises of multiple popular SDN controllers. The proposed mechanism is evaluated using a real world network topology running on top of an emulated SDN environment. The result shows that the proposed distributed controller clustering mechanism is able to significantly reduce the average latency from 8.1% to 1.6%, the packet loss from 5.22% to 4.15%, compared to distributed controller without clustering running on HP Virtual Application Network (VAN) SDN and Open Network Operating System (ONOS) controllers respectively. Moreover, proposed method also shows reasonable CPU utilization results. Furthermore, the proposed mechanism makes possible to handle unexpected load fluctuations while maintaining a continuous network operation, even when there is a controller failure. The paper is a potential contribution stepping towards addressing the issues of reliability, scalability, fault tolerance, and inter-operability.

No MeSH data available.


ONOS controller latency result.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC5383233&req=5

pone.0174715.g008: ONOS controller latency result.

Mentions: Fig 8 shows the performance chart for the ONOS controller latency test using distributed controller clustering and without clustering. When the number of switches is 25, the latency is 42,309 m for distributed controller clustering as compared to 42,706 for distributed controller without clustering. Similarly, when the number of switches is 150, the latency is 46,249 ms for distributed controller clustering as compared to 46,684 for distributed controller without clustering. This result shows that the distributed controller clustering is better than distributed controller without clustering. The distributed controller clustering reduces the latency by an average of 1.6%. This may because have clustered controllers operate in a coordinated way and each controller is aware of the network state which is shared across other clustered nodes. Besides, the clustered controllers enable a load balancing function to redistribute the controller load between different clustered nodes, therefore, offering better scalability and performance as compared to the distributed controller without clustering. Fig 9 shows the results of the packet loss test for the HP VAN SDN controllers using the distributed controller clustering and without clustering. The results show that the total number of packets loss increase when the number of packets sent increase. When 5000 UDP packets with a size of 320000 KB are sent between two end devices, the percentage of packet loss for distributed controller with clustering is 3.53% as compared to 3.99% for distributed controller without clustering. The distributed controller clustering is better than distributed controller without clustering because the proposed clustering method drops fewer packets as compared to the distributed controller without clustering. This may due to the distributed controllers having difficulty in handling a coordinated control when there was a controller failure. The clustered controllers automatically reassign controllers to the switches without interruptions when a controller fails. This enables SDN-based networks to operate reliably in the event that a controller fails and reduce the number of packets loss. Fig 10 shows the results of the packet loss test for the ONOS controllers using the distributed controller clustering and without clustering. The results show that the total number of packets loss increase when the number of packets sent increase.


Distributed controller clustering in software defined networks
ONOS controller latency result.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC5383233&req=5

pone.0174715.g008: ONOS controller latency result.
Mentions: Fig 8 shows the performance chart for the ONOS controller latency test using distributed controller clustering and without clustering. When the number of switches is 25, the latency is 42,309 m for distributed controller clustering as compared to 42,706 for distributed controller without clustering. Similarly, when the number of switches is 150, the latency is 46,249 ms for distributed controller clustering as compared to 46,684 for distributed controller without clustering. This result shows that the distributed controller clustering is better than distributed controller without clustering. The distributed controller clustering reduces the latency by an average of 1.6%. This may because have clustered controllers operate in a coordinated way and each controller is aware of the network state which is shared across other clustered nodes. Besides, the clustered controllers enable a load balancing function to redistribute the controller load between different clustered nodes, therefore, offering better scalability and performance as compared to the distributed controller without clustering. Fig 9 shows the results of the packet loss test for the HP VAN SDN controllers using the distributed controller clustering and without clustering. The results show that the total number of packets loss increase when the number of packets sent increase. When 5000 UDP packets with a size of 320000 KB are sent between two end devices, the percentage of packet loss for distributed controller with clustering is 3.53% as compared to 3.99% for distributed controller without clustering. The distributed controller clustering is better than distributed controller without clustering because the proposed clustering method drops fewer packets as compared to the distributed controller without clustering. This may due to the distributed controllers having difficulty in handling a coordinated control when there was a controller failure. The clustered controllers automatically reassign controllers to the switches without interruptions when a controller fails. This enables SDN-based networks to operate reliably in the event that a controller fails and reduce the number of packets loss. Fig 10 shows the results of the packet loss test for the ONOS controllers using the distributed controller clustering and without clustering. The results show that the total number of packets loss increase when the number of packets sent increase.

View Article: PubMed Central - PubMed

ABSTRACT

Software Defined Networking (SDN) is an emerging promising paradigm for network management because of its centralized network intelligence. However, the centralized control architecture of the software-defined networks (SDNs) brings novel challenges of reliability, scalability, fault tolerance and interoperability. In this paper, we proposed a novel clustered distributed controller architecture in the real setting of SDNs. The distributed cluster implementation comprises of multiple popular SDN controllers. The proposed mechanism is evaluated using a real world network topology running on top of an emulated SDN environment. The result shows that the proposed distributed controller clustering mechanism is able to significantly reduce the average latency from 8.1% to 1.6%, the packet loss from 5.22% to 4.15%, compared to distributed controller without clustering running on HP Virtual Application Network (VAN) SDN and Open Network Operating System (ONOS) controllers respectively. Moreover, proposed method also shows reasonable CPU utilization results. Furthermore, the proposed mechanism makes possible to handle unexpected load fluctuations while maintaining a continuous network operation, even when there is a controller failure. The paper is a potential contribution stepping towards addressing the issues of reliability, scalability, fault tolerance, and inter-operability.

No MeSH data available.