Publications

Adaptive Control Channel Traffic Shaping for Virtualized SDN in Clouds

2020 IEEE 13th International Conference on Cloud Computing (CLOUD), work-in-progress, Acceptance rate 26.3%, Virtual (originally in, Beijing, China), 2020, pp. 22-24, doi: 10.1109/CLOUD49709.2020.00013.

Yeonho Yoo, Gyeongsik Yang, Minkoo Kang, Chuck Yoo

Download PDF

Abstract

As the number of tenants in clouds grows, virtualized SDN faces a challenge of control channel fairness as the control channels interfere with each other. This paper proposes an adaptive traffic shaping scheme for control channels called “Sincon.” Through experiments, Sincon achieves traffic shaping for control channels and improves the variances of control channel throughputs and forwarding setup times up to 3.8 and 2.86 times, respectively.

Paper summary

Adapt the bandwidth of SDN control channels

A shared control path can be fairer when channel limits follow measured use.

Control channels from different tenants compete at the network hypervisor. Sincon measures their throughput and adapts per-channel limits to control interference.

Adjust limits and shape each channel

An adaptive limit calculator divides initial capacity among channels and adjusts limits according to observed usage and available capacity. A channel shaper enforces those limits on control traffic.

The adaptive limit calculator assigns channel budgets; the shaper enforces them between controllers and physical switches.View full-size figure
The adaptive limit calculator assigns channel budgets; the shaper enforces them between controllers and physical switches.

Measure fairness and forwarding setup time

Compared with OpenVirteX, the evaluation reports a reduction of up to 3.8× in the standard deviation of channel throughput. With four tenants, forwarding setup time improves by up to 2.9×. The experiments connect more balanced control-channel use to lower setup delay.

Reported improvements over OpenVirteX; lower variation indicates more balanced channel use.
MetricImprovement
Channel-throughput standard deviationUp to 3.8× lower
Forwarding setup time, four tenantsUp to 2.9× faster

Key takeaways

  1. Shared control channels need isolation that follows demand. Updating shaping limits from observed throughput addresses interference without treating every channel’s workload as constant.