Efficient elastic bulky traffic transfer with a new pricing scheme based on number of flows

Abstract

Emerging data center and cloud services carry large blocks of data, which leads to intensive bandwidth competition on today’s Internet, particularly during peak hours. Congestion pricing is considered to be an effective means to avoid peak-hour competition. Typical congestion pricing schemes proposed in the past require significant enhancements to existing network infrastructures. For instance, complex congestion information collection and price announcement mechanisms must be in place, hindering the deployments of many congestion pricing schemes. In this study, we present one simple yet effective congestion pricing scheme based on the number of flows in single-bottleneck networks carrying elastic traffic. During congested hours, data transmitters share bandwidth on the bottleneck and are also charged according to the number of flows they use. Under this pricing scheme, network capacity is used efficiently, and all users can achieve maximum utility with increasing and strictly concave utility functions. We investigated how a single congestion control parameter, the price index, affects the degree of network congestion. Through network simulation, we proved that under the proposed scheme, the big data transmitters sharply reduce their transfer during congested hours.