A switch-based approach to throughput collapse and starvation in data centers

Abstract

Abstract-Data center switches need to satisfy stringent lowdelay and high-capacity requirements. To do so, they rely on small switch buffers. However, in case of congestion, data center switches can incur throughput collapse for short TCP flows as well as temporary starvation for long TCP flows.In this paper, we introduce a lightweight hash-based algorithm called HCF (Hashed Credits Fair) to solve these problems at the switch level while being transparent to the end users. We show that it can be readily implemented in data center switches with O(1) complexity and negligible overhead. We illustrate using simulations how HCF mitigates the throughput collapse of short flows. We also show how HCF reduces unfairness and starvation for long-lived TCP flows as well as for short TCP flows, yet maximizes the utilization on the congested link. Last, even though HCF can store packets of a same flow in different queues, we also prove that it prevents packet reordering.

Simulation Study for the Performance Impacts of Video Multicast Caused by In-Cast in Cloud Computing Environment

“…And the performance may be incurred due to the congestion and retransmission mechanism provided by transmission control protocol (TCP). In order to overcome this problem, several schemes were proposed [4][5][6]. Among them, the HCF scheme proposed the credit based concept to balance the transmission rate of storage servers so that the congestion condition can be smoothed.…”

“…From the service point of view, the cloud services can be divided into three models: the software as a service (SaaS), the platform as a service (PaaS), and the infrastructure as a service (IaaS). Although the cloud computing provides flexible resource allocation, recently, it was found that, due to the characteristics of TCP protocol, the existence of in-cast issue in the storage area network greatly affects the processing performance of the cloud computing [1][2][3][4]. Thus, to adopt more servers for parallel downloading does not mean higher received throughput.…”