Bufferbloat systematic analysis

Abstract

There is a recent interest in the bufferbloat phenomenon as a possible explanation for the increased network latency observed in the Internet. The bufferbloat phenomenon is related to the excessive packet queuing in over-sized buffers inside the network that may lead to network performance degradation. In this context, we observe a lack of experimental results considering the practical aspects of off-the-shelf network devices. Therefore, in this paper, we present a systematic analysis of the bufferbloat phenomenon considering the microscopic view of the insides in the buffer architecture of typical network devices. Our experimental results show that bufferbloat might not be a significant problem in practice. First, the phenomenon is only observed in specific cases. Second, changes made to the queues under the control of the operating system have negligible effects. Finally, recent versions of the Linux kernel avoid bufferbloat, even in the specific cases in which it was observed.Index Terms-Increased latency; network traffic; buffering 978-1-4799-3743-

Buffer management in wireless full-duplex systems

“…In a systemic evaluation of the bufferbloat effect, Cardozo et al [8] suggested that bufferbloat might not be a significant problem in practice. Considering the microscopic view of the buffer architecture of typical network devices, they drew attention to the impact of varying the buffer size of various buffers in the network transmit stack.…”

On lower estimating internet queuing delay

“…Based on their work, [30] tried to capture this delay type and measure its significance by passively monitoring BitTorrent traffic. [31] analyzed bufferbloat in a testbed with typical network device buffer architecture. Although research results typically suggest that the extent of bufferbloat is not so significant, the question is still open and it remains an actively researched area (e.g.…”

On the Interoperability of Bufferbloat Solutions

“…In our previous work [4], we have presented a systematic analysis of bufferbloat considering a microscopic view of the buffer architecture of typical network devices. We have varied the sizes of two main network queues on Unix-based systems and we have notice the bufferbloat phenomenon only in specific cases.…”

“…We are particularly interested in the system behavior while we vary qdisc and ring buffer queue sizes. 4 These queues store packets in the network and link layers, respectively. Further, we focus our analysis on the output from the ring buffer and qdisc queues since their input queues commonly present sufficient packet processing capacity.…”

“…Due to space constraints, we only show results while varying both ring buffer and qdisc sizes. The detailed impact of each one of these queues, with independent size variation, has been reported on our previous work [4].…”