Showing papers by "Damon Wischik published in 2005"

Part I: buffer sizes for core routers

Abstract: In this article we describe recent work on buffer sizing for core Internet routers. This work suggests that the widely-used rule of thumb leads to buffers which are much larger than they need to be. For example, the buffer in a backbone router could be reduced from 1,000,000 packets to 10,000 without loss in performance. It could be reduced even further, perhaps to 10-20 packets, at the cost of a small amount of bandwidth utilization. This tradeoff is worth considering, for example for a possible future all-optical router.

Buffer sizes for large multiplexers: TCP queueing theory and instability analysis

Abstract: In large multiplexers with many TCP flows, the aggregate traffic flow behaves predictably; this is a basis for the fluid model of Misra, Gong and Towsley V. Misra et al., (2000) and for a growing literature on fluid models of congestion control. In this paper we argue that different fluid models arise from different buffer-sizing regimes. We consider the large buffer regime (buffer size is bandwidth-delay product), an intermediate regime (divide the large buffer size by the square root of the number of flows), and the small buffer regime (buffer size does not depend on number of flows). Our arguments use various techniques from queueing theory. We study the behaviour of these fluid models (on a single bottleneck Kink, for a collection of identical long-lived flows). For what parameter regimes is the fluid model stable, and when it is unstable what is the size of oscillations and the impact on goodput? Our analysis uses an extension of the Poincare-Linstedt method to delay-differential equations. We find that large buffers with drop-tail have much the same performance as intermediate buffers with either drop-tail or AQM; that large buffers with RED are better at least for window sizes less than 20 packets; and that small buffers with either drop-tail or AQM are best over a wide range of window sizes, though the buffer size must be chosen carefully. This suggests that buffer sizes should be much much smaller than is currently recommended.

Part II: control theory for buffer sizing

Abstract: This article describes how control theory has been used to address the question of how to size the buffers in core Internet routers. Control theory aims to predict whether the is stable, i.e. whether TCP flows are desynchronized. If flows are desynchronized then small buffers are sufficient [14 ]; the theory here shows that small buffers actually promote desynchronization--a virtuous circle.

SHRiNK: a method for enabling scaleable performance prediction and efficient network simulation

Abstract: As the Internet grows, it is becoming increasingly difficult to collect performance measurements, to monitor its state, and to perform simulations efficiently. This is because the size and the heterogeneity of the Internet makes it time-consuming and difficult to devise traffic models and analytic tools which would allow us to work with summary statistics. We explore a method to side step these problems by combining sampling, modeling, and simulation. Our hypothesis is this: if we take a sample of the input traffic and feed it into a suitably scaled version of the system, we can extrapolate from the performance of the scaled system to that of the original. Our main findings are as follows. When we scale an IP network which is shared by short- and long-lived TCP-like and UDP flows and which is controlled by a variety of active queue management schemes, then performance measures such as queueing delay and drop probability are left virtually unchanged. We show this in theory and in simulations. This makes it possible to capture the performance of large networks quite faithfully using smaller scale replicas.

Buffer requirements for high-speed routers

Abstract: A core Internet router can typically buffer 250 ms worth of data (1.25 GByte at 40 Gb/s). This would be challenging for an all-optical router. Happily, recent theory suggests that the optimal buffer size is around 30 kByte. (4 pages)

Making router buffers much smaller

Abstract: • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • The following three papers are invited editorials on the topic of Buffer Sizing. Currently, router manufacturers typically use a rule of thumb to choose buffer sizes: this rule says to provide at least one round trip time's worth of buffer-ing. These three papers reflect a feeling in some parts of the research community that this rule of thumb is out of date and needs to be revisited. Part I: Buffer Sizes for Core Routers (D. Wischik and N. McKeown) introduces the problem. It presents some simple models which suggest that, in routers with a high degree of statistical multiplexing, buffers can be much much smaller than the rule of thumb suggests. A key part of the analysis is synchronization. Broadly speaking, if TCP flows are synchronized then large buffers are needed; if TCP flows are unsynchronized then small buffers are sufficient. outlines a control-theoretic analysis of synchronization. This analysis indicates that small buffers actually induce desynchronization. It also suggests how TCP and/or active queue management schemes might be redesigned to be less prone to synchronization. why very small buffers are sufficient, so long as we are willing to sacrifice a small amount of throughput. The analysis shows that if the packets belonging to a TCP flow are sufficiently spaced out, then 10–20 packet buffers are enough, independent of the link speed. Packets are spaced naturally when flows enter through access links that are much slower than the backbone links. Alternatively, packets can be spaced out by making a small modification to TCP, described elsewhere as TCP Pacing. The models described in these three papers are supported by mathematical theory and by simulation, but so far there has not been any systematic experimental investigation. In parallel, we are starting to work with network operators to conduct experiments, and will report our results separately. And we hope that these papers will encourage more network operators to experiment with small buffers. We also hope to attract the attention of researchers who work on all-optical …