Load-sensitive routing of long-lived IP flows

Abstract

Internet service providers face a daunting challenge in provisioning network resources, due to the rapid growth of the Internet and wide fluctuations in the underlying traffic patterns. The ability of dynamic routing to circumvent congested links and improve application performance makes it a valuable traffic engineering tool. However, deployment of load-sensitive routing is hampered by the overheads imposed by link-state update propagation, path selection, and signaling.Under reasonable protocol and computational overheads, traditional approaches to load-sensitive routing of IP traffic are ineffective, and can introduce significant route flapping, since paths are selected based on out-of-date link-state information.Although stability is improved by performing load-sensitive routing at the flow level, flapping still occurs, because most IP flows have a short duration relative to the desired frequency of link-state updates. To address the efficiency and stability challenges of load-sensitive routing, we introduce a new hybrid approach that performs dynamic routing of long-lived flows, while forwarding shortlived flows on static preprovisioned paths. By relating the detection of long-lived flows to the timescale of link-state update messages in the routing protocol, route stability is considerably improved. Through simulation experiments using a one-week ISP packet trace, we show that our hybrid approach significantly outperforms traditional static and dynamic routing schemes, by reacting to fluctuations in network load without introducing route flapping.

Workload Modeling for Computer Systems Performance Evaluation

“…• The distribution of flows in the Internet, both in terms of size and in terms of duration [609,215].…”

A New Algorithm for Long Flows Statistics—MGCBF

“…And sampling satisfies the needs of performance by losing the precision. A. Shaikh, J. Rexford and K. G. Shin [3] applied a method to realize the load balance by keeping the information of every flow, and judged the flow belonging to long or not by the packet number it arrived in a fixed time unit. This is a direct but not efficient way.…”

“…Considering the performance and costing, we set the MGCBF two stages (h=2), and the second stage CBF uses recurring minimum method to reduce the error probability. Because the flows whose length is smaller than 16 packets take above 90 percents of total flows [3]. If the first stage CBF used the counter whose maximum value is 16 (c 1 =16), the second stage CBF vector V 2 can be as small as 1/10 of the first stage CBF vector V 1 .…”

“…A flow is defined as a stream of packets subject to flow specification and timeout. The flow definition can be changed according to its usage, but recent studies show that a very small percentage of flows carry the majority of the packets and bytes [1][2] [3][8] regardless of their definition. And so it is very important to improve the network performance by finding out these heavy-hitter flows (called long flows).…”

On class-based isolation of UDP, short-lived and long-lived TCP flows

“…This percentage reflects the fact that only a small percentage of flows are long-lived, however they carry the majority of total traffic bytes in the Internet [17]. …”