Measuring network throughput | EPFL Graph Search

Throughput of a network can be measured using various tools available on different platforms. This page explains the theory behind what these tools set out to measure and the issues regarding these measurements. Reasons for measuring throughput in networks. People are often concerned about measuring the maximum data throughput in bits per second of a communications link or network access. A typical method of performing a measurement is to transfer a 'large' file from one system to another system and measure the time required to complete the transfer or copy of the file. The throughput is then calculated by dividing the file size by the time to get the throughput in megabits, kilobits, or bits per second. Unfortunately, the results of such an exercise will often result in the goodput which is less than the maximum theoretical data throughput, leading to people believing that their communications link is not operating correctly. In fact, there are many overheads accounted for in throughput in addition to transmission overheads, including latency, TCP Receive Window size and system limitations, which means the calculated goodput does not reflect the maximum achievable throughput. The Maximum bandwidth can be calculated as follows: where RWIN is the TCP Receive Window and RTT is the round-trip time for the path. The Max TCP Window size in the absence of TCP window scale option is 65,535 bytes. Example: Max Bandwidth = 65,535 bytes / 0.220 s = 297886.36 B/s * 8 = 2.383 Mbit/s. Over a single TCP connection between those endpoints, the tested bandwidth will be restricted to 2.376 Mbit/s even if the contracted bandwidth is greater. Bandwidth test software is used to determine the maximum bandwidth of a network or internet connection. It is typically undertaken by attempting to download or upload the maximum amount of data in a certain period of time, or a certain amount of data in the minimum amount of time. For this reason, Bandwidth tests can delay internet transmissions through the internet connection as they are undertaken, and can cause inflated data charges.

Enhancing Quality of Service metrics for high fan-in Node.js applications by optimising the network stack: Leveraging IX: The Dataplane Operating System

Fredrik Peter Lilkaer

This thesis investigates the feasibility of porting Node.js, a JavaScript web application framework and server, to IX, a dataplane operating system specifically developed to meet the needs of high performance microsecond-computing type of applications in a datacentre setting. We show that porting requires extensions to the IX kernel to support UDS polling, which we implement. We develop a distributed load generator to benchmark the framework. The results show that running Node.js on IX improves throughput by up to 20.6%, latency by up to 5.23×, and tail latency by up to 5.68× compared to a Linux baseline. We show how server side request level reordering affect the latency distribution, predominantly in cases where the server is load saturated. Finally, due to various limitations of IX, we are unable at this time to recommend running Node.js on IX in a production environment, despite improved metrics in all test cases. However, the limitations are not fundamental, and could be resolved in future work.

2015

Protocol overheads in networks

Prasenjit Dey

To increase network capacity, it may not be sufficient to just increase the capacity of the networks. It may be necessary to reduce the network overhead bits, to leave more room for the data bits, to increase the network capacity. As network functions become more complicated, more elaborate protocols become necessary to manage the networks. Without a proper and fundamental understanding of the overheads resulting from protocol bits, we may have protocols with huge overheads, which diminish the network efficiency, resulting in high cost and low value networks. Most effort has been dedicated to finding higher capacity links to carry more data. Little, or no effort has been spent to define a framework and analyze the overheads in the networks with the aim of understanding their behavior and thus reducing them. In this thesis, we give some perspectives on quantifying network overheads and some policies to reduce them. In a fixed network example, we show that, one can provide a lower bound on the number of bits needed to encode each message, exchanged by a node with its neighbors, to know the connectivity information of the node to others, i.e. its network neighborhood. The lower bound also provides insights into the dynamics of protocol information and its connection to the link statistics. We then examine the overheads arising out of the network's effort to keep track of a user, equivalent to knowing the connectivity information in fixed networks, in the case of two-tier hierarchical networks in general, and cellular and peer-to-peer networks in particular. We find that there exists an optimal threshold-based announcing policy of the user, about its location, which minimizes the network overheads. Clustering of users is also found to be a good strategy, in two-tier hierarchical networks, to minimize user tracking overheads. We find the optimal clustering which minimizes the tracking overhead, and that the optimal clustering is crucial to finding the aforementioned optimal announcing policy. As a next strategy, we show that clustering of user location information can reduce the number of bits needed to represent the information. The location information updates can be sent periodically, and the necessary bits needed to un-cluster the information can be fetched on a requirement basis. Based on the location statistics of the user, one can find a clustering of the location information, such that, given the clustered information, the number of bits that need to be fetched is minimized. Bandwidth is a very important attribute of a network, which if not allocated optimally, can be an "overhead" in terms of reduced network efficiency and high service costs. We take the viewpoint that, the optimal allocation should be based on users' behavior, who are the final consumers of the bandwidth. We study the user's psychology for media and its connection to bandwidth, and show a way to optimize bandwidth allocation. The conclusions from the optimal allocation, lead us to an architecture of a low cost, high value, and fast Internet access solution.

EPFL2004

Protocol overheads in networks

Prasenjit Dey

EPFL2004