Detecting method for network link band width based on hop-by-hop time stamp label

Abstract

The method comprises: at first, using a test message and a load message to form a test message queue; using the detection source end to send the test message queue to network by using back to back way; each router in the transmission link records the current time stamp of each router in each test message; so the detection destination end gets a sample set comprising multi test data samples; finally, using bandwidth algorithm to make filter for the test data sample set to get physical bandwidth of the network link.

Description

technical field [0001] The invention relates to a bandwidth measurement method of a network link, in particular to a method for measuring the bandwidth of a network link based on a hop-by-hop time stamp label, and belongs to the technical field of delay measurement of communication networks. Background technique [0002] Since the birth of communication networks, the measurement and analysis of network performance has been an important research direction. It is of great significance for people to understand network behavior, improve network performance, and improve service quality. At the same time, for various low-level protocols and high-level applications, knowing the end-to-end performance parameters on the network path is helpful to achieve service quality assurance goals such as load balancing, congestion control, and QoS routing. However, with the increase of network complexity, accurate measurement of network performance parameters is often difficult to achieve. Net...

AI Technical Summary

Problems solved by technology

[0006] The disadvantage of this prior art is: the measurement method of obtaining BTC by simulating a TCP connection will inevitably consume a large amount of bandwidth, and if used improperly, it may also cause network overload, thus affecting the normal operation of user services; therefore, this method is called as "Bandwidth Killer"

However, many routers and hosts in the network may give these two types of packets a lower processing level, or even not process them, which will inevitably lead to inaccurate measurement results

[0011] 3. Store-and-forward nodes such as switches, bridges, repeaters, etc. in the network generally do not have IP addresses, so the tools for measuring them with decreasing IP TTL (Time To Live) values ​​are "hidden" nodes that are not visible, but actually take a certain amount of processing time as packets pass through them, which also skews the measurements

[0012] 4. This model uses two-way delay RTT instead of one-way delay OWD (One-Way Delay). Since two-way links are generally asymmetrical, in this model, two-way links are usually regarded as symmetrical links processing, which can also lead to deviations in the measurement results

[0013] 5. This method has a measurement upper limit. The MTU (Maximum Transport Unit) of the channel and the time granularity that the end system can provide determine the upper limit of the bandwidth that can be measured. Limited by the current hardware and software technology, tools based on this measurement principle Can not be used for high-speed network measurement

[0015] The packet pair method has the following main disadvantages: 1. Queuing failure: measurement packets may not be queued at the bottleneck link

3. Packet loss: Some packets may be lost directly, and other packets will be lost after the router overflows, resulting in wrong results

4. Congestion of the downstream link: If two packets are queued again on the link after the bottleneck link, the time interval between the two packets will change, thus making the result deviate from the true value

5. The measurement accuracy of multi-channel (Multi-channel). For example, a channel is composed of multiple channels. Since the packet pair in a measurement can only be transmitted through one channel, the measurement result is only the bottleneck bandwidth of a single channel. non-path bottleneck bandwidth

Images