Token-based Internet traffic control method

Abstract

This invention discloses a Token flow control method based on open Internet, involving packet switching network flow control technology. The methods include: user terminal equipment installed in accordance with network congestion data packets to be sent at the token; The end-user network access control into the network router the token rate; Intermediate routers in the network on the output interface controller Congestion has a physical, physical congestion controller bandwidth resources under the Congestion Adaptive extent to adjust physical congestion index, calculated in accordance with the fair probability of packet loss, random discarded token physical congestion level is less than the index data packets; Domain router interconnection between the output interface not only with physical congestion controller also has a virtual Congestion controller, controller under virtual Congestion token resources to adjust the degree virtual adaptive congestion index, lowering all output data packets token level.

Description

technical field [0001] The invention relates to the flow control technology of the packet switching network, in particular to a token-based Internet flow control method. Background technique [0002] At present, the flow control function of routers in the Internet is weak, and only provides network congestion information to user terminal equipment through packet loss signals. The traffic entering the network is mainly controlled by user terminal equipment through the transport layer or application layer, and the network lacks corresponding monitoring capabilities. . With the emergence of network technologies such as resumable uploads and multi-threaded downloads, the fairness and stability of the Internet have been challenged. Core-Stateless Fair Queue CSFQ (Core-Stateless Fair Queueing) is a method to enhance network flow control capabilities. It measures and marks the sending rate of session flow by the network access router. Usually, the data with the same source address...

AI Technical Summary

Problems solved by technology

This algorithm has three defects: it cannot deal with the fairness problem caused by Bit-Torrent download, it needs to re-mark the session rate when crossing different management domains, and the weight information cannot cross management domains

Due to the multi-destination address of BT download stream, the classification method of single-source and single-destination address of CSFQ session stream cannot control the aggregate transmission volume of BT business, so BT business will reduce the service quality of other traditional businesses and threaten the fairness and stability of the Internet

Since a complete trust relationship cannot be established between different management domains, its data flow needs to be remarked, which is the main performance bottleneck of the implementation of this method

Since the weight information cannot be reliably transmitted to other management domains, the weight information must be lost when the session rate is re-measured and marked between domains, so the weight function is only valid within one management domain, which limits the flexibility of the method

Images