Method for lowering buffered packet number in Multilink PPP waiting queue

Abstract

The invention discloses a method of reducing the number of cache slicing in a Multilink PPP waiting queue. The technical scheme of the invention is extended on the basis of a standard RFC receiving and transmitting mechanism. A slicing packaging timer is defined in a Multilink group, slicing of which the message packaging have not finished for a long time in a waiting queue can be cleaned up in time by the timer, therefore, caching time of the slicing is avoided from being too long. Furthermore, aiming at the out-of-sequence processing of message slicing, then invention can detect a message that can be packaged perfectly in time, extracts the message from the waiting queue and give the message to subsequent flows for processing, at the same time, the integrated message is replaced by a false message and the false message participates in the original out-of-sequence slicing processing flow.

Description

Technical field: [0001] The invention relates to Multilink PPP (Multilink Point-to-Point Protocol) in the field of data communication, and particularly relates to a method for reducing the number of buffered fragments in a Multilink PPP waiting queue. Background technique: [0002] Multilink PPP provides a method of decomposing, reorganizing and sorting datagrams on multiple logical data links. The purpose is to coordinate multiple independent links between two systems and provide higher bandwidth than any single link Virtual link. [0003] However, compared to single-link PPP (Point-to-Point Protocol), it is precisely because multiple single links are bundled together, Multilink PPP must pay for link management and message decomposition and reorganization. A considerable part of the additional cost is that in the process of receiving fragments and assembling fragments, the out-of-sequence processing of message fragments will inevitably cause a certain number of out-of-sequence fr...

AI Technical Summary

Problems solved by technology

There is a way to send an empty segment through an idle link to make the receiving end clear the unfinished assembly segment (see RFC1990 for details), but this method requires both ends to be able to send and process this empty segment, such a condition In practical application is not guaranteed

[0005] In addition, in the standard Multilink PPP receiving mechanism, due to the processing of out-of-order fragments (the sequence number carried by the fragment is not equal to the sequence number currently waiting and within the current receiving window), the method of buffering and waiting is adopted. Often because of waiting for an unreceived fragment, the fragments that have been received and can be assembled into a complete message are blocked in the Multilink waiting queue (a queue used to cache waiting for assembly or waiting for out-of-order processing fragments in the Multilink group ), cannot be handed over to the follow-up process in time

Images