CN-122027577-A - Method, equipment and medium for scheduling multilink traffic based on convergence and distribution

CN122027577ACN 122027577 ACN122027577 ACN 122027577ACN-122027577-A

Abstract

The invention discloses a method, equipment and medium for multi-link flow scheduling based on convergence and distribution, which relate to the technical field of flow scheduling and comprise the steps of collecting access messages, merging flow clusters, counting total traffic and message quantity of the flow clusters, generating a service observation row and average message length, executing link light detection and interface counting, generating a link observation row, generating a normalized traffic sequence and a service discrimination threshold value, acquiring a service phase row, generating a bearing state quantity and a link discrimination threshold value, acquiring a link phase row, generating a phase conflict value sequence and a time limit bearing quantity sequence, forming a candidate link set, solving the synchronization completion moment, generating link traffic, generating the number of final continuous message blocks according to the candidate link set and the link traffic, combining the average message length and the message quantity, executing distribution transmission, enabling link selection and service distribution to be more targeted, and improving the consistency, the executable and the stability of multi-link flow scheduling.

Inventors

DING QUANFU
YANG ZHENGXING
ZHANG MENG

Assignees

北京百卓网络技术有限公司

Dates

Publication Date: 20260512
Application Date: 20260415

Claims (10)

1. The method for scheduling the multi-link traffic based on the convergence and distribution is characterized by comprising the following steps of: Collecting access messages, merging flow clusters, counting total traffic and message quantity of the flow clusters, generating a traffic observation column and average message length, executing link light detection and interface counting, and generating a link observation column; Generating a normalized traffic sequence and a traffic discrimination threshold based on the traffic observation column, acquiring a traffic phase column, generating a bearing state quantity and a link discrimination threshold based on the link observation column, and acquiring a link phase column; Generating a phase conflict value sequence and a time limit receiving quantity sequence according to the total traffic, the allowed time delay, the link observation sequence, the service phase sequence and the link phase sequence to form a candidate link set, solving the synchronization completion time and generating the link traffic; And according to the candidate link set and the link traffic, combining the average message length and the message number, generating the final number of continuous message blocks, segmenting the continuous message blocks, and executing shunt transmission.
2. The method for aggregate-shunt based multilink traffic scheduling of claim 1, wherein the generating a link observation column comprises: establishing a unified time reference, configuring a clock, correcting and aligning, generating a scheduling sheet number, analyzing an access message, and acquiring a basic field; Merging is carried out on the basic field, and a stream cluster is formed according to the destination address, the transmission type and the allowed time delay; counting total traffic and message quantity of the stream clusters, writing the total traffic and message quantity into an arrival bit quantity formed by the stream clusters on the current dispatching sheet according to the sequence of the dispatching sheets, generating a traffic observation column, and calculating average message length according to the total traffic and message quantity; And executing link light detection and interface counting, generating unidirectional time delay, jitter, available rate and packet loss rate, writing a link state record into a link observation column, and generating the link observation column.
3. The method for aggregate-shunt based multilink traffic scheduling of claim 1, wherein the obtaining the traffic phase column comprises: Extracting a service observation column according to the stream cluster number, calculating normalized traffic, generating a normalized traffic sequence, and acquiring a service discrimination threshold; Based on the normalized traffic sequence, comparing the change value of the adjacent dispatching slices of the normalized traffic with the traffic discrimination threshold value bit by bit, marking the traffic change trend bit by bit, and generating a traffic phase sequence.
4. The method for aggregate-split-based multilink traffic scheduling of claim 3, wherein the acquiring the link phase column comprises: Reading a link observation column according to the link number, calculating the bearing state quantity of each link corresponding to the scheduling piece, and generating a link discrimination threshold according to the variation amplitude of the adjacent scheduling piece in the bearing state quantity; and comparing the change value of the adjacent scheduling slices in the bearing state quantity with a link discrimination threshold value bit by bit, marking the link state trend, and generating a link phase column.
5. The method for aggregate-shunt-based multilink traffic scheduling according to claim 1 or 4, wherein the generating a phase conflict value sequence and a time limit reception quantity sequence includes: According to the total traffic volume, the allowed time delay, the traffic phase column, the link phase column and the link observation column of the flow cluster, comparing the traffic phase column and the link phase column bit by bit, counting the number of conflict positions, and generating a phase conflict value sequence; and calculating the time limit bearing capacity according to the flow cluster and link combination corresponding to the phase conflict value sequence and combining the allowable time delay and the tail record of the link observation sequence, and generating a time limit bearing capacity sequence.
6. The convergence offload-based multilink traffic scheduling method of claim 1, wherein the generating link traffic comprises: the phase conflict value sequence and the time limit bearing quantity sequence are screened in a combined mode, double-layer sequencing is carried out, the time limit bearing quantity of the links is accumulated until the total traffic is covered, and a candidate link set is generated; When the accumulated value of the candidate link time limit bearing capacity reaches or exceeds the total traffic for the first time, performing binary search in a search interval formed by a search lower bound of the synchronization completion time and the allowable time delay, and obtaining the synchronization completion time to generate link traffic; and when the accumulated value of the time limit receiving quantity of the candidate link set is always smaller than the total traffic quantity, recording the allowable time delay as the synchronous completion time, and generating the link traffic quantity according to the time limit receiving quantity corresponding to each link in the candidate link set.
7. The method for scheduling multi-link traffic based on convergence and offloading as set forth in claim 6, wherein said generating the final number of consecutive message blocks comprises: Converting the number of the continuous message blocks according to the link traffic and the average message length to generate a number initial value of the continuous message blocks; Accumulating the initial value of the number of the continuous message blocks, comparing the number of the messages, executing final subtraction and first cyclic complement, eliminating discretization allowance, and generating the final number of the continuous message blocks.
8. The method for aggregate-split-based multilink traffic scheduling of claim 7, wherein the performing split transmission comprises: And sequentially segmenting continuous message blocks on an original message sequence in the stream cluster according to the number of the final continuous message blocks and the candidate link sets, and writing the continuous message blocks into corresponding transmission queues according to the fixed sequence of the candidate link sets to generate a transmission record.
9. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor implements the steps of the aggregate-split-based multilink traffic scheduling method of any one of claims 1 to 8 when the processor executes the computer program.
10. A computer readable storage medium having stored thereon a computer program, wherein the computer program when executed by a processor implements the steps of the aggregate-split based multilink traffic scheduling method of any one of claims 1 to 8.

Description

Method, equipment and medium for scheduling multilink traffic based on convergence and distribution Technical Field The invention relates to the technical field of flow scheduling, in particular to a method, equipment and medium for multi-link flow scheduling based on convergence and distribution. Background In the multilink transmission and flow scheduling technology, in order to improve the utilization rate of link resources and meet the transmission requirements of differentiated services, a method of flow classification, link state monitoring and shunt transmission synergy for access messages is often adopted, and the method generally counts parameters such as traffic, message scale, link bandwidth, time delay, jitter, packet loss rate and the like, so that link selection, service allocation and queue forwarding are performed, and load sharing, transmission guarantee and time delay control among a plurality of outgoing links are realized. However, under the scene of continuous fluctuation of the service arrival rate and the link bearing state, the conventional method is mostly dependent on single moment or static indexes for matching, and is difficult to reflect consistency between the service change trend and the link change trend, so that the scheduling judgment accuracy is affected. Disclosure of Invention The present invention has been made in view of the above-described problems occurring in the prior art. Therefore, the invention provides a multi-link flow scheduling method based on convergence and distribution, which solves the problems of insufficient cooperative matching of the business change trend and the link state change trend and insufficient fineness of continuous transmission organization after message distribution in the prior art. In order to solve the technical problems, the invention provides the following technical scheme: The invention provides a multi-link flow scheduling method based on convergence and distribution, which comprises the steps of collecting access messages, merging flow clusters, counting total traffic and message quantity of the flow clusters, generating a service observation column and average message length, executing link light detection and interface counting, generating a link observation column, generating a normalized traffic sequence and a service discrimination threshold value based on the service observation column, obtaining a service phase column, generating a bearing state quantity and a link discrimination threshold value based on the link observation column, obtaining the link phase column, generating a phase conflict value sequence and a time limit receiving quantity sequence according to the total traffic, the allowed time delay, the link observation column, the service phase column and the link phase column, forming a candidate link set, solving a synchronization completion time, generating link traffic according to the candidate link set and the link traffic, combining the average message length and the message quantity, generating a final continuous message block number, segmenting continuous message blocks, and executing distribution and transmission. The method comprises the steps of establishing a unified time reference, configuring a clock, correcting and aligning, generating a scheduling sheet number, analyzing an access message, obtaining a basic field, performing merging processing on the basic field, forming a flow cluster according to a destination address, a transmission type and an allowed time delay, counting total traffic and message quantity of the flow cluster, writing the total traffic and message quantity of the flow cluster into an arrival bit quantity formed by the flow cluster on a current scheduling sheet according to the scheduling sheet sequence, generating a traffic observation column, calculating an average message length according to the total traffic and message quantity, executing link light detection and interface counting, generating unidirectional time delay, jitter, available rate and packet loss rate, writing a link state record into the link observation column, and generating the link observation column. The method for dispatching the multi-link traffic based on the convergence and distribution comprises the steps of extracting a traffic observation column according to a stream cluster number, calculating normalized traffic, generating a normalized traffic sequence, obtaining a traffic judgment threshold value, comparing the change value of adjacent dispatching pieces of the normalized traffic with the traffic judgment threshold value bit by bit based on the normalized traffic sequence, marking the traffic change trend bit by bit, and generating the traffic phase column. The method for dispatching the multi-link traffic based on the convergence and diversion comprises the steps of reading a link observation column according to a link number, calculating a bearing state quantity of each link corresponding to a d