WO2022160994A1

WO2022160994A1 - Load balancing method and apparatus, communication device, and storage medium

Info

Publication number: WO2022160994A1
Application number: PCT/CN2021/138452
Authority: WO
Inventors: 翟玉洲
Original assignee: 中兴通讯股份有限公司
Priority date: 2021-01-29
Filing date: 2021-12-15
Publication date: 2022-08-04
Also published as: CN114827039A

Abstract

A load balancing method and apparatus, a communication device, and a storage medium. The load balancing method comprises the following steps: receiving a new network session service (101); searching for a first non-overloaded link from a plurality of uplinks according to a set search order, and allocating the first non-overloaded link to the new network session service (102); sequentially searching for a first idle link from the plurality of uplinks according to the reverse order of the set search order, and migrating a network session on a previous non-idle link of the first idle link to the first idle link (103); and when there is no session service on the previous non-idle link of the first idle link found, disabling the previous non-idle link (104).

Description

A load balancing method, device, communication device and storage medium

CROSS-REFERENCE TO RELATED APPLICATIONS

This application is based on the Chinese patent application with the application number of 202110129844.6 and the filing date of January 29, 2021, and claims the priority of the Chinese patent application. The entire content of the Chinese patent application is incorporated herein by reference.

technical field

The present application relates to the field of communication technologies, and in particular, to communication device load judgment, session path selection, and load balancing strategies.

Background technique

CPE (Customer Premise Equipment) refers to the equipment directly connected to the operator's front-end network, including but not limited to telephones, wireless routers, firewalls, computers, optical modems, 4G-to-WiFi wireless routers, etc. , which is often deployed on the client side of FTTX home-customer services, and is used to provide comprehensive access to services such as wired broadband, IPTV, and VOIP for home customers. The CPE connects to the PON transmission network and connects to the specific service equipment to complete the connection between the access network and the user equipment. Existing CPEs generally have only one uplink, or multiple links but only one link is working at the same time. Therefore, the implementation strategy of load balancing is realized by restricting the traffic of some low-priority services. The normal operation of high-priority services will inevitably lead to performance degradation of this part of low-priority services, affecting user experience.

SUMMARY OF THE INVENTION

In view of this, an embodiment of the present application proposes a method for implementing load balancing of multiple uplink network links.

The embodiment of the present application first provides a load balancing method, which includes the following steps: receiving a new network session service; searching for a first non-overloaded link from a plurality of uplinks according to a set search sequence; The first non-overloaded link is allocated to the new network session service; the first idle link is found from the plurality of uplinks in turn according to the reverse order of the set search order, and the A network session on the last non-idle link of an idle link is migrated to the first idle link; when there is no session service on the last non-idle link of the first idle link found , close the last non-idle link.

The embodiment of the present application further provides a load balancing device, including: a network session service receiving module, configured to receive new network session services; a non-overloaded link search module, configured to search from multiple The first non-overloaded link is found in the uplink; the network session service allocation module is configured to allocate the first non-overloaded link to the new network session service; the idle link search module is set to is set to find the first idle link from the plurality of uplinks in turn according to the reverse order of the set search order; the network session migration module is set to search the uplink of the first idle link; A network session on a non-idle link is migrated to the first idle link; the link closing module is set to no session service on the non-idle link on the first idle link found. , close the last non-idle link.

An embodiment of the present application further provides a communication device, including a memory, a processor, and a program stored on the memory and executable on the processor, wherein the program is executed by the processor to achieve the desired Describe the steps of the load balancing method.

An embodiment of the present application further provides a storage medium, wherein the storage medium stores at least one program, and the at least one program can be executed by at least one processor to implement the steps of the load balancing method.

Description of drawings

FIG. 1 is a flowchart of a load balancing method according to Embodiment 1 of the present application.

FIG. 2 is a schematic diagram of a multi-link link according to Embodiment 2 of the present application.

FIG. 3 is a schematic diagram of a link load rate curve.

FIG. 4 is a structural diagram of a load balancing apparatus according to Embodiment 3 of the present application.

FIG. 5 is a structural diagram of a communication device according to Embodiment 4 of the present application.

Detailed ways

It should be understood that the specific embodiments described herein are only used to explain the present application, but not to limit the present application.

Example 1

This embodiment provides a load balancing method. The process is shown in FIG. 1 and includes the following steps:

101: Receive a new network session service;

102: Find out the first non-overloaded link from the multiple uplinks according to the set search sequence, and assign the first non-overloaded link to the new network session service;

103: Find out the first idle link from the multiple uplinks in turn according to the reverse order of the set search order, and select one of the non-idle links on the previous non-idle link of the first idle link. The network session is migrated to the first idle link;

104: When there is no session service on the last non-idle link of the found first idle link, the last non-idle link is closed.

The above load balancing method is applied to the communication with multiple uplinks. Through the reverse link search sequence, the first non-overloaded link is searched for the new session service, and the first idle link is searched for the old session service. Realize the dynamic selection of new session service links and the dynamic migration of old session service links, realize that multiple links work at the same time, and all session service flows work at full speed.

Wherein, the search sequence set in step 102 includes the number sequence of the plurality of uplinks. The number sequence of the links is used to search, and the search method is simple and easy to implement. The set search sequence may also be the sequence of randomly selecting links or other possible sequences, as long as it is ensured that the search sequence for the first idle link in step 103 is opposite to the search sequence set in step 102 . For example, assuming that the search sequence in step 102 is: 3 1 4 2 according to the link number, then the search sequence in step 103 is: 2 4 1 3 according to the link number. This operation is convenient for management and can quickly find new sessions for new sessions. At the same time, the migration of old sessions can be quickly realized, and the idle link can be quickly located and closed. In step 102, when a non-overloaded link cannot be found from the multiple uplinks according to the set search sequence, the last link in the multiple uplinks is allocated to the new network session. The session is attributed to the last link, so that it is convenient to search for idle links in reverse order to realize the migration of old session services.

Before the above-mentioned

steps

102 and 103, the method further includes the step of updating the load states of the multiple uplinks, and the step of updating the load states of the multiple uplinks includes:

updating the load status of each link except the last link in the plurality of uplinks;

When the link load rate is 0, update the load state of the link to a closed state;

When W1≥link load rate>0, update the load state of the link to an idle state;

When Wh≥link load rate>W1, update the load state of the link to a light load state;

When the link load rate>Wh, update the load state of the link to the overload state, until the link load rate drops to ≤W1 and continue for a period of time, update the load state of the link to the idle state ;

Wherein W1 represents the preset load idle reference value; Wh represents the preset load overload reference value.

Before step 102, the steps of updating the load states of the plurality of uplinks have no absolute sequence with step 101.

The above step 103 may not be based on the premise of adding a new session service, and a period may be set to periodically search for an idle link and process link migration of the old session service. The link migration of the old session service and the link selection of the new session service can be performed independently of each other (this is the meaning expressed by the dotted arrow in FIG. 1 ).

Embodiment 2

This embodiment provides a CPE device load balancing method. As shown in FIG. 2, the CPE device has n uplink ports, and has n uplink links L1, L2...Ln. The CPE internally allocates a corresponding net_device interface to each uplink interface. When there is data interaction, the strategy of this application is used to select the corresponding egress link through routing. The load balancing method specifically includes the following steps:

1. The CPE device receives a new network session service;

2. Update the load status of each link of L1, L2...Ln-1 in turn, whether it is overloaded, lightly loaded, idle or closed. Ln is the last link. When the previous link is overloaded, the rest of the data must be sent from this link, so there is no need to pay attention to the load of Ln. The load status is determined according to the load rate of the link. The load rate = the current bandwidth of the link/the maximum bandwidth of the link * 100%. The maximum bandwidth of the link is obtained by the CPE through negotiation with the network operator. The CPE calculates the load rate of each link every 1s. Current bandwidth. Figure 3 shows the load rate curve of a link:

1.1 The initial state link load rate is 0, and the updated link state is the closed state.

1.2 In the T1 time period, W1≥load rate>0, update the link state to idle state.

1.3 In the T2 time period, Wh≥load rate>Wl, the update link state is light load state.

1.4 During the T3+T4 time period, when the load rate > Wh, update the link state to the overload state, until the load rate drops to ≤ Wl, and continue for the preset stable time T4, update the link state to the idle state.

1.5 T5 time period, Wl ≥ load rate, update link state to idle state.

1.6 During T6 time period, Wh≥load rate>Wl, update link state to light load state.

3. Find the first non-overloaded link in the order of L1, L2...Ln-1, and assign the found first non-overloaded link to the new network session service. In this step, if no non-overloaded link is found, the last link Ln is allocated to a new network session.

4. Find the first idle link Lx in the order of Ln-1, Ln-2...L1, and migrate a network session on the last non-idle link Lx+1 of the first idle link Lx to link Lx. When there are multiple sessions on the link Lx+1, one is randomly selected for migration.

The CPE device creates a record for each network session service, including the selected link information. When there is data exchange in the session service, the CPE device searches for the corresponding link interface according to the record. If no session traffic record is found, the set default link is selected and the record is created. Therefore, when the session service needs to be migrated, the default link is set to Lx, and the record entry of the current session service is deleted. At this time, the CPE device treats the session service as a new session service, re-initiates it, and allocates it to the link Lx. Record the session. Repeat step 4 until there is no session service on the link Lx+1, and close the link Lx+1. The purpose of migrating old session services is to aggregate session services and then close empty links to save link resources.

In the above step 4, the premise of adding a new session service may not be required. The CPE device can set a period, regularly search for idle links, and process link migration of old session services. The link migration of the old session service and the link selection of the new session service can be performed independently of each other.

When the user's service traffic is heavy and the load capacity of one link cannot meet the demand, for example, the CPE has only one 10M link, when the user is watching a high-definition movie and temporarily uses 8M bandwidth, the user is keeping the first movie In the case of normal playback, a new video service with a bandwidth of 4M is opened. At this time, the required bandwidth is 12M, which is greater than the maximum value of the link of 10M, which will definitely cause the two services to slow down and reduce the user experience. Using the above load balancing implementation method, when the second 4M video service is started, the CPE will judge that the current first link is overloaded and start the second link, and the 4M video service will also use the second link. , which ensures that 8M movies and 4M videos work at full speed at the same time, and there will be no slowdown or freeze. When the 8M movie service is closed and the load drops, the first link becomes idle, so the 4M video service on the second link is switched back to the first link, and then the second link is closed. By analogy, it can be extended to multiple links and achieve load balancing of multiple links.

Embodiment 3

This embodiment provides a load balancing device, as shown in FIG. 4 , including:

The network session service receiving module 1 is set to receive a new network session service;

The non-overloaded link search module 2 is set to search out the first non-overloaded link from the multiple uplinks according to the set search sequence;

A network session service allocation module 3, configured to allocate the first non-overloaded link to the new network session service;

The idle link search module 4 is set to search out the first idle link from the multiple uplinks in turn according to the reverse order of the set search sequence;

The network session migration module 5 is configured to migrate a network session on the last non-idle link of the first idle link to the first idle link;

The link closing module 6 is configured to close the last non-idle link when there is no session service on the last non-idle link of the found first idle link.

In some embodiments, the load balancing apparatus further includes a link load status update module, which is configured to search out the first link from the plurality of uplinks in the non-overloaded link search module 2 according to a set search sequence. Before the step of finding a non-overloaded link and the step of finding the first idle link from the plurality of uplinks according to the reverse order of the set searching sequence, the idle link searching module 4 updates the Multiple uplink load status.

In some embodiments, when the non-overloaded link search module 2 cannot find a non-overloaded link from the plurality of uplinks according to the set search sequence, the last link is allocated to the new network session .

In some embodiments, the network session migration module 5 sets the found first idle link as the default link of the network session that needs to be migrated, deletes the record entry of the network session that needs to be migrated, and the network session needs to be migrated. The session is re-initialized as a new network session and assigned to the default link.

Through the cooperation of the above modules, the load balancing device in this embodiment realizes the dynamic selection of new session service links and the dynamic migration of old session service links, so that multiple links work simultaneously, and all session service flows work at full speed. .

Embodiment 4

This embodiment provides a communication device. As shown in FIG. 5 , the communication device includes a memory 7 and a processor 8 that are connected and communicated with each other. The memory 7 stores a program that can run on the processor 8, and the program is executed by the processor. The steps for implementing the load balancing method shown in Figure 1 include the following steps:

101: Receive a new network session service;

Through the implementation of the above method, the communication device of this embodiment realizes dynamic selection of new session service links and dynamic migration of old session service links, so that multiple links work simultaneously, and all session service flows work at full speed.

Embodiment 5

This embodiment provides a storage medium, where the storage medium stores at least one program, and the at least one program can be executed by at least one processor to implement the steps of the load balancing method shown in FIG. 1 , including the following steps:

101: Receive a new network session service;

Through the implementation of the above method, the dynamic selection of the new session service link and the dynamic migration of the old session service link are realized, multiple links work simultaneously, and all session service flows work at full speed.

In this embodiment of the present application, by setting up multiple uplinks, judging the link status according to the load rate of each link, and dynamically selecting an uplink for a new network session service, multiple links can work at the same time, which increases the load capacity , while ensuring that all business flows work at full speed, and there is no need to limit the speed of some data flows. In this application, by monitoring the link traffic and performing link selection or switching in units of sessions, the maximum possible guarantee is that the service will not be interrupted and the user experience will be good.

Several embodiments of the present application have been described above with reference to the accompanying drawings, which are not intended to limit the scope of the right of the present application. Any modifications, equivalent substitutions and improvements made by those skilled in the art without departing from the scope and essence of the present application shall fall within the scope of the right of the present application.

Claims

A load balancing method includes the following steps:

Receive new network session services;

Find out the first non-overloaded link from the multiple uplinks according to the set search sequence, and assign the first non-overloaded link to the new network session service;

According to the reverse order of the set search sequence, the first idle link is sequentially searched from the multiple uplinks, and a network session on a non-idle link preceding the first idle link is migrating to the first idle link;

When there is no session service on the last non-idle link of the found first idle link, the last non-idle link is closed.
The load balancing method according to claim 1, wherein the set search order includes a numbering order of the plurality of uplinks.
The load balancing method of claim 1, further comprising:

When no non-overloaded link is found from the plurality of uplinks according to the set search sequence, the last link of the plurality of uplinks is allocated to the new network session.
The load balancing method according to claim 1, wherein in the step of finding the first non-overloaded link from the plurality of uplinks according to the set search sequence and the setting according to the setting Before the step of finding out the first idle link from the plurality of uplinks in the reverse order of the search sequence, the method further includes the step of updating the load status of the plurality of uplinks, the updating of the plurality of uplinks. Steps to check the uplink load status, including:

updating the load status of each link except the last link in the plurality of uplinks;

When the link load rate is 0, update the load state of the link to a closed state;

When W1≥link load rate>0, update the load state of the link to an idle state;

When Wh≥link load rate>W1, update the load state of the link to a light load state;

When the link load rate>Wh, update the load state of the link to the overload state, until the link load rate drops to ≤W1 and continue for a period of time, update the load state of the link to the idle state ;

Wherein W1 represents the preset load idle reference value; Wh represents the preset load overload reference value.
The load balancing method according to claim 1, wherein the first idle link is searched out from the plurality of uplinks according to the reverse order of the set search order, and the first idle link is The migration of a network session on the last non-idle link of the link to the first idle link includes:

Set the first idle link as the default link of the network session to be migrated, delete the record entry of the network session to be migrated, re-initiate the network session to be migrated as a new network session, and assign to the default link.
The load balancing method according to claim 1, wherein the first idle link is found from the plurality of uplinks according to the reverse order of the set search order, and the first idle link is The step of migrating a network session on the last non-idle link of the link to the first idle link is performed periodically according to a set period.
A load balancing device, comprising:

a network session service receiving module, configured to receive a new network session service;

The non-overloaded link search module is set to find the first non-overloaded link from multiple uplinks according to the set search sequence;

a network session service allocation module, configured to allocate the first non-overloaded link to the new network session service;

an idle link search module, configured to search for the first idle link from the plurality of uplinks in turn according to the reverse order of the set search sequence;

a network session migration module, configured to migrate a network session on the last non-idle link of the first idle link to the first idle link;

The link closing module is configured to close the last non-idle link when there is no session service on the last non-idle link of the found first idle link.
The load balancing apparatus of claim 7, further comprising:

a link load status update module, configured to find the first non-overloaded link from the plurality of uplinks and the idle link in the step of the non-overloaded link search module according to a set search sequence Before the step of finding the first idle link from the plurality of uplinks according to the reverse order of the set search sequence, the road searching module updates the load status of the plurality of uplinks.
A communication device, comprising a memory, a processor, and a program stored on the memory and executable on the processor, wherein, when the program is executed by the processor, any one of claims 1-6 is implemented The steps of the load balancing method.
A storage medium, wherein the storage medium stores at least one program, and the at least one program can be executed by at least one processor to implement the steps of the load balancing method according to any one of claims 1-6.