WO2017024824A1

WO2017024824A1 - Aggregated link-based traffic management method and device

Info

Publication number: WO2017024824A1
Application number: PCT/CN2016/081761
Authority: WO
Inventors: 胡军; 王亚朋; 姜振宇
Original assignee: 中兴通讯股份有限公司
Priority date: 2015-08-11
Filing date: 2016-05-11
Publication date: 2017-02-16
Also published as: CN106453111B; CN106453111A

Abstract

Provided are an aggregated link-based traffic management method and device. After allocating a rate-limited rate to each of member ports of a link aggregation group with respect to a feature flow of the link aggregation group, statistics are compiled on traffic of the feature flow of each of the member ports. If it is determined, according to a statistical result, that an actual sudden abnormality probability η_sudden of the feature flow of the link aggregation group is greater than a sudden abnormality probability threshold λ_sudden, and if the rate-limited rate of at least one of the member ports of the link aggregation group is under capacity, a rate-limited rate of a member port of the link aggregation group is adjusted. After allocating rate-limited rates to member ports of a link aggregation group, the above technical solution can reasonably re-adjust, on the basis of an actual traffic situation of each member port, a rate-limited rate of a feature flow of each member port, thereby making full use of bandwidth resources of each member port, enhancing a forwarding efficiency of the feature flow, and improving QoS guarantees.

Description

Aggregation link based traffic management method and device

Technical field

This document relates to, but is not limited to, the field of communications, and relates to a method and apparatus for traffic management based on an aggregated link.

Background technique

As the network scale continues to expand, operators are placing increasing demands on the bandwidth and reliability of backbone links. In the conventional technology, it is common to replace a high-speed interface board or replace a device that supports a high-rate interface board to increase bandwidth, but this solution requires a high cost and is not flexible enough. The link aggregation technology can be used to increase the link bandwidth by bundling multiple physical interfaces into one logical interface without hardware upgrade. Moreover, the mechanism of applying load balancing and backup links on the interface can further effectively improve the reliability of links between devices.

However, the device port resources are always limited. As long as the user robs the network resources, network congestion will occur. In severe cases, packet loss will still occur, resulting in degraded service quality or even unavailable. Therefore, how to increase the bandwidth of the link aggregation technology while ensuring the quality of service (QoS) has become a hotspot in the industry. The speed limit based on the link aggregation group is one of the important topics.

The rate limit is applied to the link aggregation group. That is, the rate of packets on each member port is monitored. The total rate is limited to a reasonable range to prevent network congestion caused by a large number of users. The link aggregation load balancing is the same as the bandwidth rate of the actual member ports. The same bandwidth is balanced on the ports of different bandwidths. In addition, the link aggregation application traffic rate limit and the total rate limit CIR (Committed Information) Rate, agreed rate) values are averaged or randomly assigned on each member port, which is bound to cause the member ports to load differently. For example, some member ports have sufficient bandwidth, but the actual data traffic exceeds the allocated rate, and the data is severely blocked, delayed, or even lost. Some member ports are idle, and the data flow carried is far lower than the allocated speed limit. The actual traffic sum of each member port passing through the link aggregation group is always lower than the total speed limit value of the set link aggregation group. Therefore, how to manage the rate limit value of each member port of the aggregation link group to effectively and reasonably utilize the broadband resources is a problem to be solved.

Summary of the invention

The following is an overview of the topics detailed in this document. This Summary is not intended to limit the scope of the claims.

The embodiments of the present invention provide a traffic management method and device based on an aggregated link, and manage the rate limit rate of the member ports of the aggregation link group, thereby realizing effective and rational use of the broadband resources.

The embodiment of the invention provides a traffic management method based on an aggregated link, including:

Assigning a rate limit rate to each member port of the link aggregation group for any characteristic flow of the link aggregation group;

Counting the characteristic flow of each member port;

Determining, according to the statistical result, that the actual abnormal burst probability η _sudden of the characteristic flow of the link aggregation group is greater than the abnormal burst probability threshold λ _sudden , and the rate limit rate of at least one member port in the link aggregation group is surplus The rate limit of the member ports of the link aggregation group is adjusted.

Optionally, the calculating the traffic of the feature flow of each member port includes:

_Counting the actual passing traffic rate R _{pass_ij} and the actual discarding traffic rate R _{discard_ij} of the feature stream of each member port in the consecutive k statistical periods T;

Where i is a member port, i is greater than or equal to 1 and less than or equal to the number of member ports n of the link aggregation group; the j represents a statistical period, j is greater than or equal to 1 and less than or equal to the k, The k is greater than or equal to 2.

Optionally, the method further includes: determining, according to the statistical result, whether the actual abnormal burst probability η _sudden of the feature stream of the link aggregation group is greater than an abnormal burst probability threshold λ _sudden includes:

The determining, according to the statistical result, whether the actual abnormal burst probability η _sudden of the feature stream of the link aggregation group is greater than an abnormal burst probability threshold λ _sudden includes:

Obtaining a total actual throughput R _{pass_j} and a total actual discard amount R _{discard_j} of the feature stream of the link aggregation group in the k statistical periods; wherein

R _{pass_j} =∑ _i=1,2,...,n R _{pass_ij} ;

R _{discard_j} =∑ _i=1,2,...,n R _{discard_ij} ;

Obtaining, in the k statistical periods, the number of times that the burst loss rate σ _{discard is} greater than the packet loss rate threshold λ _discard in the k statistical periods;

Determining whether the actual abnormal burst probability η _sudden of the characteristic stream of the link aggregation group is greater than an abnormal burst probability threshold λ _sudden ;

Optionally, the method further includes: determining, according to the statistics, whether the rate limit rate of the member ports of the link aggregation group is:

Obtain the current actual rate of the member port i

among them,

It is determined whether the R _{vc —} i is smaller than a currently set rate limit rate R _{carCur —} i of the member port i , and if yes, determining a rate limit rate surplus of the member port i.

Optionally, assigning a rate limit rate to each member port of the link aggregation group for a feature flow of the link aggregation group includes:

Initially, setting a total agreed rate R _{car_config} of the feature stream of the link aggregation group;

The total agreed rate divided by the link aggregation group _R car_config member to obtain the n number of ports each member of the rate limit _R carCur_i port, i represents the port member, the value i is greater than or equal to 1 and an integer less than or equal to the number n of member ports of the link aggregation group.

Optionally, the adjusting the rate limit of the member ports of the link aggregation group includes:

The sum of the difference between the current rate limit rate of the member ports of the link aggregation group and the current actual rate in the link aggregation group is used as the total rate limit rate of the link aggregation group, and The current rate limit rate of the member port of the link aggregation group that is limited to the rate limit is updated to the current actual rate of the member port.

Allocating the total margin rate rate to the member ports of the link aggregation group except the member ports with the rate limit rate redundancy.

Optionally, the total margin rate rate is allocated to the link aggregation group except the speed limit. Member ports outside the rate-capable member ports include:

And selecting, from the member of the link aggregation group, a member port that is greater than the current rate limit of the member port except the member port with the rate limit rate;

The total margin rate rate amount is allocated to the selected member port.

Optionally, the allocating the total margin rate rate amount to the selected member port includes:

Allocating the total surplus rate rate amount to the selected member port on average;

or,

The selected member ports are sequentially arranged in the order of the rate shortage, wherein the rate shortage is equal to the difference between the current actual rate of the member port and the currently set rate limit rate, or is a member port. Average actual drop traffic rate;

The total margin rate rate is compared with each rate shortage according to the rate shortage amount from small to large or from large to small, if the current total limited rate rate is greater than or equal to the current comparison. If the rate is short, the current rate limit rate of the member port corresponding to the rate shortage is updated to the original rate limit rate plus the rate shortage, and the current total rate limit rate is compared. After comparing the rate shortage of the current comparison with the next rate shortage; if the current ratio of the total margin rate of the comparison is less than the rate of the current comparison, or the rate of the current comparison is the last one, then The current rate limit rate of the member port corresponding to the rate shortage is directly updated to the original rate limit rate plus the total remaining rate limit rate currently being compared.

The embodiment of the invention further provides a device for managing traffic based on an aggregated link, comprising:

The configuration module is configured to allocate a rate limit rate to each member port of the link aggregation group for a feature flow of the link aggregation group;

a statistics module, configured to collect statistics on the traffic of the feature stream of each member port;

a management module, configured to determine, according to the statistical result of the statistics module, that the feature flow actual abnormality probability η _{sudden of} the link aggregation group is greater than an abnormal burst probability threshold λ _sudden , and the link aggregation group When the rate limit of at least one member port is limited, the rate limit of the member ports of the link aggregation group is adjusted.

Optionally, the statistics module includes a traffic statistics sub-module and a drop traffic statistics sub-module;

The traffic statistics sub-module is configured to count the actual traffic rate R _{pass_ij} of the feature stream of each member port in the consecutive k statistical periods T;

The discarding traffic statistics sub-module is configured to count the actual discard traffic rate R _{discard_ij} of the feature stream of each member port in the consecutive k statistical periods T;

Optionally, the management module includes an abnormality determining sub-module, configured to acquire a total actual throughput R _{pass_j} and a total actual discard amount R _{discard_j} of the feature stream of the link aggregation group in the k cycles; ,

R _{pass_j} =∑ _i=1,2,...,n R _{pass_ij} ;

R _{discard_j} =∑ _i=1,2,...,n R _{discard_ij} ;

And determining whether the actual abnormal burst probability η _sudden of the characteristic stream of the link aggregation group is greater than an abnormal burst probability threshold

The management module includes a redundancy judgment sub-module, and is configured to acquire a current actual rate of the member port i.

among them,

And determining whether the R _{vc —} i is less than a current rate limit rate R _{carCur —} i of the member port i, and if yes, determining a rate limit rate of the member port i.

Optionally, the management module includes a calculation submodule and an adjustment submodule.

The calculation submodule is configured to: the member ports of the link aggregation group with a limited rate rate redundancy The sum of the difference between the currently set rate limit rate and the current actual rate is used as the total margin rate rate of the link aggregation group, and the current limit of the member ports of the link aggregation group with the rate limit rate is redundant. The set rate limit rate is updated to the current actual rate of the member port;

The adjustment submodule is configured to allocate the total margin rate rate amount to a member port of the link aggregation group except the member port with the rate limit rate redundancy.

Optionally, the adjustment submodule includes a selection unit and an allocation unit;

The selecting unit is configured to select, from the member ports of the link aggregation group, the member ports other than the member ports of the rate limiting rate, that the current actual rate is greater than the member port whose current rate limit is set;

The allocation unit is configured to allocate the total margin rate rate amount to the selected member port.

Optionally, the allocating unit includes an average allocation subunit or a comparison allocation subunit;

The average allocation subunit is configured to evenly distribute the total margin rate rate amount to each selected member port;

The comparison allocation sub-unit is configured to: arrange the selected member ports in descending order of the rate shortage, wherein the rate shortage is equal to the current actual rate of the member port and the currently set rate limit. The difference between the rates is the average actual discarding traffic rate of the member ports; the total margin rate rate is compared with each rate shortage according to the rate shortage from small to large or from small to small. If the current rate limit rate of the current comparison is greater than or equal to the rate of the current comparison, the current rate limit of the member port corresponding to the rate shortage is updated to the original rate limit plus the rate shortage. And comparing the current surplus rate limit rate of the current comparison with the current rate rate shortage amount, and comparing with the next rate shortage amount; if the current total rate limit rate of the comparison is smaller than the current comparison The rate shortage, or the current comparison rate shortage is the last one, directly the current setting of the member port corresponding to the rate shortage The updated rate limit for the primary rate limit plus the current margin rate limit for the total amount of the comparison.

The embodiment of the invention further provides a computer readable storage medium, where the computer readable storage medium stores computer executable instructions, and the computer executable instructions are implemented based on the aggregation Link traffic management method method.

The beneficial effects of the embodiments of the present invention are:

The aggregated link-based traffic management method and device provided by the embodiment of the present invention are configured to collect a rate limit rate for each member port of the link aggregation group, and collect a rate limit for each member port of the link aggregation group. The feature stream streamA traffic, when it is determined according to the statistical result that the actual abnormal burst probability η _{sudden of} the feature stream streamA of the link aggregation group is greater than the abnormal burst probability threshold λ _sudden , and the rate limit of at least one member port in the link aggregation group When the rate is surplus, the rate limit of the member ports of the link aggregation group is adjusted. In the embodiment of the present invention, after the rate limit is configured for the member ports of the link aggregation group, the rate limit of each member port feature stream streamA is re-adjusted based on the actual traffic of each member port, so that each of the member ports can be fully utilized. Bandwidth resources of member ports improve the forwarding efficiency of feature stream streamA and improve QoS guarantee.

Other aspects will be apparent upon reading and understanding the drawings and detailed description.

DRAWINGS

1 is a schematic flowchart of a method for managing traffic based on an aggregate link according to Embodiment 1 of the present invention;

2 is a schematic structural diagram of an aggregate link-based traffic management apparatus according to Embodiment 2 of the present invention;

3 is a schematic diagram of a network topology structure according to Embodiment 3 of the present invention;

4 is a schematic diagram of statistical data storage provided by Embodiment 3 of the present invention;

FIG. 5 is a schematic diagram of an adjustment effect according to Embodiment 3 of the present invention.

Detailed ways

The embodiments of the present invention are further described in detail below with reference to the accompanying drawings.

Embodiment 1:

Referring to FIG. 1 , the aggregate link-based traffic management method provided in this embodiment includes:

Step 101: Allocating a rate limit rate to each member port of the link aggregation group for a feature stream (such as streamA) of the link aggregation group;

Before the step 101, the pre-configuration phase may be further configured to enable or disable the link aggregation group-based rate-limiting adaptive adjustment function by configuring the corresponding configuration device, that is, configuring the aggregation link-based traffic management to be enabled or disabled;

A feature stream in step 101 refers to a matching type such as a packet type (such as broadcast broadcast, unicast unicast, etc.), a packet priority (such as DSCP (Differentiated Services Code Point), priority Precedence, A series of packet flows, such as an access control list (ACL), and an access control list (ACL); in this step, initially, a feature stream streamA for the link aggregation group is the link aggregation group. The rate limit rate of each member port can be allocated in an average manner, or it can be allocated by random allocation or other allocation methods.

The following is an example of the average allocation and random allocation:

1. The method of equal distribution:

Set the total agreed rate R _{car_config of} the feature stream streamA of the link aggregation group;

_R car_config total committed rate divided by the number of members of the link aggregation group to give each member port n rate limit _R carCur_i port, where i represents the port member, the value i is greater than or equal to 1 and less than or equal to the link The integer number of member ports of the aggregation group.

2, random allocation:

Total agreed rate

Each member port assigned to the link aggregation group is randomly split to obtain the rate limit rate R _{carCur_i of} each member port.

Regardless of which distribution method is used, the following principles are followed:

R _{carCur_1} +R _{carCur_2} +...R _{carCur_n} =R _{car_config} ;

In the subsequent stage, the rate limit rate allocated to each member port of the link aggregation group in the foregoing step 101 may be the rate limit rate adjusted by the subsequent step 103;

Step 102: Count the traffic of the feature stream streamA of each member port of the link aggregation group;

The traffic of the feature stream streamA, which is calculated in this embodiment, includes, but is not limited to, the actual traffic rate and the actual traffic rate of the feature stream streamA of each member port. In this embodiment, the actual situation of the network and the adjustment are ensured. The result is in line with the actual demand, and the actual passing traffic rate R _{pass_ij} and the actual discarding traffic rate R _{discard_ij} of each member port feature stream streamA in the consecutive k statistical periods T are _counted ; where i is a member port, i is greater than or equal to 1 and less than Or equal to the number of member ports of the link aggregation group, n; j represents a statistical period, j is greater than or equal to 1 and less than or equal to k, and k is greater than or equal to 2. The consecutive k statistical periods T in this embodiment may refer to k cycles that are completely re-stated, and the k-1 cycles before statistics may be multiplexed. The following examples are respectively illustrated:

Completely re-statistics: For example, if n is 3, then after counting the three statistical periods t1, t2, and t3, the next round needs to wait for the statistics of three statistical periods t4, t5, and t6 to complete the next round of operations;

K-1 cycles before multiplexing: If n is still assumed to be 3, then after counting the three statistical periods t1, t2, and t3, the next round only needs to wait for the t4 statistical period statistics to be completed, then t2 and t3 can be used. And t4 three statistical cycles to perform the next round of operations;

The specific value of k can be flexibly set according to specific network conditions and specific application scenarios.

Step 103: When it is determined according to the statistical result, the characteristic abnormality probability η _{sudden of} the characteristic stream streamA of the link aggregation group is greater than the abnormal burst probability threshold λ _sudden , and the rate limit rate of at least one member port in the link aggregation group is surplus The rate limit of the member ports of the link aggregation group is adjusted.

Optionally, the method further includes: determining, according to the statistical result, whether the feature flow actual abnormality probability η _{sudden of the} link aggregation group is greater than an abnormal burst probability threshold λ _sudden includes:

Obtaining the total actual throughput R _{pass_j} and the total actual discard amount R _{discard_j} of the link aggregation group feature stream streamA in k periods;

among them,

R _{pass_j} =∑ _i=1,2,...,n R _{pass_ij} ;

R _{discard_j} =∑ _i=1,2,...,n R _{discard_ij} ;

Obtaining the number of occurrences of the characteristic flow of the link aggregation group in the k statistical periods, the burst loss rate σ _{discard is} greater than the number M of the packet loss threshold λ _discard ; wherein

Determining whether the actual abnormal burst probability η _sudden of the characteristic stream of the link aggregation group is greater than an abnormal burst probability threshold λ _sudden , wherein

The value of the abnormal burst probability threshold λ _sudden may also be specifically set according to specific network conditions and specific application scenarios;

Optionally, the method further includes: determining, according to the statistical result, whether a rate limit rate of the member ports of the link aggregation group is redundant;

Obtain the current actual rate R _{vc_i of the} member port i;

among them,

It is determined whether the obtained R _{vc_i} is smaller than the currently set rate limit rate R _{carCur_i of} the member port i, and if yes, the rate limit rate of the member port i is determined to be surplus.

R _{carCur_i} -R _{vc_i} is the margin of the rate limit of the member ports; the _{sum of the margins of} such member ports (member ports with limited rate limit) is used as the total margin rate rate, and the member ports are currently The difference between the set rate limit rate and the current actual rate is summed, and the summed result is used as the total surplus rate rate;

On the other hand, if R _{vc_i} >R _{carCur_i} , it is determined that the rate limit rate of the member port i is in short supply. In this example, R _{vc_i} -R _{carCur_i} can be used as the rate shortage of the member port i, and the average actual value of such member ports can also be used. Drop rate

As a rate shortage.

Optionally, in step 103, adjusting the rate limit of the member ports of the link aggregation group includes:

The sum of the difference between the current rate limit rate of the member ports of the link aggregation group and the current actual rate in the link aggregation group is used as the total rate limit rate of the link aggregation group, and The currently set rate limit rate of the member ports of the link aggregation group with the rate limit rate surplus is updated as described above. The current actual rate of the member port;

The member ports of the link aggregation group except the member ports with the rate limit rate redundancy are included in the link aggregation group:

And selecting, from the member of the link aggregation group, a member port that is greater than the current rate limit rate of the member port except the member port with the rate limit rate;

The total margin rate rate amount is allocated to the selected member port. Its distribution methods include but are not limited to:

Allocation method 1: Allocate the total surplus rate rate to the member port with the selected rate being evenly distributed;

Allocation method 2: The selected member ports are arranged in order from the smallest to the largest;

The total margin rate rate is compared with each rate shortage according to the rate shortage amount from small to large or from large to small, if the current total limited rate rate is greater than or equal to the current comparison. If the rate is short, the current rate limit rate of the member port corresponding to the rate shortage is updated to the original rate limit rate plus the rate shortage, and the current total rate limit rate is compared. After comparing the rate shortage of the current comparison with the next rate shortage;

If the current rate limit rate of the current comparison is smaller than the rate of the current comparison, the current rate limit set by the member port corresponding to the rate shortage is directly updated to the original rate limit plus the total margin limit. The speed rate, the member port of the subsequent rate shortage is no longer updated;

If the current rate of shortage is the last one, the current rate limit set by the member port corresponding to the rate shortage is updated to the original rate limit rate plus the total margin rate.

Regardless of the above allocation method, after updating, the rate limit rate after updating each member port still satisfies the following relationship:

R _{carCur_1} +R _{carCur_2} +...R _{carCur_n} =R _{car_config} .

It can be seen that for the link aggregation application port rate limit, the total rate limit CIR value based on the feature stream is used in each A member port is fixedly allocated in a certain way, causing some member ports to have sufficient bandwidth. However, the actual data traffic exceeds the allocated rate quota, the data is severely blocked, and packets are lost. Some member ports are idle, and the data flow is far lower. In this embodiment, the actual rate of traffic passing through the feature flow of the forwarding plane is periodically collected, and the traffic is discarded, and the speed limit adjustment criterion and adjustment mode of the member ports in this embodiment are used. The redistributed member port is based on the rate-based traffic of the feature stream. The reserved rate limit of the idle member port is allocated to the member ports with large traffic and heavy packet loss. The rate of bandwidth is increased and the packet loss rate is effectively reduced. When the rate limit value is unchanged, the forwarding efficiency of the feature stream is improved, and the bandwidth resource is further utilized to improve the QoS guarantee.

Embodiment 2:

This embodiment provides a traffic management device based on an aggregated link. Referring to FIG. 2, the method includes:

The configuration module 1 is configured to allocate a rate limit rate to each member port of the link aggregation group for the link aggregation group-feature stream streamA;

A feature stream streamA in this embodiment refers to a matching type such as a packet type (such as broadcast, unicast, etc.), a packet priority (such as DSCP, Precedence, 802.1p, etc.), an admission control list ACL, and the like. A series of packet flows; the configuration module 1 allocates a rate limit rate to each member port of the link aggregation group for a feature stream streamA of the link aggregation group, and may adopt an average allocation manner, or may adopt a random allocation or other allocation manner. Make an assignment.

The following is an example of the average allocation and random allocation:

1. The method of equal distribution:

First, set the total agreed rate R _{car_config of} the feature stream streamA of the link aggregation group;

_R car_config total committed rate divided by the number of members of the LAG port n to give the members of the rate limit _R carCur_i port, where i represents the port member, the value i is greater than or equal to 1 and less than or equal to the chain The number of member ports of the road aggregation group is an integer n.

2, random distribution method:

Total agreed rate

Each member port assigned to the link aggregation group after random splitting obtains the rate limit rate R _{carCur_i of} each member port.

R _{carCur_1} +R _{carCur_2} +...R _{carCur_n} =R _{car_config} ;

The statistics module 2 is configured to collect the feature stream streamA traffic of each member port; optionally, the statistics module 2 includes a traffic statistics sub-module 21 and a drop traffic statistics sub-module 22;

The traffic statistics sub-module 21 is configured to count the actual traffic rate R _{pass_ij} of the feature stream streamA of each member port in the k consecutive statistical periods T;

The discarding traffic statistics sub-module 22 is configured to count the actual discarding traffic rate R _{discard_ij} of each member port feature stream streamA in consecutive k statistical periods T;

Where i is a member port, i is greater than or equal to 1 and less than or equal to the number of member ports n of the link aggregation group; the j represents a statistical period, j is greater than or equal to 1 and less than or equal to the k, k is greater than or equal to 2.

In this embodiment, the traffic flow statistics sub-module 21 and the drop traffic statistics sub-module 22 are respectively set to count the statistical flow of each member port in the k consecutive statistical periods T, in order to ensure that the adjustment result meets the actual requirements. The actual traffic rate R _{pass_ij} and the actual discard traffic rate R _{discard_ij of the stream A} ; the consecutive k statistical periods T in this embodiment may refer to k cycles that are completely re- _stated , and the previously counted k-1 cycles may be multiplexed.

The management module 3 is configured to determine, according to the statistical result of the statistical module 2, that the feature stream streamA actual abnormal burst probability η _{sudden of the} link aggregation group is greater than the abnormal burst probability threshold λ _sudden , and the rate limit of at least one member port of the member port When the rate is surplus, adjust the rate limit of the member ports of the link aggregation group.

Optionally, the management module 3 includes an abnormality determining sub-module 31 configured to acquire the total actual throughput R _{pass_j} and the total actual discard amount R _{discard_j} of the link aggregation group feature stream streamA in k cycles;

Where R _{pass_j} = ∑ _{i = 1, 2, ..., n} R _{pass_ij} ;

R _{discard_j} =∑ _i=1,2,...,n R _{discard_ij} ;

Obtaining the number of occurrences of the burst loss rate of the characteristic flow of the link aggregation group in the k statistical periods σ _{discard is} greater than the number M of the packet loss rate threshold λ _discard ; wherein

And determining whether the actual abnormal burst probability η _sudden of the characteristic stream of the link aggregation group is greater than an abnormal burst probability threshold λ _sudden ;

The value of the abnormal burst probability threshold λ _sudden can also be specifically set according to specific network conditions and specific application scenarios.

Optionally, the management module 3 includes a redundancy judgment sub-module 32 configured to acquire the current actual rate R _{vc_i of the} member port i.

among them,

And determining whether R _{vc —} i is less than a current rate limit rate R _{carCur —} i of the member port i , and if yes, determining a rate limit rate of the member port i.

R _{carCur_i} -R _{vc_i} is the margin of the rate limit rate of the member ports with the limited rate rate; the _{sum of the margins of} the member ports (all member ports with the limited rate rate surplus) is used as the total margin rate rate The sum of the current rate limit set by the member port and the current actual rate difference is also taken as the total margin rate rate.

On the other hand, if R _{vc_i} >R _{carCur_i} , it is determined that the rate limit rate of the member port i is in short supply. In this example, R _{vc_i} -R _{carCur_i} can be used as the rate shortage of the member port i, and such member ports can also be used. Average actual drop rate for all member ports that are in short rate

As a rate shortage.

Optionally, the management module 3 further includes a calculation submodule 33 and an adjustment submodule 34.

The calculation sub-module 33 is configured to use, as the total surplus rate limit of the link aggregation group, the sum of the difference between the current rate limit rate of the member ports of the link aggregation group and the current actual rate. Rate the rate, and update the current rate limit rate of the member ports of the link aggregation group that are limited by the rate limit to the current actual rate of the member port;

The adjustment submodule is set to allocate the total margin rate rate amount to the link aggregation group except the speed limit. A member port outside the rate member port.

Optionally, the adjustment submodule specifically includes a selection unit and an allocation unit;

The selecting unit is configured to select, from the member ports of the link aggregation group, the member ports other than the member ports with the rate limit remaining; the current actual rate is greater than the member port whose current rate limit is set;

The allocation unit is arranged to allocate a total margin rate rate amount to each of the selected member ports.

Optionally, the allocation unit comprises an average allocation subunit or a comparison allocation subunit;

The average allocation subunit is set to evenly allocate the total surplus rate rate amount to each of the selected member ports;

The comparison allocation subunit is configured to sequentially arrange the selected member ports in descending order of the rate shortage, wherein the rate shortage is equal to the difference between the current actual rate of the member port and the currently set rate limit rate. The value is the average actual discarding traffic rate of the member port; the total margin rate rate is compared with each rate shortage according to the rate shortage amount from small to large or large to small, if the current current If the ratio of the total rate limit of the comparison is greater than or equal to the rate of the current comparison, the current rate limit of the member port corresponding to the rate shortage is updated to the original rate limit plus the rate shortage, and Comparing the current surplus rate limit rate of the current comparison with the current rate rate shortage amount and comparing with the next rate shortage amount;

R _{carCur_1} +R _{carCur_2} +...R _{carCur_n} =R _{car_config} .

Embodiment 3:

The embodiments of the present invention are further described below in conjunction with a specific application scenario.

For the network topology of this application scenario, see Figure 3. Routers R1 and R2 are connected through three Gigabit Ethernet ports (gei). Terminals PC1 and PC2 are connected to R1 and R2 through the megabit Ethernet port (fei). The network segment where PC1 and R1 are configured is 1.1.1.x/24, the network segment for PC2 and R2 is 3.3.3.x/24, and the network segment for routing between R1 and R2 is 2.2.2.x/24. The (Open Shortest Path First) protocol is connected, and the link aggregation between the ports is Inf _SG , thus realizing the data transmission between PC1 and PC2. The traffic management process for the network topology shown in Figure 3 is as follows:

In the configuration phase, configure the terminal to enable the rate-based adaptive adjustment function based on the link aggregation group.

2. The user configures the rate limit based on the aggregation port (Inf _SG ). The total agreed rate (CIR) of the feature stream streamA of the port is R _{car_config} ; then the average pre-allocation according to the number of member ports of the link aggregation group is n. The rate limit of each member port, including the following:

Where i is a member port, i is greater than or equal to 1 and less than or equal to the number of member ports of the link aggregation group n;

3. For the statistical node based on the feature flow that is established, periodically calculate the actual traffic rate R _{pass_ij} and the actual drop traffic rate R _{discard_ij} of each member port in the link aggregation group. For the collection interval T, the specific value can be flexibly set according to the actual application scenario, but it should not be set too small. Otherwise, the occasional traffic burst will cause frequent adjustment of the rate limit, which will cause fluctuations in traffic forwarding. Large, otherwise the long-term traffic burst will not be adjusted and the traffic will be seriously blocked. In this embodiment, the actual traffic rate R _{pass_ij} and the actual discard traffic rate R _{discard_ij} are obtained by the ratio of the data packet actually forwarded and discarded in each period T to the set period time interval T, _respectively .

4. Obtain statistics of consecutive k feature streams streamA. The statistical information and storage form to be stored is roughly as shown in Figure 4:

In the storage structure shown in area A of FIG. 4, the total agreed rate R _{car_config of the} stream type streamA configuration under the aggregation port Inf _SG is stored;

In the storage structure shown in area B of FIG. 4, the total actual throughput R _{pass_j} of the forwarding surface feature stream streamA collected by the storage device for the last consecutive k periods T, the total actual discard amount R _{discard_j} , where j represents a statistical period, and j is greater than Or equal to 1 and less than or equal to k; the total actual throughput R _{pass_j} collected in any period, the total actual discard amount R _{discard_j} is the passing rate and discard rate of the feature stream streamA collected by each member port in the current period. And, as follows:

R _{pass_j} =∑ _i=1,2,...,n R _{pass_ij} ;

R _{discard_j} =∑ _i=1,2,...,n R _{discard_ij} ;

In the storage structure shown in area C of FIG. 4, the rate limit rate R _{carCur_i of the} stream type streamA is defined for each member port of the aggregation port Inf _SG , and an initial pre-allocation phase is created.

See the formula.

The storage structure shown in area D of FIG. 4, the actual passing traffic rate R _{pass_ij} and the actual discarding traffic rate R _{discard_ij} of the forwarding surface feature stream streamA collected by the storage device in the latest consecutive k statistical periods T on each member port;

As shown in Figure 5, the aggregation port Inf _SG is bound to three member ports inf ₁ , inf ₂ , and inf ₃ . In the pre-adjustment state, the average rate of the aggregation port is averaged to three member ports, that is, each. The rate limit for each port is R _{car_config} /3. When the adjustment condition is met, the internal rate limit adjustment algorithm is enabled, and the rate limit value R _{carCur_i} of each member port in the link aggregation group is redistributed, and the excess rate rate quota of the idle member port is allocated to the traffic volume and packet loss. Severe member ports, after adjustment, still need to meet R _{carCur_0} + R _{carCur_1} +...R _{carCur_n} =R _{car_config} , that is, the total speed limit rate of each member port is unchanged.

Here you need to meet the following two conditions to trigger the adjustment:

1) Aggregation port Inf _SG member port stream streamA forwarding delay, blocking, and packet loss are serious;

In this embodiment, the following measurement methods can be adopted:

When k cycles are continuously collected, whether the actual abnormal burst probability η _{sudden of the} feature stream streamA is greater than the abnormal burst probability threshold λ _sudden ;

When the actual abnormal burst probability η _{sudden is} greater than the abnormal burst probability threshold λ _sudden exceeds this threshold, it indicates that the current feature stream streamA is continuously bursting, and data packet loss is serious; M is the link aggregation group at k The number of times when the burst loss rate σ _{discard is} greater than the packet loss rate threshold λ _discard occurs in the period.

2) The aggregation port Inf _SG has one or more member ports. The original finite rate rate quota has a vacancy. That is, the actual semaphore stream stream forwarding rate of some member ports is lower than the allocated rate limit rate.

In this embodiment, it can be measured in the following ways:

During the k acquisition periods T, there is a rate limit rate R _{carCur_i of the} member port feature stream _streamA compared to the average _throughput rate

Average drop rate

And R _{vc_i is} still abundant, among them,

Among them, the speed limit adjustment algorithm logic involved is as follows:

Step 1. For traversing the member ports, the member ports are averaged at each discard rate.

Size, from small to large.

Step 2.R _abundant =0; / / each member port surplus speed limit quota

In the present example, the above adjustment process, the state before adjustment shown in FIG. 5 represent the rate limit pre-assigned port on the front of each member of _R carCur_i = average rate adjustment speed _R car_avr; k statistics statistical state is shown in the periods T Each member port collects the actual passing traffic rate R _{pass_ij} and the actual discarding traffic rate R _{discard_ij} ; the adjusted state indicates that when the data displayed in the statistical result state satisfies the two conditions required for the adjustment, the above adjustment algorithm is used for each After the member port rate limit rate is redistributed, R _{carCur_1} +R _{carCur_2} +R _{carCur_3} = R _{car_config is} still satisfied after the allocation.

It can be seen that, in the embodiment of the present invention, the rate limit of the associated link aggregation application is fixed, and the CIR value of the feature stream is fixedly allocated on each member port in any manner, so that some member ports have sufficient bandwidth, but actual data. The traffic exceeds the allocated rate quota, the data is severely blocked, the packet is lost, some member ports are idle, and the data flow carried is far lower than the allocated rate limit. By periodically collecting the traffic passing through the feature flow of the forwarding plane, discarding the traffic, and combining the rate limit adjustment rule and the adjustment mode of the member in this embodiment, the rate limit of the member port based on the feature flow is reasonably redistributed, and the margin of the idle member port is limited. The fast traffic quota is allocated to the member ports with large traffic and heavy packet loss, increasing the bandwidth of the pass rate, and effectively reducing the packet loss rate. This improves the forwarding efficiency of the feature stream when the overall rate limit is unchanged. Utilize bandwidth resources to improve QoS guarantee.

The embodiment of the present invention further provides a computer readable storage medium, where the computer readable storage medium stores computer executable instructions, and when the computer executable instructions are executed, implements an aggregate link based traffic management method.

One of ordinary skill in the art will appreciate that all or a portion of the above steps may be performed by a program to instruct related hardware, such as a processor, which may be stored in a computer readable storage medium, such as a read only memory, disk or optical disk. Wait. Alternatively, all or part of the steps of the above embodiments may also be implemented using one or more integrated circuits. Correspondingly, each module/unit in the above embodiment may be implemented in the form of hardware, for example, by implementing an integrated circuit to implement its corresponding function, or may be implemented in the form of a software function module, for example, executing a program stored in the memory by a processor. / instruction to achieve its corresponding function. This application is not limited to any specific combination of hardware and software. A person skilled in the art should understand that the technical solutions of the present application can be modified or equivalent, without departing from the spirit and scope of the technical solutions of the present application, and should be included in the scope of the claims of the present application.

Industrial applicability

The above technical solutions achieve efficient and rational use of broadband resources.

Claims

A method for traffic management based on an aggregated link, comprising:

Assigning a rate limit rate to each member port of the link aggregation group for any characteristic flow of the link aggregation group;

Counting the traffic of the feature stream of each member port;

Determining, according to the statistical result, that the actual abnormal burst probability η sudden of the characteristic flow of the link aggregation group is greater than the abnormal burst probability threshold λ sudden , and the rate limit rate of at least one member port in the link aggregation group is surplus The rate limit of the member ports of the link aggregation group is adjusted.
The aggregate link-based traffic management method according to claim 1, wherein the calculating the traffic of the feature stream of each member port comprises:

Counting the actual passing traffic rate R pass_ij and the actual discarding traffic rate R discard_ij of the feature stream of each member port in the consecutive k statistical periods T;

Where i is a member port, i is greater than or equal to 1 and less than or equal to the number of member ports n of the link aggregation group; the j represents a statistical period, j is greater than or equal to 1 and less than or equal to the k, The k is greater than or equal to 2.
The aggregate link-based traffic management method according to claim 2, wherein the method further comprises: determining, according to a statistical result, whether an actual abnormal burst probability η sudden of the characteristic stream of the link aggregation group is greater than an abnormal burst The probability threshold λ sudden includes:

The determining, according to the statistical result, whether the actual abnormal burst probability η sudden of the feature stream of the link aggregation group is greater than an abnormal burst probability threshold λ sudden includes:

Obtaining a total actual throughput R pass_j and a total actual discard amount R discard_j of the feature stream of the link aggregation group in the k statistical periods;

Where R pass_j = ∑ i = 1, 2, ..., n R pass_ij ;

R discard_j =∑ i=1,2,...,n R discard_ij ;

Obtaining, in the k statistical periods, the number of times that the burst loss rate σ discard is greater than the packet loss rate threshold λ discard in the k statistical periods;

Determining whether the actual abnormal burst probability η sudden of the characteristic stream of the link aggregation group is greater than an abnormal burst probability threshold λ sudden , wherein
The aggregate link-based traffic management method of claim 2, the method further comprising: judging whether a rate limit rate of the member ports of the link aggregation group is redundant according to a statistical result; Whether the rate limit rate of the member ports of the link aggregation group is sufficient includes:

Obtain the current actual rate R vc_i of the member port i;

among them,

It is determined whether the R vc — i is smaller than a currently set rate limit rate R carCur — i of the member port i , and if yes, determining a rate limit rate surplus of the member port i.
The aggregate link-based traffic management method according to any one of claims 1 to 4, wherein the characteristic flow for the link aggregation group is a rate limit rate for each member port of the link aggregation group. :

Initially, setting a total agreed rate R car_config of the feature stream of the link aggregation group;

The total agreed rate divided by the link aggregation group R car_config member to obtain the n number of ports each member of the rate limit R carCur_i port, wherein, i represents the port member, the value i is greater than Or an integer equal to 1 and less than or equal to the number n of member ports of the link aggregation group.
The aggregate link-based traffic management method according to any one of claims 1 to 4, wherein the adjusting the rate limit rate of the member ports of the link aggregation group includes:

The sum of the difference between the current rate limit rate of the member ports of the link aggregation group and the current actual rate in the link aggregation group is used as the total rate limit rate of the link aggregation group, and The current rate limit rate of the member port of the link aggregation group that is limited to the rate limit is updated to the current actual rate of the member port.

Allocating the total margin rate rate to the member ports of the link aggregation group except the member ports with the rate limit rate redundancy.
The aggregate link-based traffic management method according to claim 6, wherein the total surplus rate rate is allocated to a member port of the link aggregation group except the member port with the rate limit remaining. include:

And selecting, from the member of the link aggregation group, a member port that is greater than the current rate limit of the member port except the member port with the rate limit rate;

The total margin rate rate amount is allocated to the selected member port.
The aggregate link-based traffic management method of claim 7, wherein the allocating the total margin rate rate amount to the selected member port comprises:

Allocating the total surplus rate rate amount to the selected member port on average;

or,

The selected member ports are sequentially arranged in the order of the rate shortage, wherein the rate shortage is equal to the difference between the current actual rate of the member port and the currently set rate limit rate, or is a member port. Average actual drop traffic rate;

The total margin rate rate is compared with each rate shortage according to the rate shortage amount from small to large or from large to small, if the current total limited rate rate is greater than or equal to the current comparison. If the rate is short, the current rate limit rate of the member port corresponding to the rate shortage is updated to the original rate limit rate plus the rate shortage, and the current total rate limit rate is compared. After comparing the rate shortage of the current comparison with the next rate shortage; if the current ratio of the total margin rate of the comparison is less than the rate of the current comparison, or the rate of the current comparison is the last one, then The current rate limit rate of the member port corresponding to the rate shortage is directly updated to the original rate limit rate plus the total remaining rate limit rate currently being compared.
A traffic management device based on an aggregated link, comprising:

The configuration module is configured to allocate a rate limit rate to each member port of the link aggregation group for a feature flow of the link aggregation group;

a statistics module, configured to collect statistics on the traffic of the feature stream of each member port;

a management module, configured to determine, according to the statistical result of the statistics module, that the feature flow actual abnormality probability η sudden of the link aggregation group is greater than an abnormal burst probability threshold λ sudden , and in the link aggregation group When the rate limit of at least one member port is sufficient, the rate limit of the member ports of the link aggregation group is adjusted.
The aggregate link-based traffic management device of claim 9, wherein the statistics module comprises a traffic statistics sub-module and a drop traffic statistics sub-module;

The traffic statistics sub-module is configured to count the actual traffic rate R pass_ij of the feature stream of each member port in the consecutive k statistical periods T;

The discarding traffic statistics sub-module is configured to count the actual discard traffic rate R discard_ij of the feature stream of each member port in the consecutive k statistical periods T;

Where i is a member port, i is greater than or equal to 1 and less than or equal to the number of member ports n of the link aggregation group; the j represents a statistical period, j is greater than or equal to 1 and less than or equal to the k, The k is greater than or equal to 2.
The aggregate link-based traffic management apparatus according to claim 10, wherein said management module comprises an abnormality determining sub-module, configured to acquire a total of said characteristic flows of said link aggregation group in said k cycles Actual throughput R pass_j and total actual discard amount R discard_j ;

R pass_j =∑ i=1,2,...,n R pass_ij ;

R discard_j =∑ i=1,2,...,n R discard_ij ;

Obtaining, in the k statistical periods, the number of times that the burst loss rate σ discard is greater than the packet loss rate threshold λ discard in the k statistical periods;

And determining whether the actual abnormal burst probability η sudden of the characteristic stream of the link aggregation group is greater than an abnormal burst probability threshold λ sudden , wherein
The aggregate link-based traffic management device of claim 10, wherein the management module includes a redundancy judgment sub-module configured to acquire a current actual rate R vc — i of the member port i;

And determining whether the R vc — i is less than a current rate limit rate R carCur — i of the member port i, and if yes, determining a rate limit rate of the member port i.
The aggregate link-based traffic management apparatus according to any one of claims 9 to 12, wherein the management module comprises a calculation submodule and an adjustment submodule,

The calculation sub-module is configured to use, as a total surplus of the link aggregation group, a sum of a difference between a current rate limit rate of the member ports of the link aggregation group and the current actual rate The rate limit rate is updated, and the current rate limit rate of the member ports of the link aggregation group with the rate limit rate is updated to the current actual rate of the member port;

The adjustment submodule is configured to allocate the total margin rate rate amount to a member port of the link aggregation group except the member port with the rate limit rate redundancy.
The aggregate link-based traffic management apparatus according to claim 13, wherein the adjustment submodule comprises a selection unit and an allocation unit;

The selecting unit is configured to select, from the member ports of the link aggregation group, the member ports other than the member ports of the rate limiting rate, that the current actual rate is greater than the member port whose current rate limit is set;

The allocation unit is configured to allocate the total margin rate rate amount to the selected member port.
The aggregate link-based traffic management apparatus according to claim 14, wherein said allocating unit comprises an average allocation subunit or a comparison allocation subunit;

The average allocation subunit is configured to evenly distribute the total margin rate rate amount to each selected member port;

The comparison allocation sub-unit is configured to: arrange the selected member ports in descending order of the rate shortage, wherein the rate shortage is equal to the current actual rate of the member port and the currently set rate limit. The difference in rate or the average actual drop rate of the member ports; And the total margin rate rate is compared with each rate shortage amount according to the rate shortage amount from small to large or from large to small, if the current ratio of the total surplus rate limit is greater than or equal to the current If the rate is short, the current rate limit of the member port corresponding to the rate shortage is updated to the original rate limit plus the rate shortage, and the current total rate limit is compared. Subtracting the current rate vacancy amount and comparing it with the next rate shortage; if the current total rate limit rate is less than the current rate lag, or the current comparison rate is the last one, Then, the current rate limit rate of the member port corresponding to the rate shortage is directly updated to the original rate limit rate plus the total remaining rate limit rate currently being compared.