WO2016000326A1

WO2016000326A1 - Method and apparatus for scheduling user queue and computer storage medium

Info

Publication number: WO2016000326A1
Application number: PCT/CN2014/087134
Authority: WO
Inventors: 李娟�
Original assignee: 深圳市中兴微电子技术有限公司
Priority date: 2014-06-30
Filing date: 2014-09-22
Publication date: 2016-01-07
Also published as: CN105323189A

Abstract

Disclosed are a method and apparatus for scheduling a user queue, wherein the method comprises: a first scheduler is articulated to a second scheduler and a third scheduler, and a user queue is simultaneously articulated to the second scheduler and the third scheduler; the first scheduler judges whether the second scheduler allows scheduling, and schedules the second scheduler when the second scheduler allows scheduling, so that the second scheduler schedules the articulated user queue via a corresponding shaper; after the user queue articulated under the second scheduler is completely scheduled, the first scheduler determines that the third scheduler allows scheduling, and schedules the third scheduler, so that the third scheduler schedules the articulated user queue via a corresponding shaper; and when the second scheduler does not allow scheduling, the first scheduler determines that the third scheduler allows scheduling, and schedules the third scheduler, so that the third scheduler schedules the articulated user queue via a corresponding shaper.

Description

Method, device and computer storage medium for user queue scheduling

Technical field

The present invention relates to data communication technologies, and in particular, to a method, an apparatus, and a computer storage medium for user queue scheduling.

Background technique

At present, with the rapid development of network communication technology and the deepening and popularization of multimedia technology, higher and more differentiated requirements are put forward for the service quality, performance and diversity flexibility of the network. In order to support services or users with different service requirements, the network is required to distinguish different communications and provide corresponding services. The emergence of Quality of Service (QOS) technology is aimed at solving this problem.

There are many scheduling algorithms for QOS. The most simple polling schedule (RR) is that all users are on a turn-by-turn basis. Everyone has equal opportunities. The polling schedule is suitable for all users and the traffic is exactly the same. Case. For more advanced and flexible applications, a more perfect scheduling algorithm is needed. For example, each user's weight is different, or the allowed traffic is different, and even each user supports different priority queues. A variety of scheduling algorithms have emerged for various applications. For example, for different user weights, a weighted round Robin (WRR) scheduling algorithm, a Deficit Weighted Round Robin (DWRR), and a modified deficit weighted round robin (MDRR) are available. Modified Deficit Round Robin), WRED, and more balanced WFQ scheduling algorithms and more.

In general, scheduling and shaping cooperate to satisfy QOS, scheduling ensures minimum bandwidth, and shaping limits the maximum bandwidth. When the network is congested, the different input service flows of the switching node are respectively scheduled and served according to the specified service rules, so that all the input service flows can share the output link bandwidth of the switching node in a predetermined manner, even if different services are used. Different users and different priorities have agreed opportunities to get the corresponding services and use the public network. But when the network When it is loose, I hope that the users who need it can make full use of the network resources and maximize the benefits. At this time, I hope to give the users who are currently using the maximum traffic and enjoy better services, and the general implementation will be There is shaping to limit the traffic of each queue.

Summary of the invention

In order to solve the existing technical problems, the embodiments of the present invention mainly provide a method, a device, and a computer storage medium for scheduling user queues.

The technical solution of the embodiment of the present invention is implemented as follows:

An embodiment of the present invention provides a method for scheduling a user queue, where the method includes:

The first scheduler is connected to the second scheduler and the third scheduler, and the user queue is simultaneously connected to the second scheduler and the third scheduler;

The first scheduler determines whether the second scheduler allows scheduling, and when the second scheduler allows scheduling, scheduling the second scheduler, so that the second scheduler schedules the connected user queue through the corresponding shaper, in the second After the user queues connected to the scheduler are all scheduled, the first scheduler determines that the third scheduler allows scheduling, and schedules the third scheduler, so that the third scheduler schedules the hooks through the corresponding shaper. User queue

When the second scheduler does not allow scheduling, the first scheduler determines that the third scheduler allows scheduling, and schedules the third scheduler, so that the third scheduler schedules the attached user queue through the corresponding shaper.

In the above solution, the first scheduler is connected to the second scheduler and the third scheduler, and the user queue is simultaneously connected to the second scheduler and the third scheduler, including:

Configuring the first scheduler to schedule the scheduling algorithm of the second scheduler and the third scheduler to be strict priority (SP) scheduling, where scheduling the second scheduler has a higher priority than scheduling the third scheduler;

The scheduling algorithm for configuring the second scheduler and the third scheduler to schedule the user queue is SP, or deficit weighted polling (DWRR), or DWRR+SP scheduling.

In the above solution, the first scheduler is connected to the second scheduler and the third scheduler, and the user queue is simultaneously connected to the second scheduler and the third scheduler, and further includes:

And configuring an shaping leaky bucket filling rate and a leaking bucket capacity of the shaper corresponding to the first scheduler, the second scheduler, and the third scheduler, wherein the fill rate size is an allowable average traffic, and the leaky bucket capacity is configured. For the allowed burst traffic.

In the above solution, the first scheduler is connected to the second scheduler and the third scheduler, and the user queue is simultaneously connected to the second scheduler and the third scheduler, and further includes: setting corresponding to each user queue A deficit counter, the shaper determines a queue of users allowed to send packets based on the deficit counter.

In the above solution, the second scheduler is configured to provide a commitment rate to each user queue, and the third scheduler is configured to provide an additional rate to each user queue.

In the foregoing solution, the determining, by the first scheduler, whether the second scheduler allows scheduling includes: the first scheduler determines, according to whether there is a user queue that needs to send a data packet and the current configuration rate is lower than a committed rate of the second scheduler. Whether the scheduler allows scheduling.

In the above solution, the first scheduler determines that the third scheduler allows the scheduling to include: the first scheduler sends the data packet according to the need and the current configuration rate is lower than the third scheduler's additional rate and the second scheduler's committed rate. The sum of the user queues determines that the third scheduler allows scheduling.

The embodiment of the invention further provides a device for scheduling a user queue, the device comprising: a first scheduler, a second scheduler, a third scheduler and one or more shapers; wherein

The first scheduler is configured to determine whether the second scheduler allows scheduling, and when the second scheduler allows scheduling, scheduling the second scheduler, after the user queues connected under the second scheduler are all scheduled, determining The third scheduler allows scheduling and scheduling the third scheduler; when the second scheduler does not allow scheduling, determining that the third scheduler allows scheduling, and scheduling the third scheduler;

The second scheduler is configured to schedule the connected user queue by using a corresponding shaper, and notify the first scheduler after the connected user queues are all scheduled;

The third scheduler is configured to schedule a connected user queue by using a corresponding shaper;

A shaper that determines the queue of users that send packets.

In the above solution, the scheduling algorithm of the first scheduler scheduling the second scheduler and the third scheduler is an SP schedule, where scheduling the second scheduler has a higher priority than scheduling the third scheduler The scheduling algorithm of the second scheduler and the third scheduler scheduling user queue is SP, or DWRR, or DWRR+SP scheduling.

The computer storage medium provided by the embodiment of the present invention stores a computer program for executing the method for scheduling the user queue.

The embodiment of the invention provides a method and a device for scheduling a user queue. The first scheduler is connected to the second scheduler and the third scheduler, and the user queue is simultaneously connected to the second scheduler and the third scheduler. A scheduler determines whether the second scheduler allows scheduling, and when the second scheduler allows scheduling, scheduling a second scheduler, the second scheduler scheduling the attached user queue by using a corresponding shaper, in the second scheduler After the downlinked user queues are all scheduled, the first scheduler determines that the third scheduler allows scheduling, and schedules a third scheduler, and the third scheduler schedules the connected user queues by using the corresponding shaper; When the second scheduler does not allow scheduling, the first scheduler determines that the third scheduler allows scheduling, and schedules a third scheduler, and the third scheduler schedules the connected user queue through the corresponding shaper; thus, the network load When it is relatively small (for example, the device processing capacity is 1G, but the input traffic is less than 1G), the maximum bandwidth occupied by each user queue is limited by scheduling and shaping each user queue. When the load is large (such as device processing capabilities are 1G, but the input flow is greater than 1G), by scheduling to ensure minimum bandwidth for each queue is occupied by the user.

DRAWINGS

1 is a schematic diagram of a single bucket scheduling in the prior art;

2 is a schematic diagram of dual bucket scheduling adopted by an embodiment of the present invention;

3 is a schematic flowchart of a method for implementing user queue scheduling according to an embodiment of the present invention;

FIG. 4 is a schematic structural diagram of an apparatus for implementing user queue scheduling according to an embodiment of the present invention.

detailed description

Single-bucket scheduling is to configure only one rate for a user queue. The part of the user queue that is larger than the configured rate will not be scheduled. If the configured rate is smaller than the configured rate, it can be scheduled to implement the corresponding QOS. Dual-bucket scheduling configures two rates for the user queue, namely, CIR (Committed Information Rate) and Extra Rate (EIR). The CIR is the rate that must be guaranteed regardless of whether the current network is congested. The EIR is the extra rate that is allowed to be scheduled when the current network is not congested.

The existing single-bucket scheduling is shown in Figure 1. The user queue is attached to the PQ scheduler, and the PQ scheduler is attached to the time slot (T-CONT) scheduler. There are two PQ schedulers in the figure, which can be called scheduler A and scheduler B. Eight user queues are attached to one PQ scheduler. Scheduler A and Scheduler B are attached to the T-CONT scheduler. A queue is only attached to a scheduler and is limited by only one rate. This is called single-bucket scheduling. The scheduling rate of the user queue 0 under the scheduler A is only one. If the rate of the user queue 0 is small and does not reach the limited rate, the scheduling can be continued. Once the limited speed is reached, the scheduling is limited and the scheduling cannot be continued. Even if the entire network is not congested and other user queues do not have a packet request, the user queue 0 is strictly limited and cannot exceed the limit rate.

The dual bucket scheduling adopted by the embodiment of the present invention is shown in FIG. 2. Hereinafter, the CIR scheduler is referred to as a C scheduler, and the EIR scheduler is an E scheduler. Each user queue is not only attached to the C scheduler, but also to the E scheduler. A user queue is limited by two rates: committed rate and extra rate. First, taking user queue 0 as an example, user queue 0 must get the committed rate. If the network is congested and other user queues also require a large amount of traffic, user queue 0 uses the committed rate. If the network is not congested and other user queues have less traffic demand, user queue 0 can get extra rate and enjoy faster. Rate, get better extra services. It is very flexible without affecting other users and getting better service.

The scheduling model in FIG. 2 uses a DWRR scheduling algorithm and shaping to implement QOS, wherein the funnel pattern represents a shaper.

In the embodiment of the present invention, the first scheduler is connected to the second scheduler and the third scheduler, and the user queue is simultaneously connected to the second scheduler and the third scheduler, and the first scheduler determines whether the second scheduler allows Scheduling, when the second scheduler allows scheduling, scheduling the second scheduler, the second scheduler scheduling the connected user queue by using the corresponding shaper, after the user queues connected under the second scheduler are all scheduled The first scheduler determines that the third scheduler allows scheduling, and schedules a third scheduler, where the third scheduler schedules the attached user queue by using a corresponding shaper; when the second scheduler does not allow scheduling, the first scheduler The scheduler determines that the third scheduler allows scheduling, and schedules a third scheduler, which schedules the attached user queue through the corresponding shaper. Here, the first scheduler may be a T-CONT scheduler, the second scheduler may be a CIR scheduler, and the third scheduler may be an EIR scheduler.

The invention will be further described in detail below with reference to the drawings and specific embodiments.

The embodiment of the invention implements a method for scheduling a user queue. As shown in FIG. 3, the method includes the following steps:

Step 301: The first scheduler is connected to the second scheduler and the third scheduler, and the user queue is simultaneously connected to the second scheduler and the third scheduler.

Here, before the step 301 is performed, the scheduling manners of the first scheduler, the second scheduler, and the third scheduler are first configured, specifically: configuring the first scheduler to schedule the second scheduler and the third scheduler to be strict. Priority (SP, Strict-Priority) scheduling; wherein scheduling the second scheduler has a higher priority than scheduling the third scheduler, thereby ensuring that the second scheduling is scheduled first And scheduling the third scheduler; configuring the second scheduler and the third scheduler to schedule the user queues as SP, or DWRR, or DWRR+SP.

Before performing step 301, the shaping bucket filling rate and the leaking bucket capacity of the shaper corresponding to the first scheduler, the second scheduler, and the third scheduler are further configured; wherein the fill rate size indicates the allowed average traffic, The leaky bucket capacity size indicates the allowed burst traffic. The shaper's leaky bucket fill rate and leaky bucket capacity can be configured according to the number of bytes represented by each shaped token and the rate that needs to be limited. In this embodiment, the leaky bucket adopts a fixed-period filling technique, and adds a fixed number of shaping tokens to the leaky bucket according to the interval period. The interval period and the number of bytes added each time the shaping token is added may be configured by the user. Different fill rates can be achieved by configuring different interval periods and padding bytes. The formula is as follows:

Fill rate = (clock frequency / interval period) * padding bytes.

The step further includes: setting a corresponding deficit counter for each user queue, and the shaper determines a user queue that allows the data packet to be sent according to the deficit counter, and the deficit counter is initialized to a fixed value. Specifically, when an active user queue is scheduled, if the corresponding deficit counter is less than 0, the service quantum of the user queue and the deficit counter are added, and the accumulated result is stored in the deficit counter, and the current scheduling is skipped; If the corresponding deficit counter is greater than 0, it is determined whether the user queue satisfies the allowable outbound condition, and when satisfied, the user queue is allowed to send a data packet, the deficit counter minus the packet length of the data packet, and the corresponding shaping token is also To reduce the amount. The allowed outbound condition includes: a, the user queue is in an active state, that is, there is a packet to be scheduled and the shaping condition is satisfied; b, the user queue has sufficient credit.

Step 302: The first scheduler determines whether the second scheduler allows scheduling, when the second scheduler allows scheduling, step 303 is performed, and when the second scheduler does not allow scheduling, step 304 is performed;

Specifically, the first scheduler determines, according to whether there is a user queue that needs to send a data packet and the current configuration rate is lower than the committed rate of the second scheduler, whether the second scheduler allows scheduling, and if yes, determines that the second scheduler allows scheduling. If not, it is determined that the second scheduler does not allow scheduling.

Step 303: scheduling the second scheduler, the second scheduler scheduling the connected user queue through the corresponding shaper, after the second queue is connected to the user queue is scheduled, step 302 is performed;

Step 304: The first scheduler determines whether the third scheduler allows scheduling. When allowed, step 305 is performed. When not allowed, the process ends.

Specifically, the first scheduler determines, according to whether there is a user queue that needs to send a data packet and the current configuration rate is lower than a sum of the additional rate of the third scheduler and the committed rate of the second scheduler, whether the third scheduler allows scheduling, If so, it is determined that the third scheduler allows scheduling, and if not, it is determined that the third scheduler does not allow scheduling.

Step 305: The third scheduler is scheduled, and the third scheduler schedules the connected user queue by using a corresponding shaper. After the third scheduler schedules the process, step 302 is performed.

In order to implement the foregoing method, an embodiment of the present invention further provides a device for scheduling a user queue. As shown in FIG. 4, the device includes: a first scheduler 41, a second scheduler 42, a third scheduler 43, and one or more shapings. 44; wherein

The first scheduler 41 is connected to the second scheduler 42 and the third scheduler 43, and the user queue is simultaneously attached to the second scheduler 42 and the third scheduler 43;

The first scheduler 41 determines whether the second scheduler 42 allows scheduling, and when the second scheduler 42 allows scheduling, schedules the second scheduler 42 after the user queues attached to the second scheduler 42 are scheduled. Determining that the third scheduler 43 allows scheduling, and scheduling the third scheduler 43; when the second scheduler 42 does not allow scheduling, determining that the third scheduler 43 allows scheduling, and scheduling the third scheduler 43;

The second scheduler 42 schedules the connected user queue through the corresponding shaper 44, and after the hooked user queues are all scheduled, notify the first scheduler 41;

The third scheduler 43 schedules the connected user queue through the corresponding shaper 44;

A shaper 44 is configured to determine a user queue for transmitting data packets.

The first scheduler 41 is a T-CONT scheduler, the second scheduler 42 is a CIR scheduler, and is configured to provide a commitment rate to each user queue, and the third scheduler 43 is an EIR scheduler. Used to provide additional rates to individual user queues.

The first scheduler 41 schedules the scheduling algorithms of the second scheduler 42 and the third scheduler 43 to be SP scheduling, wherein the scheduling second scheduler 42 has a higher priority than the scheduling third scheduler 43. Priority, in order to ensure that the second scheduler 42 is scheduled first, and then the third scheduler 43 is scheduled; the scheduling algorithm of the second scheduler and the third scheduler scheduling the user queue is SP, or DWRR, or DWRR+ SP.

Here, the shaper 44 is specifically configured to determine a user queue that is allowed to send a data packet according to a deficit counter corresponding to each user queue.

Preferably, the first scheduler 41 determines whether the second scheduler 42 allows scheduling according to whether there is a user queue that needs to send a data packet and the current configuration rate is lower than the committed rate of the second scheduler 42.

Preferably, the first scheduler 41 determines the third scheduler according to a user queue that needs to send a data packet and the current configuration rate is lower than the sum of the additional rate of the third scheduler 43 and the committed rate of the second scheduler 42. 43 allows scheduling.

The first scheduler, the second scheduler, the third scheduler, and the one or more shapers may be configured by a central processing unit (CPU), or a digital signal processor (DSP), or a programmable gate. Array (FPGA, Field-Programmable Gate Array) implementation.

In summary, the embodiment of the present invention uses dual-bucket scheduling to limit the maximum bandwidth occupied by each user queue by scheduling and shaping each user queue when the network load is relatively small, and when the network load is large, Through scheduling, the minimum bandwidth occupied by each user queue can be guaranteed.

In the several embodiments provided by the present application, it should be understood that the disclosed apparatus and method may be implemented in other manners. The device embodiments described above are merely illustrative For example, the division of the unit is only a logical function division, and the actual implementation may have another division manner, such as: multiple units or components may be combined, or may be integrated into another system, or some features may be Ignore, or not execute. In addition, the coupling, or direct coupling, or communication connection of the components shown or discussed may be indirect coupling or communication connection through some interfaces, devices or units, and may be electrical, mechanical or other forms. of.

The units described above as separate components may or may not be physically separated, and the components displayed as the unit may or may not be physical units, that is, may be located in one place or distributed to multiple network units; Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of the embodiment.

In addition, each functional unit in each embodiment of the present invention may be integrated into one processing module, or each unit may be separately used as one unit, or two or more units may be integrated into one unit; the above integration The unit can be implemented in the form of hardware or in the form of hardware plus software functional units.

A person skilled in the art can understand that all or part of the steps of implementing the above method embodiments may be completed by using hardware related to the program instructions. The foregoing program may be stored in a computer readable storage medium, and the program is executed when executed. The foregoing storage device includes the following steps: the foregoing storage medium includes: a mobile storage device, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk. A medium that can store program code.

Correspondingly, an embodiment of the present invention further provides a computer storage medium, wherein a computer program is stored, and the computer program is used to execute a method for user queue scheduling according to an embodiment of the present invention.

The above is only a specific embodiment of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art can easily think of changes or substitutions within the technical scope of the present invention. It should be covered by the scope of the present invention. Therefore, this The scope of the invention should be determined by the scope of the appended claims.

Industrial applicability

The invention discloses a method and a device for scheduling a user queue and a computer storage medium. The first scheduler can be connected to the second scheduler and the third scheduler, and the user queue is simultaneously connected to the second scheduler and the third scheduler. First, the second scheduler is scheduled to schedule the connected user queue through the corresponding shaper, and the third schedule is scheduled after the user queues attached to the second scheduler are all scheduled or when the second scheduler does not allow scheduling. And causing the third scheduler to schedule the connected user queue by using a corresponding shaper. As such, with this dual bucket scheduling, a user queue is limited by two rates. When the network load is relatively small, the maximum bandwidth occupied by each user queue is limited by scheduling and shaping the user queues. When the network load is large, the minimum bandwidth occupied by each user queue can be ensured by scheduling.

Claims

A method for user queue scheduling, the method comprising:

The first scheduler is connected to the second scheduler and the third scheduler, and the user queue is simultaneously connected to the second scheduler and the third scheduler;

The first scheduler determines whether the second scheduler allows scheduling, and when the second scheduler allows scheduling, scheduling the second scheduler, so that the second scheduler schedules the connected user queue through the corresponding shaper, in the second After the user queues connected to the scheduler are all scheduled, the first scheduler determines that the third scheduler allows scheduling, and schedules the third scheduler, so that the third scheduler schedules the hooks through the corresponding shaper. User queue

When the second scheduler does not allow scheduling, the first scheduler determines that the third scheduler allows scheduling, and schedules the third scheduler, so that the third scheduler schedules the attached user queue through the corresponding shaper.
The method of claim 1, wherein the first scheduler is connected to the second scheduler and the third scheduler, and the user queue is simultaneously attached to the second scheduler and the third scheduler, including:

The scheduling algorithm for configuring the first scheduler to schedule the second scheduler and the third scheduler is a priority SP scheduling, where scheduling the second scheduler has a higher priority than scheduling the third scheduler;

The scheduling algorithm for configuring the second scheduler and the third scheduler to schedule the user queue is SP, or a deficit weighted polling DWRR, or a DWRR+SP scheduling.
The method of claim 2, wherein the first scheduler is connected to the second scheduler and the third scheduler, and the user queue is simultaneously attached to the second scheduler and the third scheduler, and further includes:

And configuring an shaping leaky bucket filling rate and a leaking bucket capacity of the shaper corresponding to the first scheduler, the second scheduler, and the third scheduler, wherein the fill rate size is an allowable average traffic, and the leaky bucket capacity is configured. For the allowed burst traffic.
The method of claim 3, wherein the first scheduler hooks up the second schedule And the third scheduler, the user queue is simultaneously connected to the second scheduler and the third scheduler, and further includes: setting a corresponding deficit counter for each user queue, and the shaper determines, according to the deficit counter, that the data packet is allowed to be sent. User queue.
The method of claim 1 wherein said second scheduler is operative to provide a committed rate to each user queue, said third scheduler for providing an additional rate to each user queue.
The method according to claim 5, wherein the determining, by the first scheduler, whether the second scheduler allows scheduling comprises: the first scheduler according to whether there is a need to send a data packet and the current configuration rate is lower than the commitment of the second scheduler The rate user queue determines if the second scheduler allows scheduling.
The method of claim 5, wherein the first scheduler determines that the third scheduler allows scheduling comprises: the first scheduler transmitting a data packet according to need and the current configuration rate is lower than an additional rate of the third scheduler The user queue of the sum of the committed rates of the second scheduler determines that the third scheduler allows scheduling.
An apparatus for scheduling a user queue, the apparatus comprising: a first scheduler, a second scheduler, a third scheduler, and one or more shapers; wherein

The first scheduler is connected to the second scheduler and the third scheduler, and the user queue is simultaneously connected to the second scheduler and the third scheduler;

The first scheduler is configured to determine whether the second scheduler allows scheduling, and when the second scheduler allows scheduling, scheduling the second scheduler, after the user queues connected under the second scheduler are all scheduled, determining The third scheduler allows scheduling and scheduling the third scheduler; when the second scheduler does not allow scheduling, determining that the third scheduler allows scheduling, and scheduling the third scheduler;

The second scheduler is configured to schedule the connected user queue by using a corresponding shaper, and notify the first scheduler after the connected user queues are all scheduled;

The third scheduler is configured to schedule the connected user queue by using a corresponding shaper;

A shaper that is configured to determine the user queue that sent the packet.
The apparatus of claim 8, wherein the second scheduler is configured to provide a committed rate to each user queue; the third scheduler is configured to provide an additional rate to each user queue.
The apparatus according to claim 8, wherein the scheduling algorithm of the first scheduler scheduling the second scheduler and the third scheduler is an SP schedule, wherein scheduling the second scheduler has a higher priority The priority of the third scheduler is scheduled; the scheduling algorithm of the second scheduler and the third scheduler scheduling the user queue is SP, or DWRR, or DWRR+SP scheduling.
A computer storage medium having stored therein computer executable instructions for performing the method of user queue scheduling of any one of claims 1 to 7.