WO2019109902A1

WO2019109902A1 - Queue scheduling method and apparatus, communication device, and storage medium

Info

Publication number: WO2019109902A1
Application number: PCT/CN2018/119110
Authority: WO
Inventors: 杜文华; 吕晖; 郑勤; 王馨宇
Original assignee: 华为技术有限公司
Priority date: 2017-12-08
Filing date: 2018-12-04
Publication date: 2019-06-13
Also published as: CN109905331A; CN109905331B

Abstract

The present application relates to the technical field of data transmission. Disclosed are a queue scheduling method and apparatus, a communication device, and a storage medium. The method comprises: when a length difference of a first queue changes, determining a target probability, based on the length difference of the first queue and a queue length of the first queue; randomly generating a scheduling request based on the target probability, so that a proportion of a possibility of generating the scheduling request to a possibility of not generating the scheduling request is the target probability, the scheduling request being used for indicating a scheduling state for scheduling a first queue; and when the scheduling request is generated, sending the scheduling request to a queue scheduler, the first queue being any one of at least one queue managed by the scheduler, the length difference being a difference between a current queue credit value of the first queue and the queue length of the first queue, and the current queue credit value being a data amount quota of the first queue, which can be used for transmitting data. By means of the present application, the probability of great scheduling delay in the scheduling of a queue by a scheduler is reduced.

Description

Queue scheduling method and device, communication device, storage medium

The present application claims priority to Chinese Patent Application No. PCT Application No. No. No. No. No. No. No. No. No. No. in.

Technical field

The present invention relates to the field of data transmission technologies, and in particular, to a queue scheduling method and apparatus, a communication device, and a storage medium.

Background technique

Quality of Service (QoS) is an indicator that reflects the state of the network. It can reflect the network status such as network delay and congestion. Queue scheduling technology is the key technology to guarantee QoS. Among them, in the queue scheduling technology, a communication device responsible for queue scheduling such as a router or a switch needs to schedule a data packet according to a certain rule, so that each queue in which the data packet is cached can obtain a fair scheduling opportunity, so that each queue is in a queue. Packets can be sent out.

In the related art, a Queue Manager (QM) and a Queue SCheduler (QSC) are disposed in the communication device, and for each queue managed by the queue manager, the queue manager can use the data according to the queue. The difference between the transmitted data amount (ie, the queue credit) and the queue length of the queue, sends a scheduling request to the queue scheduler, so that the queue scheduler allocates the queue for data transmission. The amount of data to achieve the transfer of packets in the queue.

However, when the queue scheduling is performed by this method, when the difference corresponding to any queue managed by the queue manager changes, the queue manager sends a scheduling request to the queue scheduler, so that when the queue manager manages When a multi-queue is used, the queue manager sends a scheduling request with a high frequency. The queue scheduler has a large load and cannot process the scheduling request in a timely manner. As a result, the scheduling delay of the scheduling of the queue is large.

Summary of the invention

The application provides a queue scheduling method, device, system and computer readable storage medium, which can solve the problem that when the queue manager manages more queues, the queue manager sends a scheduling request with a higher frequency, and the queue scheduler has a larger load. The problem that the scheduling request cannot be processed in time may result in a large scheduling delay for the scheduling of the queue. The technical solution provided by the present application is as follows:

In a first aspect, an exemplary embodiment of the present application provides a queue scheduling method, where the method is applied to a queue manager, and the method includes:

Determining a target probability based on a length difference of the first queue and a queue length of the first queue, when a difference in length of the first queue changes;

And performing, according to the target probability, a random generation of the scheduling request, such that a ratio of the possibility of generating the scheduling request to an ungenerated possibility is the target probability, where the scheduling request is used to indicate that the first queue is performed Scheduled scheduling status;

When the scheduling request is generated, the scheduling request is sent to a queue scheduler;

The first queue is any one of at least one queue managed by the queue manager, and the length difference is a current queue credit value of the first queue and a queue length of the first queue. The difference between the current queue credits is the amount of data that the first queue is currently available for data transmission.

It should be noted that, by determining the target probability when the length difference of the first queue changes, the random generation of the scheduling request is performed based on the target probability, so that the ratio of the probability of generating the scheduling request to the probability of not generating is the target probability. And when the scheduling request is generated, the scheduling request is sent to the queue scheduler, which can effectively reduce the frequency at which the queue manager sends the scheduling request to the queue scheduler, thereby reducing the scheduling load of the queue scheduler. To reduce the probability that the queue scheduler has a large scheduling delay for the scheduling of the queue.

When the length difference of the first queue is changed by the queue length of the first queue, the determining the target probability based on the length difference of the first queue and the queue length of the first queue may include:

And acquiring, when the queue length of the first queue changes, a first difference after the change in the length difference, and a second difference before the change;

The target probability is determined based on at least one of the first difference and the second difference, and a queue length of the first queue.

As an implementation manner, the determining, according to at least one of the first difference value and the second difference value, and a queue length of the first queue, the target probability includes:

Comparing the first difference with a reference credit threshold and a scheduling granularity, where the reference credit threshold is smaller than the scheduling granularity, where the scheduling granularity is that when the queue scheduler schedules the first queue, The nominal amount of data allocated for data transmission allocated by the first queue;

When the first difference is greater than the reference credit threshold and less than the scheduling granularity, the target probability is determined based on the first difference and the scheduling granularity.

When the first difference is greater than the reference credit threshold and less than the scheduling granularity, it may be determined that the first queue still has a certain queue credit, but the remaining queue credit is not enough for the data packet in the first queue to be dequeued. In order to ensure the effective dequeue of the data packets in the first queue, the first queue may request the queue scheduler to schedule them, but in order to prevent the queue manager from managing multiple queues to send scheduling requests to the queue scheduler, At this time, the target probability of generating the scheduling request may be determined, so that the scheduling request is randomly generated according to the target probability, so as to reduce the frequency at which the queue manager sends the scheduling request to the queue scheduler, and reduce the scheduling load of the queue scheduler, and further Ensure the effective dequeue of the data packet.

Optionally, the determining the target probability based on the first difference value and the scheduling granularity, including:

Determining, by the probability that the target difference is greater than the first random number, the target probability, where the target difference is a difference between the scheduling granularity and the first difference, the first random number is randomly generated and not A positive number greater than the scheduling granularity.

In another implementation manner, the determining the target probability based on at least one of the first difference value and the second difference value, and a queue length of the first queue, includes:

Comparing the second difference with a reference credit threshold and a scheduling granularity, where the reference credit threshold is smaller than the scheduling granularity, where the scheduling granularity is when the queue scheduler schedules the first queue, The nominal amount of data allocated for data transmission allocated by the first queue;

Determining the target probability based on the second difference and a target packet length when the second difference is greater than the reference credit threshold and less than the scheduling granularity, where the target packet length is the first queue The length of the packet whose length has changed.

When the second difference is greater than the reference credit threshold and less than the scheduling granularity, it may be determined that the first queue still has a certain queue credit, but the remaining queue credit is not enough for the data packet in the first queue to be dequeued. In order to ensure the effective dequeue of the data packets in the first queue, and to prevent the queue manager from managing the multiple queues to send the scheduling request to the queue scheduler, the target probability of generating the scheduling request may be determined at this time, so as to facilitate A scheduling request is randomly generated according to the target probability, thereby ensuring effective dequeue of the data packet.

Optionally, the determining the target probability based on the second difference and the target packet length, including:

The probability that the target packet length is greater than the second random number is determined as the target probability, and the second random number is a positive number that is randomly generated and not greater than the second difference.

In this case, when the target probability is determined based on the second difference and the target packet length, the determined target probability exhibits a more uniform change, which can further reduce the frequency at which the queue manager sends the scheduling request to the queue scheduler, thereby reducing The scheduling load of the small queue scheduler can further reduce the probability that the queue scheduler has a large scheduling delay for scheduling of the queue.

Comparing the first difference with the reference credit threshold and the scheduling granularity, where the reference credit threshold is smaller than the scheduling granularity, and the scheduling granularity is that when the queue scheduler schedules the first queue, The nominal amount of data allocated for data transmission allocated by the first queue;

Comparing the second difference with the reference credit threshold and the scheduling granularity, respectively;

And determining, when the first difference value and the second difference value are greater than the reference credit threshold and less than the scheduling granularity, determining the target probability based on the second difference value and a target packet length, The target packet length is the length of the data packet that changes the queue length of the first queue.

It should be noted that, in this case, the target probability is determined based on the second difference and the target packet length, and the queue manager can be further reduced to the queue according to the processing method for determining the target probability based on the first difference and the scheduling granularity. The frequency at which the scheduler sends the scheduling request, thereby reducing the scheduling load of the queue scheduler, can further reduce the probability that the queue scheduler has a large scheduling delay for scheduling the queue.

A ratio of the target packet length to the second difference is determined as the target probability.

Comparing the first difference with a reference credit threshold and a preset first credit threshold, where the first credit threshold is less than the reference credit threshold;

When the first difference is greater than the first credit threshold and less than the reference credit threshold, determining that the target probability is 1.

When the first difference is greater than the first credit threshold and less than the reference credit threshold, the queue credit is deficit, but the deficit is not serious. At this time, it can be determined that the first queue does need the queue scheduler to schedule it. And in order to ensure the normal dequeue of the data packets in the first queue, the first queue needs the queue scheduler to schedule it at a relatively small rate. Therefore, when the first difference is greater than the first credit threshold and less than the reference credit threshold, the target probability may be determined to be one.

Comparing the first difference with a preset first credit threshold;

When the first difference is less than the first credit threshold, the target probability is determined to be 1.

When the first difference is less than the first credit threshold, the queue credit of the first queue has a severe deficit, and the first queue urgently needs the queue scheduler to schedule it. For example, the first queue needs the queue scheduler to have a relatively large rate pair. It performs scheduling to ensure the normal dequeue of data packets in the first queue.

When the length difference of the first queue is changed by the current queue credit value of the first queue, the target probability is determined based on the length difference of the first queue and the queue length of the first queue, including:

When the current queue credit value of the first queue changes, the target probability is determined according to the length difference.

As an implementation manner, the determining the target probability according to the length difference includes:

Comparing the length difference with the reference credit threshold and the scheduling granularity, where the reference credit threshold is smaller than the scheduling granularity, and the scheduling granularity is when the queue scheduler schedules the first queue, The nominal amount of data allocated for data transmission by the first queue;

When the length difference is greater than the reference credit threshold and less than the scheduling granularity, the target probability is determined based on the length difference and the scheduling granularity.

In the first case, when the length difference is greater than the reference credit threshold and less than the scheduling granularity, it may be determined that the first queue still has a certain queue credit, but the remaining queue credit is not enough for the first queue. The data packet is dequeued. At this time, the target probability of generating the scheduling request may be determined to randomly generate the scheduling request, reduce the frequency at which the queue manager sends the scheduling request to the queue scheduler, and reduce the scheduling load of the queue scheduler, and further Ensure the effective dequeue of the data packet.

Optionally, the determining the target probability based on the length difference and the scheduling granularity, including:

Determining, by the probability that the length difference is smaller than the first random number, is the target probability, and the first random number is a positive number that is randomly generated and not greater than the scheduling granularity.

As another implementation manner, the determining the target probability according to the length difference includes:

Comparing the length difference with a reference credit threshold and a preset first credit threshold, where the first credit threshold is less than the reference credit threshold;

When the length difference is greater than the first credit threshold and less than the reference credit threshold, determining that the target probability is 1.

When the length difference is greater than the first credit threshold and less than the reference credit threshold, the queue credit is deficit, but the deficit is not serious. At this time, it can be determined that the first queue does need the queue scheduler to schedule it, and In order to ensure the normal dequeue of data packets in the first queue, the first queue needs the queue scheduler to schedule it at a relatively small rate. Therefore, when the length difference is greater than the first credit threshold and less than the reference credit threshold, the target probability may be determined to be one.

In another implementation manner, the determining the target probability according to the length difference includes:

Comparing the length difference with a preset first credit threshold;

When the length difference is less than the first credit threshold, the target probability is determined to be 1.

When the length difference is less than the first credit threshold, the queue credit of the first queue has a severe deficit, and the first queue urgently needs the queue scheduler to schedule it.

Correspondingly, the performing random generation of the scheduling request based on the target probability includes:

When the scheduling request is generated based on the target probability, the scheduling request is generated based on the selection result, where the scheduling request includes a first status identifier, where the first status identifier indicates that the queue scheduler performs queue scheduling The first rate, the first rate is less than a preset scheduling rate threshold.

Or performing random generation of the scheduling request based on the target probability, including:

When the scheduling request is generated based on the target probability, the scheduling request is generated based on the selection result, where the scheduling request includes a second status identifier, where the second status identifier indicates that the queue scheduler performs queue scheduling The second rate is greater than a preset scheduling rate threshold.

It should be noted that when the random generation of the scheduling request is performed based on the target probability, since the ratio of the possibility of generating the scheduling request to the probability of not generating is the target probability, the queue manager is reduced based on the plurality of queues managed by the queue manager. The number of scheduling requests is such that the number of scheduling requests generated by the queue manager is as uniformly distributed as possible, which reduces the frequency at which the queue manager sends scheduling requests to the queue scheduler, and can reduce the queue. The scheduling load of the scheduler is such that the update rate of the scheduling state does not exceed the scheduling capability of the queue scheduler, thereby reducing the probability that the queue scheduler has a large scheduling delay for scheduling of the queue. In addition, the number of scheduling requests generated by the queue manager can be uniformly distributed as shown in the queue manager, so that the current queue credits of all the queues are evenly distributed within the preset range corresponding to the scheduling granularity. The sum of the credits is close to the total scheduling bandwidth that the queue scheduler can schedule, so that the actual total scheduling bandwidth is more accurate, and the scheduling bandwidth caused by the high frequency of sending the scheduling request and the scheduling bandwidth between the queues are not solved. Accurate and other issues.

Further, after the scheduling request is sent to the queue scheduler, the method further includes:

Receiving a scheduling response sent by the queue scheduler, where the scheduling response includes a scheduling granularity, where the scheduling granularity is a rated data amount allocated by the queue scheduler for the first queue for data transmission;

Updating the current queue credit value of the first queue based on the scheduling response.

After receiving the scheduling request sent by the queue manager, the queue scheduler may run a scheduling algorithm according to the scheduling state indicated by the scheduling request, allocate a rated data amount for data transmission to the first queue, and send the same to the queue manager. A scheduling response of the amount of data, so that the queue manager performs a queue credit update according to it, thereby performing a dequeuing operation on the data packets in the first queue.

Optionally, after the updating the current queue credit value of the first queue based on the scheduling response, the method further includes:

Dequeuing the data packet of the corresponding length in the first queue based on the updated current queue credit value.

After updating the current queue credit value, if the current queue credit value satisfies the dequeue data volume of the data packet in the first queue, the dequeuing operation may be performed on the data packet of the corresponding length in the first queue to implement data transmission.

Further, the sending the scheduling request to the queue scheduler includes:

Determining whether the scheduling state indicated by the scheduling request is the same as the current scheduling state corresponding to the first queue;

And when the scheduling state indicated by the scheduling request is different from the current scheduling state corresponding to the first queue, the scheduling request is sent to the queue scheduler.

After the scheduling request is generated, the queue manager may compare the current scheduling state corresponding to the first queue with the scheduling state indicated by the scheduling request, when the scheduling state indicated by the scheduling request is the same as the current scheduling state corresponding to the first queue. The queue scheduler can continue to schedule the first queue according to the current scheduling state. At this time, it is not necessary to send a scheduling request to the queue scheduler, which can save system resources; when the scheduling status indicated by the scheduling request is related to the current queue When the scheduling status is different, the scheduling request is sent to the queue scheduler to update the scheduling state of the queue queue to the first queue.

Optionally, before the determining the target probability based on the difference between the length of the first queue and the queue length of the first queue, when the length difference of the first queue is changed, the method further includes:

When the queue length of the first queue changes, determining that the length difference of the first queue changes;

And/or, when the current queue credit value of the first queue changes, determining that the length difference of the first queue changes.

In a second aspect, an exemplary embodiment of the present application further provides a queue scheduling apparatus, where the apparatus includes:

a first determining module, configured to determine a target probability based on a length difference of the first queue and a queue length of the first queue when a difference in length of the first queue changes;

a generating module, configured to perform random generation of the scheduling request based on the target probability, such that a ratio of a probability of generating the scheduling request to an ungenerated possibility is the target probability, and the scheduling request is used to indicate a target Decoding the scheduling state of the first queue;

a sending module, configured to send the scheduling request to a queue scheduler when the scheduling request is generated;

Optionally, the first determining module includes:

Obtaining a sub-module, configured to: when the queue length of the first queue changes, obtain a first difference after the length difference is changed, and a second difference before the change;

a first determining submodule, configured to determine the target probability based on at least one of the first difference value and the second difference value, and a queue length of the first queue.

As an implementation manner, the first determining submodule includes:

a first comparing unit, configured to compare the first difference with a reference credit threshold and a scheduling granularity, where the reference credit threshold is smaller than the scheduling granularity, and the scheduling granularity is that the queue scheduler pairs the first The rated data amount allocated for the first queue for data transmission when the queue is scheduled;

a first determining unit, configured to determine the target probability based on the first difference value and the scheduling granularity when the first difference value is greater than the reference credit threshold and less than the scheduling granularity.

The first determining unit is configured to:

As another implementation manner, the first determining submodule includes:

a first comparing unit, configured to compare the second difference with a reference credit threshold and a scheduling granularity, where the reference credit threshold is smaller than the scheduling granularity, and the scheduling granularity is that the queue scheduler pairs the first The rated data amount allocated for the first queue for data transmission when the queue is scheduled;

a first determining unit, configured to determine, according to the second difference and a target packet length, the target probability, when the second difference is greater than the reference credit threshold and less than the scheduling granularity, the target packet length The length of the data packet for which the queue length of the first queue changes.

The first determining unit is configured to:

In another implementation manner, the first determining submodule includes:

a first comparing unit, configured to compare the first difference value with a reference credit threshold and a scheduling granularity, where the reference credit threshold is smaller than the scheduling granularity, and the scheduling granularity is that the queue scheduler pairs the first The rated data amount allocated for the first queue for data transmission when the queue is scheduled;

The first comparing unit is configured to compare the second difference value with the reference credit threshold and the scheduling granularity respectively;

a first determining unit, configured to determine, when the first difference value and the second difference value are greater than the reference credit threshold and less than the scheduling granularity, based on the second difference value and a target packet length The target probability, the target packet length is a length of a data packet that changes a queue length of the first queue.

The first determining unit is configured to:

In another implementation manner, the first determining submodule includes:

a first comparison unit, configured to compare the first difference value with a reference credit threshold and a preset first credit threshold, where the first credit threshold is smaller than the reference credit threshold;

The first determining unit is configured to determine that the target probability is 1 when the first difference is greater than the first credit threshold and less than the reference credit threshold.

In another implementation manner, the first determining submodule includes:

a first comparing unit, configured to compare the first difference with a preset first credit threshold;

a first determining unit, configured to determine that the target probability is 1 when the first difference is less than the first credit threshold.

Optionally, the first determining module includes:

a second determining submodule, configured to determine the target probability according to the length difference when a current queue credit value of the first queue changes.

As an implementation manner, the second determining submodule includes:

a second comparing unit, configured to compare the length difference value with a reference credit threshold and a scheduling granularity, where the reference credit threshold is smaller than the scheduling granularity, where the scheduling granularity is that the queue scheduler is to the first queue The amount of rated data allocated for the first queue for data transmission when scheduling is performed;

a second determining unit, configured to determine the target probability based on the length difference and the scheduling granularity when the length difference is greater than the reference credit threshold and less than the scheduling granularity.

The second determining unit is configured to:

As another implementation manner, the second determining submodule includes:

a second comparing unit, configured to compare the length difference value with a reference credit threshold and a preset first credit threshold, where the first credit threshold is smaller than the reference credit threshold;

And a second determining unit, configured to determine that the target probability is 1 when the length difference is greater than the first credit threshold and less than the reference credit threshold.

In another implementation manner, the second determining submodule includes:

a second comparing unit, configured to compare the length difference with a preset first credit threshold;

a second determining unit, configured to determine that the target probability is 1 when the length difference is less than the first credit threshold.

Optionally, the generating module is configured to:

Further, the device further includes:

a receiving module, configured to receive a scheduling response sent by the queue scheduler, where the scheduling response includes a scheduling granularity, where the scheduling granularity is a rated data volume allocated by the queue scheduler for the first queue for data transmission ;

And an update module, configured to update a current queue credit value of the first queue based on the scheduling response.

Optionally, the device further includes:

And an execution module, configured to perform a dequeuing operation on the data packet of the corresponding length in the first queue based on the updated current queue credit value.

The sending module is configured to:

Optionally, the device further includes:

a second determining module, configured to determine, when the queue length of the first queue changes, a change in a length difference of the first queue;

And/or, the second determining module is configured to determine that a difference in length of the first queue changes when a current queue credit value of the first queue changes.

In a third aspect, an exemplary embodiment of the present application further provides a communication device, where the communication device includes: a queue scheduler and a queue manager;

The queue manager includes the queue scheduling device of any of the second aspects;

The queue scheduler is configured to send a scheduling response to the queue manager after receiving the scheduling request sent by the queue manager, where the scheduling response is used to allocate rated data for data transmission based on the scheduling request. the amount.

In a fourth aspect, an exemplary embodiment of the present application further provides a queue scheduling apparatus, where the apparatus includes:

Processing component

a memory for storing executable instructions of the processing component;

Wherein the processing component is configured to:

In a fifth aspect, the exemplary embodiment of the present application further provides a storage medium, when the instructions in the storage medium are executed by a processing component of the terminal, enabling the terminal to perform the queue scheduling according to any one of the first aspects. method.

The technical solutions provided by the exemplary embodiments of the present application have the following beneficial effects:

The queue scheduling method and apparatus, the communication device, and the storage medium provided by the exemplary embodiments of the present application determine a target probability by changing a length difference of the first queue, and perform random generation of the scheduling request based on the target probability, so that the scheduling request is performed. The ratio of the generated possibility to the ungenerated possibility is the target probability, and when the scheduling request is generated, the scheduling request is sent to the queue scheduler, which reduces the plurality of queue managers based on the related management techniques. The number of scheduling requests generated by the queue is such that the number of scheduling requests generated by the queue manager is uniformly distributed as shown in the queue, which can effectively reduce the frequency at which the queue manager sends scheduling requests to the queue scheduler, thereby reducing The scheduling load of the small queue scheduler reduces the probability that the queue scheduler has a large scheduling delay for the scheduling of the queue.

DRAWINGS

FIG. 1A is a schematic diagram of an implementation environment involved in a queue scheduling method according to an exemplary embodiment of the present disclosure.

FIG. 1B is a schematic diagram of a principle of queue scheduling provided by an exemplary embodiment of the present application.

FIG. 2 is a flowchart of a queue scheduling method provided by an exemplary embodiment of the present application.

FIG. 3 is a flowchart of a method for detecting whether a queue length of a first queue changes according to an exemplary embodiment of the present application.

FIG. 4 is a flowchart of a method for determining a target probability based on a target length difference of a first queue and a queue length of a first queue when a queue length of a first queue changes, according to an exemplary embodiment of the present application.

FIG. 5 is a flowchart of a method for determining a target probability based on at least one of a first difference value and a second difference value, and a queue length of a first queue, according to an exemplary embodiment of the present application.

FIG. 6 is a flowchart of a method for determining a target probability according to a length difference when a current queue credit value of a first queue changes according to an exemplary embodiment of the present application.

FIG. 7A is a block diagram of a queue scheduling apparatus provided by an exemplary embodiment of the present application.

FIG. 7B is a block diagram of a first determining module provided by an exemplary embodiment of the present application.

FIG. 7C is a block diagram of a first determining sub-module provided by an exemplary embodiment of the present application.

FIG. 7D is a block diagram of another first determining module provided by an exemplary embodiment of the present application.

FIG. 7E is a block diagram of a second determining sub-module provided by an exemplary embodiment of the present application.

FIG. 7F is a block diagram of another queue scheduling apparatus provided by an exemplary embodiment of the present application.

FIG. 8 is a schematic structural diagram of a queue scheduling apparatus according to an exemplary embodiment of the present application.

Detailed ways

In order to make the objects, technical solutions and advantages of the present application more clear, the embodiments of the present application will be further described in detail below with reference to the accompanying drawings.

Referring to FIG. 1A, a schematic diagram of an implementation environment involved in a queue scheduling method provided in an exemplary embodiment of the present application is shown. As shown in FIG. 1A, the implementation environment may include a communication device 110, a first terminal 120, and a second terminal 130.

The communication device 110 can be a router or a switch. The first terminal 120 and the second terminal 130 can be a smart phone, a computer, a multimedia player, an e-reader, or a wearable device. The communication device 110 and the first terminal 120 The connection between the communication device 110 and the second terminal 130 can be established through a wired network or a wireless network. During communication between the first terminal 120 and the second terminal 130, the communication device 110 can set the information of the first terminal 120. Forwarded to the second terminal 130. This information is usually transmitted in the form of a data packet.

Generally, the communication device 110 includes a queue manager 1101 and a queue scheduler 1102, and the queue manager 1101 and the queue scheduler 1102 can be connected by wire or wirelessly.

The queue manager 1101 manages a plurality of queues, and the plurality of queues are divided by the queue manager 1101 according to the description information of the received data packet (for example, output port, transmission priority, and user identity (ID), etc.) The obtained, and the plurality of data packets in each queue are arranged in chronological order according to the queue manager 1101 receiving the data packets.

The queue manager 1101 manages a plurality of queues, which is actually a queue scheduling process. The queue scheduling process mainly includes an enrollment management process and a dequeue management process. The following two aspects are described in the following embodiments:

First, the enqueue management process: after receiving the data packet, the queue manager 1101 performs an enqueue operation on the data packet to add the data packet to the end of the queue. The enqueue operation may be: adding the descriptor of the data packet to the end of the linked list corresponding to the queue.

Second, the dequeue management process: first, the queue manager 1101 sends a scheduling request to the queue scheduler 1102 according to the queue status of the queue it manages; then, the queue scheduler 1102 is configured according to the scheduling request sent by the queue manager 1101 and the user configuration. The QoS policy generates a scheduling response and sends the scheduling response to the queue manager 1101. Finally, after receiving the scheduling response sent by the queue scheduler 1102, the queue manager 1101 responds to the corresponding one of the queues managed by the queue based on the scheduling response. The packet performs a dequeue operation. The dequeue operation may be: reading the descriptor of the packet located at the beginning of the queue from the linked list.

Further, referring to FIG. 1B, the queue manager 1101 may include: an enqueue module, a queue management module, and a dequeue module. The three modules cooperate to implement a queue scheduling process. The specific process of the queue scheduling may refer to the following process:

After the queue manager 1101 receives a certain data packet, the enqueue module may divide the data packet into a corresponding queue according to the description information of the certain data packet (ie, perform a queue operation), and then the queue management module The managed queue performs enqueue update. Then, the queue manager 1101 can perform queue status update according to the queue length of the queue it manages and the current queue credit value of the queue, and send a schedule to the queue scheduler 1102 based on the updated queue status. a request, the scheduling request indicates a target queue that needs to be scheduled, and a scheduling state of the target queue. After receiving the scheduling request, the queue scheduler 1102 may run a preset scheduling algorithm according to the scheduling state indicated by the scheduling request, where The target queue provides a dequeue credit, which is usually equal to the scheduling granularity, and the scheduling granularity is the rated data amount allocated for the data transmission by the queue scheduler when scheduling the queue (for example, the scheduling granularity may be It is 1 kilobyte (KB)). After receiving the scheduling response with the dequeue credit returned by the queue scheduler, the queue management module updates the current queue credit value of the target queue according to the dequeue credit, and sends a dequeue request to the dequeue module to request the dequeue module pair. The data packet performs the dequeue operation. After receiving the dequeue request, the dequeue module may dequeue the data packet of the head of the target queue based on the current queue credit value of the updated target queue, thereby implementing data transmission. Moreover, after the data packet is dequeued, the queue management module can dequeue the queue managed by the queue according to the dequeue operation of the data packet.

The current queue credit value is a data amount quota that the queue is currently available for data transmission. The scheduling status is used to identify the rate at which the queue scheduler schedules queues. The normal scheduling state has two states: a slow scheduling state and a fast scheduling state, where the slow scheduling state indicates the first rate, and the fast scheduling state indicates the second rate, where the first rate is smaller than the second rate. The rate is less than the preset scheduling rate threshold, and the second rate is greater than the preset scheduling rate threshold. For example, the first rate may be: 200 clock cycles, and the second rate may be: 10 clock cycles.

FIG. 2 is a flowchart of a queue scheduling method according to an exemplary embodiment of the present disclosure. The queue scheduling method may be applied to the queue manager shown in FIG. 1A or FIG. 1B. The queue scheduling method takes the first queue as an example. To illustrate, the first queue is any one of at least one queue managed by the queue manager. As shown in FIG. 2, the queue scheduling method may include:

Step 201: When the length difference of the first queue changes, the target probability is determined based on the length difference of the first queue and the queue length of the first queue.

The length difference is the difference between the current queue credit value of the first queue and the queue length of the first queue. Moreover, there are many factors that cause the difference in the length of the first queue to change, for example, the following two factors: First, the queue length of the first queue changes, for example, when a new data packet is added to the first queue, The queue length of the first queue is updated according to the length of the data packet and the historical length of the first queue. Correspondingly, the length difference of the first queue changes. Second, the current queue credit value of the first queue changes. For example, when the queue scheduler sends the scheduling result to the first queue, the queue manager updates the current queue credit value based on the scheduling result, and the current queue of the first queue. The credit value will change. Correspondingly, the difference in the length of the first queue will change.

Therefore, it can be determined whether the length difference of the first queue changes by detecting whether the queue length of the first queue and/or the current queue credit value changes, for example, in the queue length of the first queue and the current queue credit value. When at least one of the changes occurs, it is determined that the difference in length of the first queue changes.

Optionally, as shown in FIG. 3, the process of detecting whether the queue length of the first queue changes may include:

Step a: Detect whether a new data packet is added to the first queue.

The implementation process of the step a may include: detecting whether the length of the linked list corresponding to the first queue is increased, and when the length of the linked list corresponding to the first queue is increased, determining that a descriptor of a new data packet is added to the linked list, That is, it can be determined that a new data packet is added to the first queue, and accordingly, it can be determined that the queue length of the first queue changes.

Step b: When a new data packet is added to the first queue, it is determined that the queue length of the first queue changes.

It should be noted that the action of detecting the queue length of the first queue and determining that the queue length of the first queue changes may be performed by the queue management module in FIG. 1B, and in actual implementation, the queue management module may be configured with The module dedicated to detecting the length of the queue, or the queue management module may also determine that a new data packet is added to the first queue after receiving the enqueue update request sent to the enqueue module, and determine that the queue length of the first queue occurs. The embodiment of the present invention does not specifically limit it.

For the above two factors that cause the difference in the length of the first queue to change, the exemplary embodiments of the present application correspondingly provide the following two ways to determine the target probability:

In the first implementation manner, when the queue length of the first queue changes, the target probability is determined based on the target length difference of the first queue and the queue length of the first queue.

The target length difference may be at least one of a first difference value and a second difference value, where the first difference value is a difference between a current queue credit value of the first queue and a queue length of the first queue, where the The difference is the difference between the current queue credit value of the first queue and the queue length of the first queue.

Optionally, referring to FIG. 4, the implementation process of the first implementation manner may include:

Step 2011: When the queue length of the first queue changes, the first difference after the change of the length difference is obtained, and the second difference before the change.

When a new data packet is added to the first queue, the current queue length of the first queue may be updated according to the length of the data packet and the historical length of the first queue, and the current queue length may be the length of the data packet and the first The sum of the historical lengths of the queues. Correspondingly, the first difference is the difference between the current queue credit value of the first queue and the current queue length of the first queue, and the second difference is the difference between the current queue credit value of the first queue and the historical length of the first queue. value.

For example, suppose that a packet is added to the first queue at a certain time, the size of the enqueue packet is 128 bytes (Byte, B), and the history queue length of the first queue before the packet is added to the first queue. For 128B, the current queue credit value is 1KB. Then, after the data packet is added, the current queue length of the first queue is 128B+128B=256B, the first difference is 1KB-256B=768B, and the second difference is 1KB. -128B=896B.

Step 2012: Determine a target probability based on at least one of the first difference and the second difference, and a queue length of the first queue.

In practical applications, when the sizes of the first difference and the second difference are different, the method for determining the target probability is different. As shown in FIG. 5, according to the size of the first difference and the second difference, the step 2012 The implementation process can include the following situations:

In the first case, the first difference is compared with the reference credit threshold and the scheduling granularity respectively. When the first difference is greater than the reference credit threshold and less than the scheduling granularity, the target probability is determined based on the first difference and the scheduling granularity.

Optionally, the implementing method for determining the target probability based on the first difference value and the scheduling granularity may include: determining a probability that the target difference value is greater than the first random number as the target probability.

The target difference is a difference between the scheduling granularity and the first difference. The first random number is a positive number that is randomly generated and is not greater than the scheduling granularity, and the reference credit threshold is smaller than the scheduling granularity. The reference credit threshold and the scheduling granularity may be set according to actual needs. For example, the reference credit threshold may be 0, and the scheduling granularity may be 1 KB.

In the first case, when the first difference is greater than the reference credit threshold and less than the scheduling granularity, it may be determined that the first queue still has a certain queue credit, but the remaining queue credit is not enough for the first queue. The data packet is dequeued. At this time, in order to ensure the effective dequeue of the data packet in the first queue, the first queue can request the queue scheduler to schedule it, but in order to avoid multiple queues managed by the queue manager The scheduling request is sent to the queue scheduler. At this time, the target probability of generating the scheduling request may be determined, so that the scheduling request is randomly generated according to the target probability, so as to reduce the frequency at which the queue manager sends the scheduling request to the queue scheduler, The scheduling load of the small queue scheduler, in order to ensure the effective dequeue of the data packet.

For example, assuming that the first difference is 768B, the reference credit threshold is 0, and the scheduling granularity is 1 KB. At this time, if the first difference is greater than the reference credit threshold and less than the scheduling granularity, the target may be determined based on the first difference and the scheduling granularity. Probability, and the target difference value is 1KB-768B=256B. Since the first random number is randomly generated and is not greater than the positive number of the scheduling granularity, the probability that the target difference value is greater than the first random number can be determined=256B/1KB=1 /4, the target probability is 1/4.

The second case: comparing the second difference with the reference credit threshold and the scheduling granularity respectively, and when the second difference is greater than the reference credit threshold and less than the scheduling granularity, determining the target probability based on the second difference and the target packet length.

Optionally, the method for determining the target probability based on the second difference and the target packet length may include: determining a probability that the target packet length is greater than the second random number as the target probability.

The second random number is a positive number that is randomly generated and is not greater than the second difference, and the target packet length is a length of the data packet that changes the queue length of the first queue, that is, the target packet length is newly added. The length of a queue of packets.

When the second difference is greater than the reference credit threshold and less than the scheduling granularity, it may be determined that the first queue still has a certain queue credit, but the remaining queue credit is not enough for the data packet in the first queue to be dequeued. In order to ensure the effective dequeue of the data packets in the first queue, and to avoid sending a scheduling request to the queue scheduler by multiple queues managed by the queue manager, the target probability of generating the scheduling request may be determined at this time, so that The scheduling request is randomly generated according to the target probability, thereby ensuring effective dequeue of the data packet.

For example, it is assumed that the length of the data packet newly added to the first queue (ie, the target packet length) is 128B, the second difference is 896B, the reference credit threshold is 0, and the scheduling granularity is 1 KB. At this time, the second difference is greater than the reference. The credit threshold is less than the scheduling granularity, and the target probability may be determined based on the second difference and the target packet length, and, since the second random number is randomly generated and not greater than the positive number of the second difference, the target packet length may be determined to be greater than The probability of the second random number = 128B/896B = 1/7, that is, the target probability is 1/7.

Moreover, each time the queue length of the first queue changes, the target probability determined by the second case is used to determine the target probability, and the scheduling probability may not be generated based on the target probability determined multiple times. The situation, however, the target probability determined multiple times will be incremented in order of time, and finally the target probability will be 1, and a scheduling request can be generated.

For the sake of brevity of description, the method for determining the target probability provided by the second case is simply referred to as the first determining method, and the target probability determined by the first determining method is referred to as the i-th target probability, and the i is positive. Integer. If the scheduling request is not generated based on the first target probability, and a new data packet is added to the first queue, the length of the newly added data packet is assumed to be 128B. In this case, the second determined by the first determining method is used. The difference = 1 KB - 128 B - 128 B = 768 B, the probability that the target packet length is greater than the second random number = 128 B / 768 B = 1 / 6, and the second target probability is 1/6.

If the scheduling request is not generated based on the second target probability, and a new data packet is added to the first queue, it is assumed that the length of the newly added data packet is 128B. At this time, the second difference determined based on the first determining method is determined. The value=1KB-128B-128B-128B=640B, the probability that the target packet length is greater than the second random number=128B/640B=1/5, that is, the third target probability is 1/5.

By analogy, the fourth target probability is 1/4, the fifth target probability is 1/3, the sixth target probability is 1/2, and the seventh target probability is 1, and the scheduling request can be generated once.

It can be seen that, in this case, when the target probability is determined based on the second difference and the target packet length, the determined target probability exhibits a more uniform change, which can further reduce the queue manager sending the scheduling request to the queue scheduler. The frequency, and thus the scheduling load of the queue scheduler, can further reduce the probability that the queue scheduler has a large scheduling delay for scheduling of the queue.

The third case: comparing the first difference with the reference credit threshold and the scheduling granularity respectively, comparing the second difference with the reference credit threshold and the scheduling granularity respectively, when the first difference and the second difference are both satisfied: greater than the reference When the credit threshold is less than the scheduling granularity, the target probability is determined based on the second difference and the target packet length.

Optionally, the method for determining the target probability based on the second difference and the target packet length may include: determining a ratio of the target packet length to the second difference as the target probability.

For example, assume that the length of the data packet newly added to the first queue (ie, the target packet length) is 128B, the first difference is 768B, the second difference is 896B, the reference credit threshold is 0, and the scheduling granularity is 1 KB. And the first difference value and the second difference value both satisfy: greater than the reference credit threshold and less than the scheduling granularity, the target probability may be determined based on the second difference value and the target packet length, and the ratio of the target packet length to the second difference value is= 128B/896B=1/7, the target probability can be determined to be 1/7.

In the fourth case, the first difference is compared with the reference credit threshold and the preset first credit threshold, and when the first difference is greater than the first credit threshold and less than the reference credit threshold, the target probability is determined to be 1.

The first credit threshold is less than a reference credit threshold, and the first credit threshold is a threshold corresponding to a serious deficit in the queue credit. For example, the first credit threshold may be a credit threshold corresponding to the fast dispatch waterline, when the first difference is When the value is less than the first credit threshold, the first queue urgently needs the queue scheduler to schedule it. For example, the first queue needs the queue scheduler to schedule it at a relatively large rate to ensure the normal data packets in the first queue. Departure. Moreover, the first credit threshold may be set according to actual needs, for example, the first credit threshold may be -10K.

For example, assuming that the first difference is -768B, the reference credit threshold is 0, and the first credit threshold is -10 KB. At this time, if the first difference is greater than the first credit threshold and less than the reference credit threshold, the target probability may be determined as 1, and the first queue requires the queue scheduler to schedule it at a relatively small rate.

In the fifth case, the first difference is compared with a preset first credit threshold, and when the first difference is smaller than the first credit threshold, the target probability is determined to be 1.

It should be noted that when at least one of the first difference value and the second difference value is compared according to the above five cases, when at least one of the first difference value and the second difference value does not satisfy the above five cases In the comparison result, it can be determined that the first queue does not need to be scheduled.

In a second implementation manner, when the current queue credit value of the first queue changes, the target probability is determined according to the length difference.

Optionally, referring to FIG. 6, the implementation process of the second implementation manner may include the following situations:

In the first case, the length difference is compared with the reference credit threshold and the scheduling granularity respectively. When the length difference is greater than the reference credit threshold and less than the scheduling granularity, the target probability is determined based on the length difference and the scheduling granularity.

Optionally, the method for determining the target probability based on the length difference and the scheduling granularity may include: determining a probability that the length difference is smaller than the first random number as the target probability.

The length difference is the difference between the current queue credit value of the first queue and the queue length of the first queue after the length difference is changed.

For example, assuming that the length difference is 768B, the reference credit threshold is 0, and the scheduling granularity is 1 KB. At this time, if the length difference is greater than the reference credit threshold and less than the scheduling granularity, the target probability may be determined based on the length difference and the scheduling granularity, and Since the first random number is randomly generated and is not greater than the positive number of the scheduling granularity, it may be determined that the probability that the length difference is smaller than the first random number=768B/1KB=1/4, that is, the target probability is 3/4.

In the second case, the length difference is respectively compared with the reference credit threshold and the preset first credit threshold. When the length difference is greater than the first credit threshold and less than the reference credit threshold, the target probability is determined to be 1.

In the second case, although the queue credit of the first queue has a deficit, but the deficit is not serious, at this time, the implementation process of the second case should be referred to according to the first difference and the second difference. The size of the value determines the implementation of the fourth case of the target probability, and will not be described here.

In the third case, the length difference is compared with a preset first credit threshold, and when the length difference is less than the first credit threshold, the target probability is determined to be 1.

In the third case, the queue courier of the first queue has a severe deficit, and the first queue urgently needs the queue scheduler to schedule it. At this time, the implementation process of the third case should be referred to according to the first difference sum. The size of the second difference determines the implementation process of the fifth case of the target probability, and details are not described herein again.

It should be noted that when the length difference does not satisfy the comparison result in the above five cases, it may be determined that the first queue does not need to be scheduled.

In practical applications, in the five cases of the first implementable manner described above, and in the three cases of the second implementable manner, the process of comparing the length difference with each range may be performed in parallel, or may be performed in series. The execution of the line is not specifically limited in the exemplary embodiment of the present application.

For example, in the first implementation manner, when the process of comparing the length difference with each range is performed serially, the first difference may be compared with a preset first credit threshold. When the difference is not less than the first credit threshold, the first difference is compared with the reference credit threshold, and when the first difference does not satisfy the first credit threshold and is less than the reference credit threshold, the second difference and the scheduling granularity are used. For comparison, when the second interpolation difference does not satisfy the reference credit threshold and is smaller than the scheduling granularity, the first difference is compared with the scheduling granularity, and then the target probability is determined according to the comparison result.

In the second implementation manner, when the process of comparing the length difference with each range is serial execution, the length difference may be first compared with a preset first credit threshold, when the length difference is not less than When the first credit threshold is used, the length difference is compared with the reference credit threshold. When the length difference does not satisfy the first credit threshold and is less than the reference credit threshold, the length difference is compared with the scheduling granularity, and then determined according to the comparison result. Target probability.

And, since the first credit threshold is a threshold corresponding to a serious deficit in the queue credit, the first difference (or the length difference) is first compared with a preset first credit threshold, when the first difference (or length) When the difference is less than the first credit threshold, the operation of generating the scheduling request may be performed according to the situation of the severe deficit in time, thereby timely alleviating the situation that the queue credit has a serious deficit, so as to ensure the normal dequeue of the data packet in the first queue. .

Step 202: Perform random generation of the scheduling request based on the target probability, such that the ratio of the probability of generating the scheduling request to the probability of not being generated is the target probability.

The scheduling request is used to indicate a scheduling state for scheduling the first queue.

Correspondingly, based on the different situations of the foregoing first implementable manner and the second implementable manner, the implementable manner of performing the random generation of the scheduling request based on the target probability may include: when selecting the generating the scheduling request based on the target probability, A scheduling request is generated based on the selection result. The scheduling request includes a first status identifier or a second status identifier, where the first status identifier indicates a first rate when the queue scheduler performs queue scheduling, and the second status identifier indicates a second time when the queue scheduler performs queue scheduling. rate. The first rate is less than a preset scheduling rate threshold, and the second rate is greater than a preset scheduling rate threshold.

And, referring to FIG. 5 and FIG. 6, when the target probability is based on the first one and the second case in the first to fourth cases and the second achievable manner in the first implementable manner, The first queue needs the queue scheduler to schedule it at a relatively small rate. At this time, the generated scheduling request may include the first state identifier; when the target probability is based on the fifth case of the first implementable manner described above In the third case of the second implementation, the first queue needs the queue scheduler to schedule it at a relatively large rate. In this case, the generated scheduling request may include the second status identifier.

For example, it is assumed that the target probability determined in step 201 based on the first difference value and the scheduling granularity (ie, the first case of the first implementable mode) is 1/4, and at this time, when selecting whether to generate a scheduling request, scheduling The ratio of the probability of request generation to the probability of not being generated is 1/4, and when the scheduling request is generated, a scheduling request may be generated, and the generated scheduling request includes a first state identifier to request the queue scheduler to The first rate schedules the first queue.

It should be noted that when the random generation of the scheduling request is performed based on the target probability, since the ratio of the possibility of generating the scheduling request to the probability of not generating is the target probability, the queue manager is reduced based on the plurality of queues managed by the queue manager. The number of scheduling requests is such that the number of scheduling requests generated by the queue manager is as uniformly distributed as possible, which reduces the frequency at which the queue manager sends scheduling requests to the queue scheduler, and can reduce the queue. The scheduling load of the scheduler is such that the update rate of the scheduling state does not exceed the scheduling capability of the queue scheduler, thereby reducing the probability that the queue scheduler has a large scheduling delay for scheduling of the queue.

In addition, the number of scheduling requests generated by the queue manager can be uniformly distributed as shown in the queue manager, so that the current queue credits of all the queues are evenly distributed within the preset range corresponding to the scheduling granularity, and the queue credits of all the queues are The sum of the sums is close to the total scheduling bandwidth that the queue scheduler can schedule, so that the actual total scheduling bandwidth is more accurate, and the scheduling bandwidth caused by the high frequency of sending scheduling requests and the scheduling bandwidth ratio between the queues are inaccurate. And other issues.

Step 203: When a scheduling request is generated, the scheduling request is sent to the queue scheduler.

Optionally, the sending the scheduling request to the queue scheduler includes: determining whether the scheduling status indicated by the scheduling request is the same as the current scheduling status corresponding to the first queue, and the current status corresponding to the first queue when the scheduling status indicated by the scheduling request is When the scheduling status is different, the scheduling request is sent to the queue scheduler.

Step 204: Receive a scheduling response sent by a queue scheduler.

The scheduling response includes scheduling granularity.

After receiving the scheduling request sent by the queue manager, the queue scheduler may run a scheduling algorithm according to the scheduling state indicated by the scheduling request, allocate a rated data amount for data transmission to the first queue, and send the same to the queue manager. A scheduling response of the amount of data so that the queue manager performs a dequeue operation on the packets in the first queue.

Step 205: Update a current queue credit value of the first queue based on the scheduling response.

After the queue manager receives the scheduling response sent by the queue scheduler, the current queue credit value of the first queue may be updated based on the rated data amount carried in the scheduling response, and the updating method may include: the current queue credit value when the scheduling request is generated. The sum of the rated data amount is determined as the updated current queue credit value.

It should be noted that, after updating the current queue credit value of the first queue, since the current queue credit value of the first queue changes, the length difference of the first queue also changes accordingly. The changed length difference and the queue length of the first queue determine the target probability again, so as to more accurately determine the current scheduling state, so that the queue scheduler can more effectively schedule the first queue, thereby realizing efficient use of resources.

Step 206: Perform a dequeuing operation on the data packet of the corresponding length in the first queue based on the updated current queue credit value.

After updating the current queue credit value, if the current queue credit value satisfies the dequeue data volume of the data packet in the first queue, the dequeuing operation may be performed on the data packet of the corresponding length in the first queue to implement data transmission. Moreover, since the dequeued data packet consumes the queue credit of the same data amount in the dequeuing operation, the length difference of the first queue does not change after the operation dequeuing operation is performed, and there is no need to update the scheduling at this time. status.

Optionally, in the foregoing step 201, an action action of determining a target probability based on a length difference of the first queue and a queue length of the first queue, and an action of step 202 to step 205 may be performed by the queue management module in FIG. 1B, where The actions of step 206 can all be performed by the dequeue module of Figure 1B.

It should be noted that, in the actual implementation, the foregoing process may be implemented by simple code, for example, it may be implemented by using an code written by an if statement, which is not specifically limited herein.

In summary, the queue scheduling method provided by the exemplary embodiment of the present application determines the target probability by changing the length difference of the first queue, and performs random generation of the scheduling request based on the target probability, so that the scheduling request may be generated. The ratio of the likelihood to the ungenerated probability is the target probability, and when the scheduling request is generated, the scheduling request is sent to the queue scheduler, and the queue manager is configured to generate a schedule based on the plurality of queues managed by the queue manager relative to the related art. The number of requests is such that the number of scheduling requests generated by the queue manager is uniformly distributed as shown in the queue manager, which can effectively reduce the frequency at which the queue manager sends scheduling requests to the queue scheduler, thereby reducing queue scheduling. The scheduling load of the device reduces the probability that the queue scheduler has a large scheduling delay for the scheduling of the queue.

It should be noted that the sequence of the steps of the queue scheduling method provided by the embodiment of the present invention may be appropriately adjusted, and the steps may also be correspondingly increased or decreased according to the situation. Anyone skilled in the art may be within the technical scope disclosed by the present invention. Methods that can be easily conceived of variations are encompassed within the scope of the present invention and therefore will not be described again.

The following is an apparatus embodiment of the present application, which can be used to implement the method embodiments of the present application. For details not disclosed in the device embodiments of the present application, please refer to the method embodiments of the present application.

Please refer to FIG. 7A, which is a block diagram of a queue scheduling apparatus 700 provided by an exemplary embodiment of the present application. As shown in FIG. 7A, the apparatus 700 may include:

The first determining module 701 is configured to determine a target probability based on a length difference of the first queue and a queue length of the first queue when the length difference of the first queue changes.

The generating module 702 is configured to perform random generation of the scheduling request based on the target probability, such that the ratio of the probability of generating the scheduling request to the probability of not being generated is a target probability, and the scheduling request is used to indicate a scheduling state for scheduling the first queue. .

The sending module 703 is configured to send a scheduling request to the queue scheduler when the scheduling request is generated.

The first queue is any one of the at least one queue managed by the queue manager, and the length difference is the difference between the current queue credit value of the first queue and the queue length of the first queue, and the current queue credit value is The amount of data that a queue currently has for data transmission.

In summary, the queue scheduling apparatus provided by the exemplary embodiment of the present application determines the target probability when the length difference of the first queue changes by the first determining module, and the generating module performs random generation of the scheduling request based on the target probability. The ratio of the possibility of generating the scheduling request to the probability of not being generated is the target probability. When the sending module generates the scheduling request, the sending module sends the scheduling request to the queue scheduler, which reduces the queue manager based on the related technology. The number of scheduling requests generated by the plurality of queues is such that the number of scheduling requests generated by the queue manager is uniformly distributed, which can effectively reduce the queue manager sending scheduling requests to the queue scheduler. The frequency, in turn, reduces the scheduling load of the queue scheduler, and reduces the probability that the queue scheduler has a large scheduling delay for scheduling of the queue.

As shown in FIG. 7B, the first determining module 701 may include:

The obtaining sub-module 7011 is configured to: when the queue length of the first queue changes, obtain a first difference after the length difference is changed, and a second difference before the change.

The first determining sub-module 7012 is configured to determine a target probability based on at least one of the first difference value and the second difference value, and a queue length of the first queue.

In an implementation manner, as shown in FIG. 7C, the first determining submodule 7012 may include:

The first comparison unit 7012a is configured to compare the first difference value with the reference credit threshold and the scheduling granularity respectively, where the reference credit threshold is smaller than the scheduling granularity, and the scheduling granularity is allocated by the queue scheduler to the first queue when scheduling the first queue. The nominal amount of data used for data transmission.

The first determining unit 7012b is configured to determine a target probability based on the first difference and the scheduling granularity when the first difference is greater than a reference credit threshold and less than a scheduling granularity.

Optionally, the first determining unit 7012b is configured to:

The probability that the target difference is greater than the first random number is determined as the target probability, and the target difference is the difference between the scheduling granularity and the first difference, and the first random number is a positive number that is randomly generated and not greater than the scheduling granularity.

In another implementation manner, as shown in FIG. 7C, the first determining submodule 7012 may include:

The first comparing unit 7012a is configured to compare the second difference with the reference credit threshold and the scheduling granularity respectively, where the reference credit threshold is smaller than the scheduling granularity, and the scheduling granularity is allocated by the queue scheduler to the first queue when scheduling the first queue. The nominal amount of data used for data transmission.

The first determining unit 7012b is configured to determine a target probability based on the second difference and the target packet length when the second difference is greater than the reference credit threshold and less than the scheduling granularity, and the target packet length is to change the queue length of the first queue. The length of the packet.

Optionally, the first determining unit 7012b is configured to:

In another implementation manner, as shown in FIG. 7C, the first determining sub-module 7012 may include:

The first comparing unit 7012a is configured to compare the second difference value with the reference credit threshold and the scheduling granularity, respectively.

The first determining unit 7012b is configured to: when the first difference value and the second difference value are both greater than the reference credit threshold and less than the scheduling granularity, determine the target probability based on the second difference and the target packet length, and the target packet length is The length of a packet whose queue length changes.

Optionally, the first determining unit 7012b is configured to:

The ratio of the target packet length to the second difference is determined as the target probability.

In a further implementation manner, as shown in FIG. 7C, the first determining sub-module 7012 may include:

The first comparing unit 7012a is configured to compare the first difference value with the reference credit threshold and the preset first credit threshold, where the first credit threshold is smaller than the reference credit threshold.

The first determining unit 7012b is configured to determine that the target probability is 1 when the first difference is greater than the first credit threshold and less than the reference credit threshold.

In a further implementation manner, as shown in FIG. 7C, the first determining submodule 7012 may include:

The first comparing unit 7012a is configured to compare the first difference with a preset first credit threshold.

The first determining unit 7012b is configured to determine that the target probability is 1 when the first difference is less than the first credit threshold.

As shown in FIG. 7D, the first determining module 701 may include:

The second determining sub-module 7013 is configured to determine a target probability according to the length difference when the current queue credit value of the first queue changes.

As shown in FIG. 7E, the second determining submodule 7013 may include:

The second comparison unit 7013a is configured to compare the length difference value with the reference credit threshold and the scheduling granularity respectively, and the reference credit threshold is smaller than the scheduling granularity, and the scheduling granularity is used by the queue scheduler to allocate the first queue. The amount of data that is rated for data transmission.

The second determining unit 7013b is configured to determine the target probability based on the length difference and the scheduling granularity when the length difference is greater than the reference credit threshold and less than the scheduling granularity.

Optionally, the second determining unit 7013b is configured to:

The probability that the length difference is smaller than the first random number is determined as the target probability, and the first random number is a positive number that is randomly generated and not larger than the scheduling granularity.

As shown in FIG. 7E, the second determining submodule 7013 may include:

The second comparing unit 7013a is configured to compare the length difference value with the reference credit threshold and the preset first credit threshold, where the first credit threshold is smaller than the reference credit threshold.

The second determining unit 7013b is configured to determine that the target probability is 1 when the length difference is greater than the first credit threshold and less than the reference credit threshold.

As shown in FIG. 7E, the second determining submodule 7013 may include:

The second comparing unit 7013a is configured to compare the length difference with a preset first credit threshold.

The second determining unit 7013b is configured to determine that the target probability is 1 when the length difference is less than the first credit threshold.

Optionally, the generating module 702 is configured to:

When the scheduling request is generated based on the target probability, the scheduling request is generated according to the selection result, and the scheduling request may include a first state identifier, where the first state identifier indicates a first rate when the queue scheduler performs queue scheduling, and the first rate is less than a preset. The scheduling rate threshold.

Optionally, the generating module 702 is configured to:

When the scheduling request is generated based on the target probability, the scheduling request is generated based on the selection result, the scheduling request may include a second status identifier, where the second status identifier indicates a second rate when the queue scheduler performs the queue scheduling, and the second rate is greater than the preset. The scheduling rate threshold.

Optionally, as shown in FIG. 7F, the apparatus 700 may further include:

The receiving module 704 is configured to receive a scheduling response sent by the queue scheduler, where the scheduling response may include a scheduling granularity, where the scheduling granularity is a rated data amount allocated by the queue scheduler for the first queue for data transmission.

The update module 705 is configured to update the current queue credit value of the first queue based on the scheduling response.

Optionally, as shown in FIG. 7F, the apparatus 700 may further include:

The executing module 706 is configured to perform a dequeuing operation on the data packet of the corresponding length in the first queue based on the updated current queue credit value.

Optionally, the sending module 703 is configured to:

It is determined whether the scheduling state indicated by the scheduling request is the same as the current scheduling state corresponding to the first queue.

When the scheduling status indicated by the scheduling request is different from the current scheduling status corresponding to the first queue, the scheduling request is sent to the queue scheduler.

Optionally, as shown in FIG. 7F, the apparatus 700 may further include:

The second determining module 707 is configured to determine that the length difference of the first queue changes when the queue length of the first queue changes.

And/or, the second determining module 707 is configured to determine that the length difference of the first queue changes when the current queue credit value of the first queue changes.

A person skilled in the art can clearly understand that, for the convenience and brevity of the description, the specific working process of the device, the module and the sub-module described above can refer to the corresponding process in the foregoing method embodiment, and details are not described herein again.

The exemplary embodiment of the present application further provides a communication device, where the communication device includes: a queue scheduler and a queue manager;

The queue manager includes a queue scheduling device as shown in FIG. 7A or 7F.

The queue scheduler is configured to send a scheduling response to the queue manager after receiving the scheduling request sent by the queue manager, where the scheduling response is used to allocate a nominal amount of data for data transmission based on the scheduling request.

The exemplary embodiment of the present application further provides a queue scheduling apparatus, where the apparatus includes:

Processing component

a memory for storing executable instructions of the processing component;

Wherein the processing component is configured to:

When the length difference of the first queue changes, the target probability is determined based on the length difference of the first queue and the queue length of the first queue;

Performing a random generation of the scheduling request based on the target probability, such that the ratio of the probability of generating the scheduling request to the probability of not being generated is a target probability, and the scheduling request is used to indicate a scheduling state for scheduling the first queue;

When a scheduling request is generated, the scheduling request is sent to the queue scheduler;

The exemplary embodiment of the present application provides a queue scheduling apparatus 20, including a memory, a processor, and a computer program stored on the memory and operable on the processor. When the processor executes the computer program, the foregoing embodiment of the present application is implemented. Any of the queue scheduling methods.

Specifically, the terminal may be the queue scheduling device 20 as shown in FIG. 8 . Please refer to FIG. 8 , which is a schematic structural diagram of a queue scheduling device 20 according to an exemplary embodiment of the present application. It can include a processor 22 and a signal interface 24.

Processor 22 includes one or more processing cores. The processor 22 executes various functional applications and data processing by running software programs and modules.

There may be a plurality of signal interfaces 24 for establishing connections with other devices or modules, for example, through which the speakers or headphones may be connected.

Optionally, the queue scheduling device 20 further includes components such as a memory 26, a bus 28, and the like. The memory 26 and the signal interface 24 are connected to the processor 22 via a bus 28, respectively.

Memory 26 can be used to store software programs as well as modules. In particular, memory 26 may store application modules 262 required for at least one function.

The application module 262 can include:

The first determining unit 2621 has the same or similar function as the first determining module 701.

The generating unit 2622 has the same or similar function as the generating module 702.

The transmitting unit 2623 has the same or similar function as the transmitting module 703.

The exemplary embodiments of the present application also provide a storage medium, which may be a non-transitory computer readable storage medium, enabling a terminal to execute an example of the present application when instructions in the storage medium are executed by a processing component of the terminal The queue scheduling method provided by the embodiment.

The exemplary embodiments of the present application also provide a computer program product comprising instructions that, when executed on a computer, cause the computer to perform the queue scheduling method provided by the exemplary embodiments of the present application.

It should be noted that the queue scheduling method provided in the foregoing embodiment can also be used in other scenarios that require a single individual to perform rounding up or rounding, so that the overall randomness of multiple individuals is uniform, and the overall remainder is as close as possible. The default value is, for example, the multi-port concatenation at the interface needs to be sent to a fixed length, and needs to be randomly and uniformly transmitted to avoid bursts, which is not specifically limited in the exemplary embodiment of the present application.

A person skilled in the art may understand that all or part of the steps of implementing the above embodiments may be completed by hardware, or may be instructed by a program to execute related hardware, and the program may be stored in a computer readable storage medium. The storage medium mentioned may be a read only memory, a magnetic disk or an optical disk or the like.

The above description is only an optional embodiment of the present application, and is not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application are included in the protection of the present application. Within the scope.

Claims

A queue scheduling method, wherein the method is applied to a queue manager, the method comprising:

Determining a target probability based on a length difference of the first queue and a queue length of the first queue, when a difference in length of the first queue changes;

And performing, according to the target probability, a random generation of the scheduling request, such that a ratio of the possibility of generating the scheduling request to an ungenerated possibility is the target probability, where the scheduling request is used to indicate that the first queue is performed Scheduled scheduling status;

When the scheduling request is generated, the scheduling request is sent to a queue scheduler;

The first queue is any one of at least one queue managed by the queue manager, and the length difference is a current queue credit value of the first queue and a queue length of the first queue. The difference between the current queue credits is the amount of data that the first queue is currently available for data transmission.
The method according to claim 1, wherein the determining a target probability based on a length difference of the first queue and a queue length of the first queue comprises:

And acquiring, when the queue length of the first queue changes, a first difference after the change in the length difference, and a second difference before the change;

The target probability is determined based on at least one of the first difference and the second difference, and a queue length of the first queue.
The method according to claim 2, wherein the determining the target probability based on at least one of the first difference value and the second difference value, and a queue length of the first queue, including :

Comparing the first difference with a reference credit threshold and a scheduling granularity, where the reference credit threshold is smaller than the scheduling granularity, where the scheduling granularity is that when the queue scheduler schedules the first queue, The nominal amount of data allocated for data transmission allocated by the first queue;

When the first difference is greater than the reference credit threshold and less than the scheduling granularity, the target probability is determined based on the first difference and the scheduling granularity.
The method according to claim 3, wherein the determining the target probability based on the first difference value and the scheduling granularity comprises:

Determining, by the probability that the target difference is greater than the first random number, the target probability, where the target difference is a difference between the scheduling granularity and the first difference, the first random number is randomly generated and not A positive number greater than the scheduling granularity.
The method according to claim 2, wherein the determining the target probability based on at least one of the first difference value and the second difference value, and a queue length of the first queue, including :

Comparing the second difference with a reference credit threshold and a scheduling granularity, where the reference credit threshold is smaller than the scheduling granularity, where the scheduling granularity is when the queue scheduler schedules the first queue, The nominal amount of data allocated for data transmission allocated by the first queue;

Determining the target probability based on the second difference and a target packet length when the second difference is greater than the reference credit threshold and less than the scheduling granularity, where the target packet length is the first queue The length of the packet whose length has changed.
The method according to claim 5, wherein the determining the target probability based on the second difference and a target packet length comprises:

The probability that the target packet length is greater than the second random number is determined as the target probability, and the second random number is a positive number that is randomly generated and not greater than the second difference.
The method according to claim 2, wherein the determining the target probability based on at least one of the first difference value and the second difference value, and a queue length of the first queue, including :

Comparing the first difference with the reference credit threshold and the scheduling granularity, where the reference credit threshold is smaller than the scheduling granularity, and the scheduling granularity is that when the queue scheduler schedules the first queue, The nominal amount of data allocated for data transmission allocated by the first queue;

Comparing the second difference with the reference credit threshold and the scheduling granularity, respectively;

And determining, when the first difference value and the second difference value are greater than the reference credit threshold and less than the scheduling granularity, determining the target probability based on the second difference value and a target packet length, The target packet length is the length of the data packet that changes the queue length of the first queue.
The method according to claim 7, wherein the determining the target probability based on the second difference and a target packet length comprises:

A ratio of the target packet length to the second difference is determined as the target probability.
The method according to claim 2, wherein the determining the target probability based on at least one of the first difference value and the second difference value, and a queue length of the first queue, including :

Comparing the first difference with a reference credit threshold and a preset first credit threshold, where the first credit threshold is less than the reference credit threshold;

When the first difference is greater than the first credit threshold and less than the reference credit threshold, determining that the target probability is 1.
The method according to claim 2, wherein the determining the target probability based on at least one of the first difference value and the second difference value, and a queue length of the first queue, including :

Comparing the first difference with a preset first credit threshold;

When the first difference is less than the first credit threshold, the target probability is determined to be 1.
The method according to claim 1, wherein the determining a target probability based on a length difference of the first queue and a queue length of the first queue comprises:

When the current queue credit value of the first queue changes, the target probability is determined according to the length difference.
The method according to claim 11, wherein said determining said target probability based on said difference in length comprises:

Comparing the length difference with the reference credit threshold and the scheduling granularity, where the reference credit threshold is smaller than the scheduling granularity, and the scheduling granularity is when the queue scheduler schedules the first queue, The nominal amount of data allocated for data transmission by the first queue;

When the length difference is greater than the reference credit threshold and less than the scheduling granularity, the target probability is determined based on the length difference and the scheduling granularity.
The method according to claim 12, wherein the determining the target probability based on the length difference and the scheduling granularity comprises:

Determining, by the probability that the length difference is smaller than the first random number, is the target probability, and the first random number is a positive number that is randomly generated and not greater than the scheduling granularity.
The method according to claim 11, wherein the determining the target probability according to the length difference comprises:

Comparing the length difference with a reference credit threshold and a preset first credit threshold, where the first credit threshold is less than the reference credit threshold;

When the length difference is greater than the first credit threshold and less than the reference credit threshold, determining that the target probability is 1.
The method according to claim 11, wherein the determining the target probability according to the length difference comprises:

Comparing the length difference with a preset first credit threshold;

When the length difference is less than the first credit threshold, the target probability is determined to be 1.
The method according to any one of claims 3 to 9, or the method according to any one of claims 12 to 14, wherein the performing the random generation of the scheduling request based on the target probability comprises:

When the scheduling request is generated based on the target probability, the scheduling request is generated based on the selection result, where the scheduling request includes a first status identifier, where the first status identifier indicates that the queue scheduler performs queue scheduling The first rate, the first rate is less than a preset scheduling rate threshold.
The method according to claim 10 or 15, wherein the performing the random generation of the scheduling request based on the target probability comprises:

When the scheduling request is generated based on the target probability, the scheduling request is generated based on the selection result, where the scheduling request includes a second status identifier, where the second status identifier indicates that the queue scheduler performs queue scheduling The second rate is greater than a preset scheduling rate threshold.
The method according to any one of claims 1 to 15, wherein after the scheduling request is sent to the queue scheduler, the method further comprises:

Receiving a scheduling response sent by the queue scheduler, where the scheduling response includes a scheduling granularity, where the scheduling granularity is a rated data amount allocated by the queue scheduler for the first queue for data transmission;

Updating the current queue credit value of the first queue based on the scheduling response.
The method according to claim 18, wherein after the updating the current queue credit value of the first queue based on the scheduling response, the method further comprises:

Dequeuing the data packet of the corresponding length in the first queue based on the updated current queue credit value.
The method according to any one of claims 1 to 15, wherein the sending the scheduling request to a queue scheduler comprises:

Determining whether the scheduling state indicated by the scheduling request is the same as the current scheduling state corresponding to the first queue;

And when the scheduling state indicated by the scheduling request is different from the current scheduling state corresponding to the first queue, the scheduling request is sent to the queue scheduler.
The method according to any one of claims 1 to 15, wherein when the length difference of the first queue changes, the difference between the length of the first queue and the queue of the first queue is Before the length determines the target probability, the method further includes:

When the queue length of the first queue changes, determining that the length difference of the first queue changes;

And/or, when the current queue credit value of the first queue changes, determining that the length difference of the first queue changes.
A queue scheduling device, characterized in that the device comprises:

a first determining module, configured to determine a target probability based on a length difference of the first queue and a queue length of the first queue when a difference in length of the first queue changes;

a generating module, configured to perform random generation of the scheduling request based on the target probability, such that a ratio of a probability of generating the scheduling request to an ungenerated possibility is the target probability, and the scheduling request is used to indicate a target Decoding the scheduling state of the first queue;

a sending module, configured to send the scheduling request to a queue scheduler when the scheduling request is generated;

The first queue is any one of at least one queue managed by the queue manager, and the length difference is a current queue credit value of the first queue and a queue length of the first queue. The difference between the current queue credits is the amount of data that the first queue is currently available for data transmission.
The device according to claim 22, wherein the first determining module comprises:

Obtaining a sub-module, configured to: when the queue length of the first queue changes, obtain a first difference after the length difference is changed, and a second difference before the change;

a first determining submodule, configured to determine the target probability based on at least one of the first difference value and the second difference value, and a queue length of the first queue.
The apparatus according to claim 23, wherein the first determining submodule comprises:

a first comparing unit, configured to compare the first difference with a reference credit threshold and a scheduling granularity, where the reference credit threshold is smaller than the scheduling granularity, and the scheduling granularity is that the queue scheduler pairs the first The rated data amount allocated for the first queue for data transmission when the queue is scheduled;

a first determining unit, configured to determine the target probability based on the first difference value and the scheduling granularity when the first difference value is greater than the reference credit threshold and less than the scheduling granularity.
The device according to claim 24, wherein the first determining unit is configured to:

Determining, by the probability that the target difference is greater than the first random number, the target probability, where the target difference is a difference between the scheduling granularity and the first difference, the first random number is randomly generated and not A positive number greater than the scheduling granularity.
The apparatus according to claim 23, wherein the first determining submodule comprises:

a first comparing unit, configured to compare the second difference with a reference credit threshold and a scheduling granularity, where the reference credit threshold is smaller than the scheduling granularity, and the scheduling granularity is that the queue scheduler pairs the first The rated data amount allocated for the first queue for data transmission when the queue is scheduled;

a first determining unit, configured to determine, according to the second difference and a target packet length, the target probability, when the second difference is greater than the reference credit threshold and less than the scheduling granularity, the target packet length The length of the data packet for which the queue length of the first queue changes.
The device according to claim 26, wherein the first determining unit is configured to:

The probability that the target packet length is greater than the second random number is determined as the target probability, and the second random number is a positive number that is randomly generated and not greater than the second difference.
The apparatus according to claim 23, wherein the first determining submodule comprises:

a first comparing unit, configured to compare the first difference value with a reference credit threshold and a scheduling granularity, where the reference credit threshold is smaller than the scheduling granularity, and the scheduling granularity is that the queue scheduler pairs the first The rated data amount allocated for the first queue for data transmission when the queue is scheduled;

The first comparing unit is configured to compare the second difference value with the reference credit threshold and the scheduling granularity respectively;

a first determining unit, configured to determine, when the first difference value and the second difference value are greater than the reference credit threshold and less than the scheduling granularity, based on the second difference value and a target packet length The target probability, the target packet length is a length of a data packet that changes a queue length of the first queue.
The device according to claim 28, wherein the first determining unit is configured to:

A ratio of the target packet length to the second difference is determined as the target probability.
The apparatus according to claim 23, wherein the first determining submodule comprises:

a first comparison unit, configured to compare the first difference value with a reference credit threshold and a preset first credit threshold, where the first credit threshold is smaller than the reference credit threshold;

The first determining unit is configured to determine that the target probability is 1 when the first difference is greater than the first credit threshold and less than the reference credit threshold.
The apparatus according to claim 23, wherein the first determining submodule comprises:

a first comparing unit, configured to compare the first difference with a preset first credit threshold;

a first determining unit, configured to determine that the target probability is 1 when the first difference is less than the first credit threshold.
The device according to claim 22, wherein the first determining module comprises:

a second determining submodule, configured to determine the target probability according to the length difference when a current queue credit value of the first queue changes.
The apparatus according to claim 32, wherein the second determining submodule comprises:

a second comparing unit, configured to compare the length difference value with a reference credit threshold and a scheduling granularity, where the reference credit threshold is smaller than the scheduling granularity, where the scheduling granularity is that the queue scheduler is to the first queue The amount of rated data allocated for the first queue for data transmission when scheduling is performed;

a second determining unit, configured to determine the target probability based on the length difference and the scheduling granularity when the length difference is greater than the reference credit threshold and less than the scheduling granularity.
The device according to claim 33, wherein the second determining unit is configured to:

Determining, by the probability that the length difference is smaller than the first random number, is the target probability, and the first random number is a positive number that is randomly generated and not greater than the scheduling granularity.
The apparatus according to claim 32, wherein the second determining submodule comprises:

a second comparing unit, configured to compare the length difference value with a reference credit threshold and a preset first credit threshold, where the first credit threshold is smaller than the reference credit threshold;

And a second determining unit, configured to determine that the target probability is 1 when the length difference is greater than the first credit threshold and less than the reference credit threshold.
The apparatus according to claim 32, wherein the second determining submodule comprises:

a second comparing unit, configured to compare the length difference with a preset first credit threshold;

a second determining unit, configured to determine that the target probability is 1 when the length difference is less than the first credit threshold.
The device according to any one of claims 24 to 30 or 33 to 35, wherein the generating module is configured to:

When the scheduling request is generated based on the target probability, the scheduling request is generated based on the selection result, where the scheduling request includes a first status identifier, where the first status identifier indicates that the queue scheduler performs queue scheduling The first rate, the first rate is less than a preset scheduling rate threshold.
The device according to claim 31 or 36, wherein the generating module is configured to:

When the scheduling request is generated based on the target probability, the scheduling request is generated based on the selection result, where the scheduling request includes a second status identifier, where the second status identifier indicates that the queue scheduler performs queue scheduling The second rate is greater than a preset scheduling rate threshold.
The device according to any one of claims 22 to 36, wherein the device further comprises:

a receiving module, configured to receive a scheduling response sent by the queue scheduler, where the scheduling response includes a scheduling granularity, where the scheduling granularity is a rated data volume allocated by the queue scheduler for the first queue for data transmission ;

And an update module, configured to update a current queue credit value of the first queue based on the scheduling response.
The device of claim 39, wherein the device further comprises:

And an execution module, configured to perform a dequeuing operation on the data packet of the corresponding length in the first queue based on the updated current queue credit value.
The device according to any one of claims 22 to 36, wherein the sending module is configured to:

Determining whether the scheduling state indicated by the scheduling request is the same as the current scheduling state corresponding to the first queue;

And when the scheduling state indicated by the scheduling request is different from the current scheduling state corresponding to the first queue, the scheduling request is sent to the queue scheduler.
The device according to any one of claims 22 to 36, wherein the device further comprises:

a second determining module, configured to determine, when the queue length of the first queue changes, a change in a length difference of the first queue;

And/or, the second determining module is configured to determine that a difference in length of the first queue changes when a current queue credit value of the first queue changes.
A communication device, comprising: a queue scheduler and a queue manager;

The queue manager includes the queue scheduling apparatus according to any one of claims 22 to 42;

The queue scheduler is configured to send a scheduling response to the queue manager after receiving the scheduling request sent by the queue manager, where the scheduling response is used to allocate rated data for data transmission based on the scheduling request. the amount.
A queue scheduling device, characterized in that the device comprises:

Processing component

a memory for storing executable instructions of the processing component;

Wherein the processing component is configured to:

Determining a target probability based on a length difference of the first queue and a queue length of the first queue, when a difference in length of the first queue changes;

And performing, according to the target probability, a random generation of the scheduling request, such that a ratio of the possibility of generating the scheduling request to an ungenerated possibility is the target probability, where the scheduling request is used to indicate that the first queue is performed Scheduled scheduling status;

When the scheduling request is generated, the scheduling request is sent to a queue scheduler;

The first queue is any one of at least one queue managed by the queue manager, and the length difference is a current queue credit value of the first queue and a queue length of the first queue. The difference between the current queue credits is the amount of data that the first queue is currently available for data transmission.
A storage medium characterized by enabling the terminal to execute the queue scheduling method according to any one of claims 1 to 21 when an instruction in the storage medium is executed by a processing component of the terminal.