CN102263701B - Queue regulation method and device - Google Patents

Abstract

The invention discloses a queue adjustment method and device, wherein the queue adjustment method includes: comparing the relationship between the occupied length of the memory space of the current quality of service (QOS) out of the queue and the set first threshold and the second threshold, wherein the first Both the threshold and the second threshold are less than the length of the memory space of the current QOS dequeue, and the first threshold is less than the second threshold; if the occupied length of the memory space of the current QOS dequeue is greater than the second threshold, the occupied length of other memory spaces is less than Select a queue from the QOS out-queue of the first threshold, wherein the current QOS out-of-queue and other QOS out-of-queues are set with borrow buffers; borrow the borrowed buffer of the selected QOS out-of-queue to the tail of the current QOS out-of-queue . Through the present invention, the effect of making full use of the overall resource of the QOS queue and improving the service quality is achieved.

Description

Queue adjustment method and device

technical field

The present invention relates to the communication field, in particular, to a queue adjustment method and device for quality of service (Quality Of Service, QOS) dequeue.

Background technique

Quality of Service (QOS) refers to a collection of a series of technologies that allow users to obtain predictable service levels in terms of throughput, delay, jitter, and packet loss rate.

For QOS outgoing queues, the port QOS outgoing queue lengths are evenly distributed, and the switch allocates 8 queues of the same length to all ports. At present, the QOS dequeuing technology is that when a QOS dequeue is full, the packets put in the queue will be discarded, and at this time, other QOS dequeues may be idle. For example, when a certain queue (such as queue 7) of port 1 is full, the newly added message on queue 7 will be discarded. At this time, other queues of this port may be idle or have few messages , In addition, the QOS queues of other ports may also be idle or have few packets.

It can be seen that in actual use, in most cases, fewer queues will be full, while more queues are idle or there are few packets, and the queue resources cannot be fully utilized.

Therefore, in the current QOS dequeuing technical scheme, when the queues of some ports are full and cause packet loss, other QOS dequeues may be idle or rarely used at this time, so that the overall QOS dequeue Queue resources are not fully used.

Contents of the invention

The main purpose of the present invention is to provide a queue adjustment method and device for QOS out-of-queue, so as to at least solve the above-mentioned problem that the existing QOS out-of-queue resources are not fully used as a whole.

According to one aspect of the present invention, a queue adjustment method is provided, including: comparing the relationship between the occupied length of the memory space of the current quality of service QOS out of the queue and the set first threshold and the second threshold, wherein the first threshold and The second thresholds are smaller than the length of the memory space of the current QOS queue, and the first threshold is smaller than the second threshold; if the occupied length of the memory space of the current QOS queue is greater than the second threshold, the occupied length of other memory spaces is smaller than the first threshold. Select a queue in the QOS queue of the threshold value, wherein, the current QOS queue and other QOS queues are all provided with borrow buffers; the borrow buffer of the selected QOS queue is borrowed to the tail of the current QOS queue.

Preferably, the length of the borrowed buffer is: the length of the borrowed buffer=the length of the QOS queue-the second threshold+(the second threshold-the first threshold)×N%, where N is an integer between 0-100.

Preferably, the queue adjustment method further includes: if the occupied length of the memory space of the current QOS dequeue is less than the first threshold, then setting the current QOS dequeue as a borrowable queue, allowing other QOS dequeues to supply their borrow buffers.

Preferably, the queue adjustment method also includes: if the occupied length of the memory space of the current QOS out of the queue is less than or equal to the second threshold and greater than or equal to the first threshold, then the current QOS out of the queue judges whether its borrowed buffer is borrowed; if so, Then take back its borrowed buffer, and redistribute the borrowed buffer for the QOS that borrowed its borrowed buffer to go out of the queue; if not, then set the current QOS going out of the queue as a queue that cannot be borrowed, and refuse to lend its borrowed buffer to other QOS queues.

Preferably, the queue adjustment method further includes: when the occupied length of the memory space of the current QOS dequeue decreases from greater than the second threshold to less than or equal to the second threshold, then the current QOS dequeues and returns the borrowed buffer.

Preferably, the queue adjustment method further includes: the system saves the corresponding relationship between the current QOS out-queue and the QOS out-queue to which the borrowed buffer belongs, and updates it in real time.

According to another aspect of the present invention, a queue adjustment device is provided, including: a comparison module, which is used to compare the relationship between the occupied length of the memory space of the current quality of service QOS out of the queue and the set first threshold and the second threshold, Wherein, the first threshold and the second threshold are all less than the memory space length of the current QOS out of the queue, and the first threshold is less than the second threshold; the first allocation module is used for when the comparison result of the comparison module is that the memory space of the current QOS out of the queue is When the occupied length is greater than the second threshold, a queue is selected from other QOS queues whose memory space occupied length is less than the first threshold, wherein, the current QOS queue and other QOS queues are all provided with a borrow buffer; The borrow buffer of the QOS dequeue is borrowed to the end of the current QOS dequeue.

Preferably, the length of the borrowed buffer is: the length of the borrowed buffer=the length of the QOS queue-the second threshold+(the second threshold-the first threshold)×N%, where N is an integer between 0-100.

Preferably, the queue adjustment device also includes: a second allocation module, used to set the current QOS out of the queue as a borrowable queue if the comparison result of the comparison module is that the memory space of the current QOS out of the queue is occupied by a length less than the first threshold, allowing Provide its borrowed buffer to other QOS queues; and/or, the third allocation module, if the comparison result of the comparison module is that the occupied length of the memory space of the current QOS queue is less than or equal to the second threshold and greater than or equal to the second threshold A threshold, then the current QOS out of the queue judges whether its borrowed buffer is borrowed; if it is, then withdraw its borrowed buffer, and redistribute the borrowed buffer for the QOS out of the queue that borrowed its borrowed buffer; if not, set the current QOS The outgoing queue is a non-borrowable queue, and refuses to lend its borrow buffer to other QOS queues.

Preferably, the queue adjustment device further includes: a return module, used for when the occupied length of the memory space of the current QOS dequeue is reduced from greater than the second threshold to less than or equal to the second threshold, then the current QOS dequeues and returns the borrowed buffer Area.

Through the present invention, a queue buffer that can be shared is divided into all QOS queues. When the use of a certain QOS queue exceeds the second threshold, other QOS queues whose queue length is less than the first threshold are used. The queue buffer is borrowed and added to the queue, so that the length of the queue is greater than the original initially allocated length, so that when the message in the QOS out queue exceeds the initial length of the QOS out queue, the message can still be cached in the QOS out queue. In the queue, it will not be discarded, thus solving the problem that the existing QOS queue resources are not fully used as a whole, and then achieve the effect of making full use of the overall resources of the QOS queue and improving the quality of service.

Description of drawings

The accompanying drawings described here are used to provide a further understanding of the present invention and constitute a part of the application. The schematic embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations to the present invention. In the attached picture:

FIG. 1 is a flowchart of steps of a queue adjustment method according to Embodiment 1 of the present invention;

FIG. 2 is a flowchart of steps of a queue adjustment method according to Embodiment 2 of the present invention;

Fig. 3 is a queue initial allocation diagram in a queue adjustment method according to Embodiment 3 of the present invention;

4 is a buffer queue management diagram in a queue adjustment method according to Embodiment 3 of the present invention;

FIG. 5 is a graph of queue borrowable lengths in a queue adjustment method according to Embodiment 3 of the present invention;

Fig. 6 is a buffer queue recovery diagram in a queue adjustment method according to Embodiment 3 of the present invention;

7 is a buffer queue borrowing diagram in a queue adjustment method according to Embodiment 3 of the present invention;

FIG. 8 is a buffer queue return diagram in a queue adjustment method according to Embodiment 3 of the present invention;

FIG. 9 is a management diagram of buffer queue borrowing in a queue adjustment method according to Embodiment 3 of the present invention;

Fig. 10 is a structural block diagram of a queue adjustment device according to Embodiment 4 of the present invention.

detailed description

Hereinafter, the present invention will be described in detail with reference to the drawings and examples. It should be noted that, in the case of no conflict, the embodiments in the present application and the features in the embodiments can be combined with each other.

Embodiment one

Referring to FIG. 1 , it shows a flowchart of steps of a queue adjustment method according to Embodiment 1 of the present invention.

The queue adjustment method in this embodiment includes the following steps:

Step S102: compare the relationship between the occupied length of the memory space of the current QOS queue and the set first threshold and the second threshold;

Wherein, both the first threshold and the second threshold are smaller than the memory space length of the current QOS dequeuing, and the first threshold is smaller than the second threshold.

In this embodiment, all the QOS queues of a port are provided with borrow buffers. When a QOS queue is idle or rarely used, its borrow buffer can be lent to other QOS queues; When the QOS outgoing queue is very full, other borrowable buffers can be borrowed to increase the queue length and expand the queue space.

In this embodiment, when the queue space occupied length is less than the first threshold, it indicates that the queue is relatively idle, and there is still more free queue space that can be used and can be borrowed from other queues; when the queue occupied space length is greater than or equal to the first threshold When a threshold is less than or equal to the second threshold, it means that the queue is busy and cannot lend its borrowed buffer to other queues, and it is possible to borrow borrowed buffers from other queues; when the length of the queue occupied space is greater than the second threshold , it means that the queue is very busy and needs to borrow corresponding borrow buffers from other queues that can lend buffers.

Step S104: If the occupied length of the memory space of the current QOS out-queue is greater than the second threshold, select a queue from other QOS out-queues whose memory space occupied length is less than the first threshold;

Wherein, both the current QOS exit queue and other QOS exit queues are provided with a borrow buffer, and the borrow buffer is a shared buffer that can be borrowed from other QOS exit queues when necessary.

Step S106: Borrow the borrow buffer of the selected QOS exit queue to the tail of the current QOS exit queue.

By borrowing the borrow buffer of the selected QOS dequeue to the end of the current QOS dequeue, the buffer space of the current QOS dequeue can be expanded so as to buffer more messages. The method of borrowing can be flexibly set by those skilled in the art according to the actual situation, such as using a linked list method, or using a table record method, etc., as long as the address of the provided borrowing buffer can be determined to store message data.

In related technologies, in the QOS out-of-queue technical solution, when the queues of some ports are full and cause packet loss, other QOS out-of-queues may be idle or rarely used at this time, so that the QOS out-of-queue queues in the overall queue Resources are underutilized. Through this embodiment, a queue buffer that can be shared is divided into all QOS dequeues. When the use of a certain QOS dequeue exceeds the second threshold, other QOS dequeues whose queue length is less than the first threshold are dequeued. The queue buffer of the queue is borrowed and added to the queue, so that the length of the queue is greater than the original initial allocation length, so that when the message in the QOS out-queue exceeds the initial length of the QOS out-queue, the message can still be cached in the QOS In the queue, it will not be discarded, thus solving the problem that the existing QOS queue resources are not fully used on the whole, and then achieve the effect of making full use of the overall resources of the QOS queue and improving the quality of service.

Embodiment two

Referring to FIG. 2 , it shows a flowchart of steps of a queue adjustment method according to Embodiment 2 of the present invention.

In this embodiment, a queue buffer that can be shared is divided into all QOS outgoing queues. When a queue is full, other idle queues are used to increase the length of the queue, reduce the number of lost packets, and provide better QOS Serve.

The queue adjustment method in this embodiment includes the following steps:

Step S202: For each QOS dequeue of each port, set the minimum buffer return threshold (ie, the first threshold) and the maximum buffer borrowing threshold (ie, the second threshold).

Step S204: All the queues whose queue length used in the QOS dequeue does not exceed the minimum buffer return threshold are managed by the software; for the queues that have been borrowed, the relationship between borrowing and being borrowed is recorded and managed by the software system.

Among them, the queues whose queue lengths used in all QOS output queues do not exceed the minimum buffer return threshold can be uniformly managed by means of linked lists, or can be managed uniformly by means of table registration. Of course, those skilled in the art can also use Other suitable ways, the present invention is not limited thereto. Through the unified management of the queue length used in all QOS queues that do not exceed the minimum buffer return threshold, it is possible to obtain and provide loanable cache space conveniently and quickly when there is a QOS queue that needs to borrow cache space from other queues In addition, it is also convenient to understand and obtain the situation of the queues that can lend buffer space in real time, and improve system efficiency.

For the queues that have been borrowed, there is a one-to-one correspondence between the queues that borrow the cache space and the queues that are borrowed. Record this relationship in time. On the one hand, you can keep track of the borrowing and borrowing in the queue. On the other hand, When the queue that borrows the cache space no longer needs to borrow, the borrowed buffer can be returned to the corresponding queue in a timely and effective manner.

The borrowable length of each QOS outqueue can be the remaining queue greater than the maximum buffer borrowing threshold. Preferably, the borrowable length of each QOS queue can be composed of two parts, one part is the remaining queue greater than the maximum buffer borrow threshold, and the other part is the difference between the maximum buffer borrow threshold and the minimum buffer return threshold multiplied by A configurable percentage of the space length.

Step S206: Perform corresponding queue adjustment processing according to the occupied length of the memory space of the current QOS out-queue.

specifically,

When the usage of the queue exceeds the minimum buffer return threshold, the borrow buffer of the queue is set to be unborrowable. If the borrowed buffer of the queue has been borrowed, it needs to be withdrawn, and the borrowed buffer of a queue can be reallocated from the system to the borrower.

When the use of the queue changes from the minimum buffer return threshold to exceeding the maximum buffer borrow threshold, a borrow is found from the idle buffer queue list, and the length of the queue is increased. The search for the idle buffer queue gives priority to other queues on the port where the borrowed queue is located. If other queues cannot be borrowed or have already been borrowed, the queues on other ports are used to borrow.

When the usage of the queue changes from exceeding the maximum buffer borrowing threshold to being less than the maximum buffer borrowing threshold, the borrowed buffer is returned.

When the usage of the queue changes from exceeding the minimum buffer return threshold to being less than the minimum buffer return threshold, the borrow buffer of the queue is set as available for borrowing.

Through the scheme of dynamically adjusting the queue size provided by this embodiment, for each QOS queue of each port, the minimum buffer return threshold (also referred to as the minimum threshold in this paper) and the maximum buffer borrowing threshold (this paper) are respectively set. is also referred to as the maximum threshold), and the part of the queue that exceeds the maximum threshold and the part of the queue between the minimum threshold and the maximum threshold are jointly set as queues that can be borrowed (also referred to as buffer queues in this article), When the length of the used queue of the port exceeds the minimum threshold, the reporting software indicates that the queue cannot lend out the buffer queue (by default, all ports can lend out the buffer queue). When the queue length exceeds the maximum threshold, the software can detect which other port buffer queues can be borrowed. After finding out, borrow the idle buffer queue to the busy queue, increase the length of the queue, and reduce the length of the idle queue. It starts to get busy, even if the queue length exceeds the minimum threshold, the borrowed buffer must be returned. In order to prevent buffer shocks, each queue can set a high maximum threshold and a low minimum threshold. When the queue length increases from the minimum threshold to the maximum threshold, the high level is triggered, and the queue length is increased from the maximum threshold. When it decreases to the minimum threshold, the low bit is triggered.

Compared with the prior art, this embodiment dynamically adjusts the size of the queue by sharing part of the resources of the queue, so that the resource utilization of the queue is more fully utilized. When the use of a certain queue exceeds the maximum threshold, add other buffer queues whose queue length is less than the minimum threshold to the queue, so that the length of the queue is greater than the original initial allocation length. When the length is long, the packets can still be buffered in the queue without being discarded, thereby improving the quality of service.

Embodiment Three

In this embodiment, the QOS queue adjustment method of the present invention is described in detail by taking the queue adjustment during the dynamic change process of a QOS out-queue queue usage as an example.

First, referring to FIG. 3 , it shows a queue initial allocation diagram in a queue adjustment method according to Embodiment 3 of the present invention. As shown in Figure 3, the present embodiment divides each QOS dequeuing into three parts: the part (vertical stripe part in the figure) that is less than the minimum threshold (ie the first threshold), the minimum threshold and the maximum threshold (ie the first threshold) The part between the two thresholds (the diagonal stripe part in the figure) and the part greater than the maximum threshold (the remaining queue behind the diagonal stripe part in the figure, that is, the cross stripe part).

Among them, when the occupied length of the memory space of the queue is less than the minimum threshold, that is, when the use of the queue is less than the minimum threshold, the queue is set as a borrowable queue, allowing to lend its cache to other QOS queues; when When the occupied length of the memory space of the queue is between the minimum threshold and the maximum threshold, that is, when the usage of the queue is between the minimum threshold and the maximum threshold, the queue is set as a non-borrowable queue. At this time, Refuse to lend its cache to other QOS queues; when the occupied length of the memory space of the queue is greater than the maximum threshold, that is, when the usage of the queue is greater than the maximum threshold, it means that the usage of the queue is about to reach the maximum value of the queue , so it needs to be borrowed from other queues to increase the available length of the queue.

Referring to FIG. 4 , it shows a buffer queue management diagram in a queue adjustment method according to Embodiment 3 of the present invention. As shown in Figure 4, in the initial stage of the system, this embodiment manages all idle queues (that is, queues whose memory space occupation length is less than the minimum threshold), and connects them into a linked list available for borrowing, and the linked list only records the Queues use queues smaller than the minimum threshold. At the same time, all the borrowing queues and borrowed queues are connected into a linked list, and the borrowing and borrowing status of the linked list are recorded, which is convenient for returning after borrowing.

Referring to FIG. 5 , it shows a queue borrowable length diagram in a queue adjustment method according to Embodiment 3 of the present invention. As shown in FIG. 5 , this embodiment designs the calculation of the length of the queues available for borrowing, and acquires the queues available for borrowing. The length of the queue that can be borrowed is composed of two parts, one is the remaining queue that is greater than the maximum threshold, and the other is the difference between the maximum threshold and the minimum threshold multiplied by a configurable percentage, which is configurable and determined according to the actual usage adjust. The position of the length available for borrowing is allocated from the tail of the queue forward, as shown in the horizontal stripes in the figure.

Specifically, the borrowable queue length=queue length-maximum threshold+(maximum threshold-minimum threshold)*N%, where N is an integer between 0-100 and can be flexibly configured according to actual needs.

Referring to FIG. 6 , it shows a buffer queue reclamation diagram in a queue adjustment method according to Embodiment 3 of the present invention.

As shown in FIG. 6 , when the memory space used by a queue (for example, queue 7 of port 1 ) increases from less than the minimum threshold to greater than the minimum threshold, the queue needs to be set as a non-borrowable queue. If the queue is not borrowed by other queues, that is, the queue is in the queue available for borrowing, take the queue out of the queue available for borrowing, that is, the queue can no longer be borrowed; if the queue has been borrowed by other queues, Then forcibly take back the queue borrowed by the queue, delete it from the borrowed queue list, and at the same time, re-allocate a borrowable queue from the borrowable queues for the queue that borrowed the queue.

Specifically, including:

Step S602: Search for the queue from the borrowable queue, if found, it means that the queue is not borrowed, delete the queue from the borrowable queue, and do not perform the following steps; if not found, execute step S604.

Step S604: the queue has been borrowed, the queue is searched from the borrowed relationship table, the borrower who borrows the queue is found, and the borrowed queue is withdrawn from the borrower.

Step S606: Borrow a new queue from the borrowable queue for the borrower who borrowed this queue. If the borrowable queue is empty, the borrower will not be able to borrow the queue.

Referring to FIG. 7 , it shows a buffer queue borrowing diagram in a queue adjustment method according to Embodiment 3 of the present invention.

As shown in Figure 7, when the memory space used by a queue (queue 7 of port 1 in the figure) changes from less than the maximum threshold to exceeding the maximum threshold, the usage of the queue is about to exceed the maximum value of the queue, and other queues need to be borrowed to Increase the size of the queue, find a queue that can be borrowed from the borrowable queue (queue 6 of port 1 in the figure), put the borrowed queue at the end of the current queue, increase the length of the queue, and reduce the borrowed queue at the same time length.

Specifically, including:

Step S702: Find a queue that can be borrowed from the queues that can be borrowed (queue 6 of port 1 in the figure), if no queue can be found for borrowing, do nothing, that is, the following steps do not need to be executed; If a queue available for borrowing is found, go to step S704.

Step S704: Delete the borrowing queue (queue 6 of port 1 in the figure) from the available queues to prevent other queues from borrowing.

Of course, in actual operation, other ways, such as marking the borrowing queue, can also be used to distinguish whether it is borrowed or not.

Step S706: Add the borrowing queue (queue 7 of port 1 in the figure) and the borrowed queue (queue 6 of port 1 in the figure) into the borrowing queue relationship table.

Step S708: Add the borrowed queue (such as queue 6 of port 1 in the figure) to the end of the borrowing queue (such as queue 7 of port 1 in the figure), and increase the length of the queue.

Referring to FIG. 8 , it shows a buffer queue return diagram in a queue adjustment method according to Embodiment 3 of the present invention.

As shown in Figure 8, when the memory space used by the queue (queue 7 of port 1 in the figure) is reduced from greater than the maximum threshold to less than the maximum threshold, the length of the queue is sufficient, and other queues that have been borrowed need to be returned. Update the queue borrowing relationship table.

Specifically, including:

Step S802: Check the queue borrowing relationship table to find out whether there is a record of the queue borrowing other queues, if yes, execute step S804; if not, do nothing, that is, the following steps do not need to be executed.

Step S804: Remove the borrowed queue (queue 6 of port 1 in the figure) from the end of the borrowing queue (queue 7 of port 1 in the figure) and return it to the borrowed queue.

Step S806: Delete the borrowing relationship between the borrowing queue and the borrowed queue in the borrowing queue relationship table.

Step S808: Add the queue to be borrowed (for example, the queue 6 of port 1 in the figure) to the list of queues available for borrowing, and other queues that need to borrow the queue can borrow the queue.

Referring to FIG. 9 , it shows a buffer queue borrowing management diagram in a queue adjustment method according to Embodiment 3 of the present invention.

As shown in Figure 9, when the memory space used by the queue (queue 7 of port 1 in the figure) is reduced from greater than the minimum threshold to less than the minimum threshold, the queue is relatively idle and can be set as a borrowable queue, increasing to the linked list of the borrowable queue.

Through this embodiment, the queue size is dynamically adjusted by sharing part of the resources out of each QOS queue, so that the queue resources are more fully utilized and the service quality is improved.

Embodiment four

Referring to FIG. 10 , it shows a structural block diagram of a queue adjustment device according to Embodiment 4 of the present invention.

The queue adjustment device of this embodiment includes: a comparison module 1002, which is used to compare the relationship between the occupied length of the memory space of the current QOS dequeue and the set first threshold and the second threshold, wherein the first threshold and the second threshold are both Less than the memory space length of the current QOS dequeue, the first threshold is less than the second threshold; the first allocation module 1004, when the comparison result of the comparison module 1002 is that the occupied length of the memory space of the current QOS dequeue is greater than the second threshold, Select a queue from other QOS out-queues whose memory space occupied length is less than the first threshold, wherein, the current QOS out-queue and other QOS out-queues are all provided with a borrow buffer; the borrowed buffer of the selected QOS out of the queue Borrow to the tail of the current QOS queue.

Preferably, the length of the borrow buffer is: the length of the borrow buffer=QOS queue length-the second threshold+(the second threshold-the first threshold)×N%, wherein, N is 0- Integer between 100.

Preferably, the queue adjustment device of this embodiment further includes: a second distribution module 1006, configured to set the current QOS queue out if the length of the memory space occupied by the current QOS out of the queue is less than the first threshold if the comparison result of the comparison module 1002 For a borrowable queue, it is allowed to provide its borrowed buffer to other QOS queues; and/or, the third distribution module 1008 is used for if the comparison result of the comparison module 1002 is that the occupied length of the memory space of the current QOS queue is less than or equal to The second threshold is greater than or equal to the first threshold, then the current QOS out of the queue judges whether its borrowed buffer is lent; ; If not, set the current QOS out queue as a non-borrowable queue, and refuse to lend its borrow buffer to other QOS queues.

Preferably, the queue adjustment device of this embodiment also includes: a return module, used to return the current QOS out of the queue when the occupied length of the memory space of the current QOS out of the queue is reduced from greater than the second threshold to less than or equal to the second threshold The borrow buffer for borrowing.

Preferably, the queue adjustment device of this embodiment further includes: a saving module, used for the system to save the corresponding relationship between the current QOS out-queue and the QOS out-queue to which the borrowed buffer belongs, and update it in real time.

For the method implemented in this embodiment, reference may be made to the related descriptions of the aforementioned corresponding method embodiments, and it has the beneficial effects of the aforementioned embodiments, so details are not repeated here.

From the above description, it can be seen that the QOS out-queue queue adjustment technical solution on the distributed switch provided by the present invention dynamically adjusts according to the queue usage of the QOS out-queue, and by sharing some resources of the queue, the queue Resources are more fully utilized. Especially when the use of a certain queue exceeds the maximum threshold, add other buffer queues whose queue length is less than the minimum threshold to the queue, so that the length of the queue is greater than the original initial allocation length. When the packets in the queue exceed the queue When the original length is set, the packets can still be buffered in the queue without being discarded, thus improving the quality of service.

Obviously, those skilled in the art should understand that each module or each step of the above-mentioned present invention can be realized by a general-purpose computing device, and they can be concentrated on a single computing device, or distributed in a network formed by multiple computing devices Alternatively, they may be implemented in program code executable by a computing device so that they may be stored in a storage device to be executed by a computing device, and in some cases in an order different from that shown here The steps shown or described are carried out, or they are separately fabricated into individual integrated circuit modules, or multiple modules or steps among them are fabricated into a single integrated circuit module for implementation. As such, the present invention is not limited to any specific combination of hardware and software.

The above descriptions are only preferred embodiments of the present invention, and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and changes. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included within the protection scope of the present invention.

Claims (10)

1. A queue adjustment method, characterized in that, comprising: Comparing the relationship between the occupied length of the memory space of the current quality of service QOS out of the queue and the set first threshold and the second threshold, wherein the first threshold and the second threshold are smaller than the memory of the current QOS out of the queue a spatial length, the first threshold being smaller than the second threshold; If the occupied length of the memory space of the current QOS dequeue is greater than the second threshold, select a queue from other QOS dequeues whose memory space occupied length is less than the first threshold, wherein the current QOS Borrowing buffers are set in the outgoing queue and other QOS outgoing queues; Borrowing the borrow buffer of the selected QOS dequeue to the tail of the current QOS dequeue. 2. The method according to claim 1, wherein the length of the borrow buffer is: The length of the borrow buffer=QOS dequeue length-the second threshold+(the second threshold-the first threshold)×N%, where N is an integer between 0-100. 3. The method according to claim 1 or 2, further comprising: If the occupied length of the memory space of the current QOS dequeue is less than the first threshold, then set the current QOS dequeue as a borrowable queue, allowing other QOS dequeues to supply their borrow buffers. 4. The method according to claim 1 or 2, further comprising: If the occupied length of the memory space of the current QOS dequeue is less than or equal to the second threshold and greater than or equal to the first threshold, then the current QOS dequeues to determine whether its borrowed buffer is borrowed; if so, then Take back its borrowed buffer, and redistribute the borrowed buffer for the QOS that borrowed its borrowed buffer; if not, then set the current QOS out of the queue as a non-borrowable queue, and refuse to lend its borrowed buffer to the other QOS queues buffer. 5. The method according to claim 1, further comprising: When the occupied length of the memory space of the current QOS dequeue decreases from greater than the second threshold to less than or equal to the second threshold, the current QOS dequeues and returns the borrowed buffer. 6. The method according to claim 1, further comprising: The system saves the corresponding relationship between the current QOS out queue and the QOS out queue to which the borrowed buffer belongs, and updates it in real time. 7. A queue adjustment device, characterized in that it comprises: The comparison module is used to compare the relationship between the occupied length of the memory space of the current quality of service QOS dequeuing and the set first threshold and the second threshold, wherein the first threshold and the second threshold are both smaller than the current The memory space length of the QOS dequeuing, the first threshold is less than the second threshold; The first allocation module is configured to, when the comparison result of the comparison module is that the occupied length of the memory space of the current QOS dequeuing is greater than the second threshold, the occupied length of other memory spaces is less than the first threshold Select a queue in the QOS out of the queue, wherein, the current QOS out of the queue and other QOS out of the queue are all provided with a borrow buffer; the borrowed buffer of the selected QOS out of the queue is borrowed to the current QOS Dequeue the tail of the queue. 8. The device according to claim 7, wherein the length of the borrow buffer is: The length of the borrow buffer=QOS dequeue length-the second threshold+(the second threshold-the first threshold)×N%, where N is an integer between 0-100. 9. The device according to claim 7 or 8, characterized in that the device further comprises: The second allocation module is used to set the current QOS queue as a borrowable queue if the comparison result of the comparison module is that the occupied length of the memory space of the current QOS out of the queue is less than the first threshold, allowing the queue to be borrowed. Other QOS dequeues for their borrowed buffers; and / or, The third allocation module is configured to: if the comparison result of the comparison module is that the occupied length of the memory space of the current QOS queue is less than or equal to the second threshold and greater than or equal to the first threshold, the current QOS goes out queue and judges whether its borrowed buffer is borrowed; If so, then reclaims its borrowed buffer, and redistributes borrowed buffer for the QOS that borrows its borrowed buffer goes out queue; If not, then setting described current QOS goes out queue as A non-borrowable queue refuses to lend its borrow buffer to the other QOS queues. 10. The device according to claim 7, further comprising: Returning module, for when the occupied length of the memory space of the current QOS out of the queue is reduced from greater than the second threshold to less than or equal to the second threshold, then the current QOS out of the queue returns the borrowed buffer.