CN102263701B - Queue regulation method and device - Google Patents
Queue regulation method and device Download PDFInfo
- Publication number
- CN102263701B CN102263701B CN201110240078.7A CN201110240078A CN102263701B CN 102263701 B CN102263701 B CN 102263701B CN 201110240078 A CN201110240078 A CN 201110240078A CN 102263701 B CN102263701 B CN 102263701B
- Authority
- CN
- China
- Prior art keywords
- queue
- qos
- dequeues
- threshold
- borrow
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L47/00—Traffic control in data switching networks
- H04L47/10—Flow control; Congestion control
- H04L47/30—Flow control; Congestion control in combination with information about buffer occupancy at either end or at transit nodes
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L47/00—Traffic control in data switching networks
- H04L47/50—Queue scheduling
- H04L47/62—Queue scheduling characterised by scheduling criteria
- H04L47/6215—Individual queue per QOS, rate or priority
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L49/00—Packet switching elements
- H04L49/90—Buffering arrangements
- H04L49/9084—Reactions to storage capacity overflow
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The invention discloses a queue regulation method and a queue regulation device. The queue regulation method comprises the following steps of: comparing relationships between the occupied memory space length of a current quality of service (QoS) output queue and first and second preset threshold values, wherein the first and second threshold values are both less than the memory space length of the current QoS output queue, and the first threshold value is less than the second threshold value; if the occupied memory space length of the current QoS output queue is greater than the second threshold value, selecting one of the other QoS output queues with the occupied memory space lengths which are less than the first threshold value, wherein to-be-borrowed buffers are arranged in all the current QoS output queue and the other QoS output queues; and borrowing the to-be-borrowed buffers of the selected QoS output queue to the tail of the current QoS output queue. By the method and the device, the effects of fully utilizing overall QoS output queue resources and improving quality of service are achieved.
Description
Technical field
The present invention relates to the communications field, in particular to a kind of service quality (Quality Of Service, QOS)
The queue regulation method and device of dequeue.
Background technology
Service quality QOS refers to that permission user obtains expected clothes at aspects such as handling capacity, delay, shake, packet loss
A series of set of technologies of business level.
For QOS dequeues, port-qos dequeue length is mean allocation realizing, switch is that all of the port is each
It is assigned with the queue of 8 equal lengths.At present, QOS dequeues technology is, when a QOS dequeue is full, to be put into the queue
Message will be dropped, and now, other QOS dequeues are probably then idle.For example, when port 1 certain queue (for example
When queue is 7) full, the message for newly increasing in queue 7 will be dropped, and at this moment, other queues of the port are then likely to be empty
Not busy or message seldom, additionally, the QOS dequeues of other ports are also likely to be idle or message is little.
It can be seen that, in actual use, majority of case is that less queue can be expired, and more queue be it is idle or
Seldom, queue resource can not make full use of message.
Thus, exist in current QOS dequeue technical schemes causes packet loss when the queue of section ports because of queue full
When, other QOS dequeues be probably now it is idle or use it is little so that the queue on the whole of QOS dequeues
The problem fully not used by resource.
The content of the invention
Present invention is primarily targeted at a kind of queue regulation method of QOS dequeues and device are provided, at least to solve
The above-mentioned existing QOS dequeues on the whole problem fully not used by queue resource.
According to an aspect of the invention, there is provided a kind of queue regulation method, including：Compare current service quality Q OS
The relation of the occupied length of memory headroom of dequeue first threshold and Second Threshold with setting, wherein, first threshold and the
Two threshold values are respectively less than the memory headroom length of current QOS dequeues, and first threshold is less than Second Threshold；If current QOS dequeues
The occupied length of memory headroom be more than Second Threshold, then from the occupied length of other memory headrooms less than first threshold
A queue is selected in QOS dequeues, wherein, is provided with borrow buffering in current QOS dequeues and other QOS dequeues
Area；The borrow relief area of the QOS dequeues for selecting is borrowed into the tail of the queue of current QOS dequeues.
Preferably, the length of borrow relief area is：Length=QOS dequeues length-the Second Threshold of borrow relief area+
(Second Threshold-first threshold) × N%, wherein, N is the integer between 0-100.
Preferably, queue regulation method also includes：If the occupied length of memory headroom of current QOS dequeues is less than first
Threshold value, then arrange current QOS dequeues for borrowing queue, it is allowed to confesses which to other QOS dequeues and borrows relief area.
Preferably, queue regulation method also includes：If the occupied length of memory headroom of current QOS dequeues is less than or waits
In Second Threshold and more than or equal to first threshold, then currently QOS dequeues judge which borrows whether relief area lends；If so,
Then withdraw which and borrow relief area, and borrow relief area is redistributed for borrowing its QOS dequeue for borrowing relief area；If it is not, then
Current QOS dequeues are set for borrowing queue, refuse its borrow relief area to be lent to other QOS queues.
Preferably, queue regulation method also includes：When the occupied length of memory headroom of current QOS dequeues is by more than the
Two threshold values are reduced to less than or equal to Second Threshold, then current QOS dequeues give back the borrow relief area of borrow.
Preferably, queue regulation method also includes：System preserves the borrow relief area institute of current QOS dequeues and its borrow
The corresponding relation of the QOS dequeues of category, and real-time update.
According to a further aspect in the invention, there is provided a kind of regulator device, including：Comparison module, works as comparing
The relation of first threshold and Second Threshold of the occupied length of memory headroom of front service quality QOS dequeue with setting, wherein,
First threshold and Second Threshold are respectively less than the memory headroom length of current QOS dequeues, and first threshold is less than Second Threshold；First
Distribute module, for being more than second when the occupied length of memory headroom that the comparative result of comparison module is current QOS dequeues
During threshold value, a queue is selected from QOS dequeues of the occupied length of other memory headrooms less than first threshold, wherein,
Borrow relief area is provided with current QOS dequeues and other QOS dequeues；By the borrow buffering of the QOS dequeues for selecting
Area borrows the tail of the queue of current QOS dequeues.
Preferably, the length of borrow relief area is：Length=QOS dequeues length-the Second Threshold of borrow relief area+
(Second Threshold-first threshold) × N%, wherein, N is the integer between 0-100.
Preferably, regulator device also includes：Second distribute module, if the comparative result for comparison module is current
The occupied length of memory headroom of QOS dequeues is less than first threshold, then arrange current QOS dequeues for borrowing queue, permit
Perhaps which is confessed to other QOS dequeues borrow relief area；And/or, the 3rd distribute module, if for the comparative result of comparison module
The occupied length of memory headroom for current QOS dequeues is less than or equal to Second Threshold and is more than or equal to first threshold, then
Current QOS dequeues judge which borrows whether relief area lends；If so, then withdraw which and borrow relief area, and for borrowing its borrow
The QOS dequeues of relief area redistribute borrow relief area；If it is not, then arranging current QOS dequeues for queue can not be borrowed, refuse
Which is lent to other QOS queues absolutely and borrows relief area.
Preferably, regulator device also includes：Module is given back, it is occupied for the memory headroom when current QOS dequeues
Second Threshold is reduced to less than or equal to length by more than Second Threshold, then current QOS dequeues are given back the borrow of borrow and delayed
Rush area.
By the present invention, using a queue buffer that can be shared all is marked off to all QOS dequeues, when certain
When using more than Second Threshold of QOS dequeues, the team in QOS dequeues by other using queue length less than first threshold
Row relief area borrows and increases to the queue so that length of the length of the queue more than original original allocation, so that work as should
When message in QOS dequeues exceedes the QOS dequeue initial lengths, message can be still buffered in the QOS dequeues, and not
Can be dropped, so as to solve the problems, such as that queue resource is not fully used existing QOS dequeues on the whole, and then reach
Make full use of QOS dequeues entirety resource, the effect improved service quality.
Description of the drawings
Accompanying drawing described herein is used for providing a further understanding of the present invention, constitutes the part of the application, this
Bright schematic description and description does not constitute inappropriate limitation of the present invention for explaining the present invention.In the accompanying drawings：
The step of Fig. 1 is a kind of according to embodiments of the present invention one queue regulation method flow chart；
The step of Fig. 2 is a kind of according to embodiments of the present invention two queue regulation method flow chart；
Fig. 3 is the queue original allocation figure in a kind of according to embodiments of the present invention three queue regulation method；
Fig. 4 is the buffering queue management figure in a kind of according to embodiments of the present invention three queue regulation method；
Fig. 5 is that the queue in a kind of according to embodiments of the present invention three queue regulation method can borrow depth map；
Fig. 6 is the buffer queue recovery figure in a kind of according to embodiments of the present invention three queue regulation method；
Fig. 7 is the buffer queue borrow figure in a kind of according to embodiments of the present invention three queue regulation method；
Fig. 8 is that the buffer queue in a kind of according to embodiments of the present invention three queue regulation method gives back figure；
Fig. 9 is that the buffer queue in a kind of according to embodiments of the present invention three queue regulation method borrows management figure；
Figure 10 is a kind of structured flowchart of according to embodiments of the present invention four regulator device.
Specific embodiment
Below with reference to accompanying drawing and in conjunction with the embodiments describing the present invention in detail.It should be noted that not conflicting
In the case of, the feature in embodiment and embodiment in the application can be mutually combined.
Embodiment one
With reference to Fig. 1, flow chart the step of show a kind of according to embodiments of the present invention one queue regulation method.
The queue regulation method of the present embodiment is comprised the following steps：
Step S102：The occupied length of memory headroom of the current QOS dequeues of comparison and the first threshold and second of setting
The relation of threshold value；
Wherein, first threshold and Second Threshold are respectively less than the memory headroom length of current QOS dequeues, also, the first threshold
Value is less than Second Threshold.
In the present embodiment, borrow relief area is provided with all QOS dequeues of a port, when a QOS goes out team
Row it is idle or use it is little when, which borrows relief area can lend and use to other QOS dequeues；On the contrary when a QOS goes out
When queue is very full, then can borrow borrow relief area that other can borrow to increase queue length, extension queue space.
In the present embodiment, when queue footprint length be less than first threshold when, illustrate that the queue is more idle, also compared with
Many idle queues spaces can be used, and can lend to other queues；When queue takes up room length more than or equal to the first threshold
When being worth and being less than or equal to Second Threshold, illustrate that the queue is relatively done, it is impossible to which is lent to other queues and borrow relief area, and be possible to
The borrow relief area that can be borrowed is borrowed to other queues；When queue takes up room length more than Second Threshold, then the team is illustrated
Row are extremely busy, need to borrow corresponding borrow relief area to other queues that can lend relief area.
Step S104：If the occupied length of memory headroom of current QOS dequeues is more than Second Threshold, from other interior
The occupied length in space is deposited less than one queue of selection in the QOS dequeues of first threshold；
Wherein, borrow relief area is provided with current QOS dequeues and other QOS dequeues, the borrow relief area is
One shared buffer, can be lent to other QOS dequeues when necessary.
Step S106：The borrow relief area of the QOS dequeues for selecting is borrowed into the tail of the queue of current QOS dequeues.
The tail of the queue for arriving current QOS dequeues is borrowed by the relief area that borrows of the QOS dequeues that will be selected, can be extended and be worked as
The spatial cache of front QOS dequeues, in order to cache more messages.The mode of borrow can by those skilled in the art according to
Practical situation is flexibly arranged, and such as uses chained list mode, or using charting mode etc., as long as can determine the borrow of offer
The address of relief area, in order to stored messages data.
In correlation technique, exist in QOS dequeue technical schemes causes packet loss when the queue of section ports because of queue full
When, other QOS dequeues be probably now it is idle or use it is little so that the queue on the whole of QOS dequeues
Resource is not fully used.By the present embodiment, adopt and a queue that can be shared all is marked off to all QOS dequeues
Relief area, when certain QOS dequeue is when using more than Second Threshold, by other using queue length less than first threshold
Queue buffer in QOS dequeues is borrowed increases to the queue so that length of the length of the queue more than original original allocation
Degree, so that when the message in the QOS dequeues exceedes the QOS dequeue initial lengths, message can still be buffered in this
In QOS dequeues, without being dropped, so as to solve existing QOS dequeues, queue resource is not fully made on the whole
Problem, so reached make full use of QOS dequeues entirety resource, the effect improved service quality.
Embodiment two
With reference to Fig. 2, flow chart the step of show a kind of according to embodiments of the present invention two queue regulation method.
In the present embodiment, a queue buffer that can be shared all is marked off to all QOS dequeues, in certain queue
Man Shi, borrows other idle queues to increase the length of the queue, reduces packet loss quantity, there is provided preferably QOS services.
The queue regulation method of the present embodiment is comprised the following steps：
Step S202：For each QOS dequeue of each port, it is respectively provided with minimal buffering and gives back threshold values (i.e. first
Threshold value) and maximum cushioning borrow threshold values (i.e. Second Threshold).
Step S204：The queue of threshold values is given back not less than minimal buffering by the queue length used in all QOS dequeues,
It is managed collectively by software；For the queue for having borrowed, record is borrowed and the relation being borrowed, and is managed collectively by software system.
Wherein it is possible to the queue length used in all QOS dequeues not less than minimal buffering by the way of chained list
The queue for giving back threshold values is managed collectively, it would however also be possible to employ the mode of form registration is managed collectively, certainly, this area skill
Art personnel can also adopt other appropriate ways, the invention is not limited in this regard.By to the team used in all QOS dequeues
Row length gives back the unified management of the queue of threshold values not less than minimal buffering, can have QOS dequeues to need to borrow other teams
During the spatial cache of row, conveniently and efficiently obtain and provide and can lend the queue of spatial cache, also, also allow for understanding in real time and
Acquisition can lend the situation of the queue of spatial cache, improve system effectiveness.
For the queue for having borrowed, between the queue for borrowing spatial cache and the queue for being borrowed spatial cache, have one
One-to-one correspondence, records the relation in time, the situation that on the one hand can be borrowed and be borrowed in understanding queue in time, the opposing party
Face, when the queue for borrowing spatial cache no longer needs to borrow, timely and effectively can give back the borrow relief area that is borrowed to
Corresponding queue.
It can be to borrow the residual queue of threshold values more than maximum cushioning that each QOS dequeue is available for the length for borrowing.It is preferred that
Ground, each QOS dequeue are available for the length for borrowing be made up of two parts, and a part is greater than maximum cushioning and borrows threshold values
Residual queue, another part are that the difference that maximum cushioning borrow threshold values gives back threshold values with minimal buffering is multiplied by the percentage that can be configured
The space length of ratio.
Step S206：According to the occupied length of memory headroom of current QOS dequeues, carry out at corresponding regulator
Reason.
Specifically,
When the minimal buffering that uses more than of queue gives back threshold values, it is to borrow by the borrow setting buffers of the queue.
If the borrow relief area of the queue has been lent, need to withdraw, the borrow that can redistribute a queue from system is delayed
Rush area to use to borrower.
Threshold values is given back when the use of queue from minimal buffering to be changed into borrowing threshold values more than maximum cushioning, from free buffer queue
A borrow is found out in list, increases the length of the queue.The lookup of free buffer queue is preferentially from borrow queue place end
Other queues of mouth, if other queues can not be lent or lent, then from the queue borrow of other ports.
When the use of queue becomes smaller than maximum cushioning borrow threshold values by more than maximum cushioning borrow threshold values, borrow is given back
Borrow relief area.
When the use of queue gives back threshold values by giving back threshold values more than minimal buffering and become smaller than minimal buffering, by the queue
Setting buffers are borrowed for borrowing.
The scheme with dynamic adjustment queue size provided by the present embodiment, is gone out for each QOS of each port
Queue, is respectively provided with minimal buffering and gives back threshold values (also referred to herein simply as minimum threshold values) and maximum cushioning borrow threshold values (this paper
In also referred to as maximum threshold values), and by between the part queue beyond maximum threshold values and minimum threshold values and maximum threshold values
Partition queue element is set to the queue (being also referred to as buffer queue herein) that can be borrowed jointly, when the use queue length of port
More than minimum threshold values, then report software indicate the queue can not lend buffer queue (acquiescence all of the port all can be lend it is slow
Rush queue).When queue length exceedes maximum threshold values, software can detect that the buffer queue of also which port can be borrowed,
Idle buffer queue is borrowed in busy queue after finding, increase the length of the queue, reduce idle queues length, one
The beginning of denier idle queues is busy, i.e., exceed minimum threshold values using queue length, and the buffering of borrow will be given back.For
Prevent buffering concussion, each queue from can set a high-order maximum threshold values, the minimum threshold values of a low level, using team
When row length increases to maximum threshold values by minimum threshold values, triggering is high-order, is reduced to minimum valve using queue length by maximum threshold values
During value, low level is triggered.
Compared with prior art, part resource of the present embodiment by shared queue, dynamic adjustment queue size, makes team
The row utilization of resources is more abundant.When certain queue is when using more than maximum threshold values, other are less than into minimum valve using queue length
The buffer queue of value increases to the queue so that length of the length of the queue more than original original allocation, when the report in queue
When text exceedes queue initial length, message can be still cached in the queue, without being dropped, so as to improve Service Quality
Amount.
Embodiment three
In the present embodiment, by taking the regulator that the queue of a QOS dequeue is used in dynamic changing process as an example, in detail
Illustrate QOS queue regulation methods of the present invention.
First, with reference to Fig. 3, show that the queue in a kind of according to embodiments of the present invention three queue regulation method initially divides
Figure.As shown in figure 3, each QOS dequeue is divided into three parts by the present embodiment：Less than minimum threshold values (i.e. first threshold)
Partly between (nicking part in figure), minimum threshold values and maximum threshold values (i.e. Second Threshold) partly (inclined stripe part in figure) and
More than maximal valve value part (residual queue in figure behind inclined stripe part, i.e. cross stripe portion).
Wherein, when the occupied length of the memory headroom of queue is located at less than minimum valve value part, i.e. when the use of queue
During less than minimum threshold values, queue is set to borrow queue, it is allowed to lend its caching to other QOS dequeues；When queue
When the occupied length of memory headroom is located at part between minimum threshold values and maximum threshold values, i.e. when the use of queue is between minimum valve
When between value and maximum threshold values, the queue is set to borrow queue, now, refuses to lend which to other QOS dequeues and delays
Deposit；When the occupied length of the memory headroom of queue is located at more than maximal valve value part, i.e. when the use of queue is more than maximal valve
After value, illustrate that the use of the queue soon reaches the maximum of queue, it is therefore desirable to borrow from other queue, increase the queue
Operational length.
With reference to Fig. 4, the buffering queue management figure in a kind of according to embodiments of the present invention three queue regulation method is shown.
As shown in figure 4, the present embodiment is in the system starting stage, by all idle queues, (i.e. queuing memory footprint length is less than most
The queue of little threshold values) all manage, it is linked to be a chained list for being available for borrowing, the chained list only records the queue using less than minimum
The queue of threshold values.Meanwhile, by all borrow queues and it is borrowed queue and is linked to be a chained list, the borrow situation of record chained list and quilt
Borrow situation, gives back after being easy to borrow.
With reference to Fig. 5, show that the queue in a kind of according to embodiments of the present invention three queue regulation method can borrow length
Figure.As shown in figure 5, the present embodiment design is available for the calculating of the queue length for borrowing, acquisition to be available for the queue for borrowing.Queue is available for
The length of borrow is made up of two parts, and a part is greater than the residual queue of maximum threshold values, another part be maximum threshold values with most
Little threshold difference value is multiplied by the percentage ratio that can be configured, and the percentage ratio can be configured, and is adjusted according to actually used situation.It is available for
The position of the length of borrow is distributed from the afterbody of queue forward, the horizontal stripe part such as in figure.
Specifically, queue length=queue length-maximum threshold values+(maximum threshold values-minimum threshold values) * N% can be borrowed, its
In, the size of N is the integer between 0-100, can flexible configuration according to actual needs.
With reference to Fig. 6, the buffer queue recovery figure in a kind of according to embodiments of the present invention three queue regulation method is shown.
As shown in fig. 6, when the memory headroom that queue (such as the queue 7 of port 1) is used is increased to by minimum threshold values is less than
During more than minimum threshold values, the queue is set to borrow queue by needs.If the queue is not borrowed by other queues, i.e.,
The queue is taken out from the queue for being available for borrowing in the queue for being available for borrowing by the queue, i.e. the queue can not be borrowed again；
If the queue is borrowed by other queues, the queue that the queue is lent is withdrawn by force, is deleted from borrowing in queue table
Remove, meanwhile, then redistribute one and can borrow queue from can borrow in queue for borrowing the queue of the queue.
Specifically, including：
Step S602：From borrowing in queue, the queue is searched, if it is found, illustrate that the queue is not borrowed, should
Queue from can borrow in queue delete, the step of no longer execution below；If do not found, execution step S604.
Step S604：The queue has been borrowed, and searches the queue from being borrowed in relation table, finds and borrows the queue
Borrower, withdraws the borrow queue from borrower.
Step S606：Again borrow in queue a queue to the borrower for borrowing this queue from can borrow, if can borrow
It is empty with queue, borrower will be borrowed less than queue.
With reference to Fig. 7, the buffer queue borrow figure in a kind of according to embodiments of the present invention three queue regulation method is shown.
As shown in fig. 7, the memory headroom used when queue (such as the queue 7 of figure middle port 1) by less than maximum threshold values to surpassing
After crossing maximum threshold values, the use of the queue needs to borrow other queues to increase the big of queue soon more than the maximum of queue
It is little, a queue (queue 6 of the port 1 in such as figure) for being available for borrowing is searched in queue from borrowing, and the queue of borrow is put
Enter the afterbody of current queue, increase the length of queue, while reducing the length for borrowing queue.
Specifically, including：
Step S702：A queue (queue 6 of the port 1 in such as figure) for being available for borrowing is found out in queue from borrowing,
If can not find be available for borrow queue, do not do any operation, i.e. below step without carrying out；If find being available for what is borrowed
Queue, carries out step S704.
Step S704：Queue (such as the queue 6 of the port 1 in figure) will be borrowed from being available for borrowing deletion in queue, which will be prevented
Its queue is borrowed.
Certainly, in practical operation, it is also possible to otherwise, the mode that such as the borrow queue is marked, with
Distinguish whether which is lent.
Step S706：Queue (such as the queue 7 of the port 1 in figure) will be borrowed and queue is borrowed (such as the port 1 in figure
6) queue is added to borrow queue relation table.
Step S708：Queue (such as the queue 6 of figure middle port 1) will be borrowed increases to borrow queue (such as figure middle port 1
Queue afterbody 7), increases the length of queue.
With reference to Fig. 8, show that the buffer queue in a kind of according to embodiments of the present invention three queue regulation method gives back figure.
As shown in figure 8, when the memory headroom that queue (such as the queue 7 of figure middle port 1) is used is reduced by maximum threshold values is more than
To after less than maximum threshold values, the length of the queue is used enough, and needs give back other queues for having borrowed, is updated queue borrow and is closed
It is table.
Specifically, including：
Step S802：Check that queue borrows relation table, search whether that the queue borrows the record of other queues, if
Have, execution step S804；If it did not, do not do any operation, i.e. below step without carrying out.
Step S804：To be borrowed queue (such as the queue 6 of the port 1 in figure) from borrow queue (port 1 in such as figure
Queue afterbody 7) removes, and returns to and is borrowed queue.
Step S806：To borrow in queue relation table, the borrow relation for borrowing queue and being borrowed queue is deleted.
Step S808：To be borrowed queue (such as the queue 6 of figure middle port 1) increase to be available for borrow queue linked list in, need
Other queues of queue to be borrowed can borrow the queue.
With reference to Fig. 9, show that the buffer queue in a kind of according to embodiments of the present invention three queue regulation method borrows pipe
Reason figure.
As shown in figure 9, when the memory headroom that queue (such as the queue 7 of figure middle port 1) is used is reduced by minimum threshold values is more than
To after less than minimum threshold values, the use of the queue is more idle, may be configured as borrowing queue, increases to the chain that can borrow queue
In table.
By the present embodiment, using the part resource for sharing each QOS dequeue, dynamic adjustment queue size, queue is made
The utilization of resources is more abundant, improves service quality.
Example IV
With reference to Figure 10, a kind of structured flowchart of according to embodiments of the present invention four regulator device is shown.
The regulator device of the present embodiment includes：Comparison module 1002, the internal memory for the current QOS dequeues of comparison are empty
Between first threshold and Second Threshold of the occupied length with setting relation, wherein, first threshold and Second Threshold are respectively less than worked as
The memory headroom length of front QOS dequeues, first threshold are less than Second Threshold；First distribute module 1004, compares mould for working as
When the comparative result of block 1002 is that the occupied length of memory headroom of current QOS dequeues is more than Second Threshold, from other interior
Deposit the occupied length in space less than first threshold QOS dequeues in select a queue, wherein, current QOS dequeues and its
Borrow relief area is provided with its QOS dequeue；The borrow relief area of the QOS dequeues of selection is borrowed to go out to current QOS
The tail of the queue of queue.
Preferably, the length of borrow relief area is：Borrow length=QOS dequeue length-second threshold of relief area
Value+(Second Threshold-first threshold) × N%, wherein, N is the integer between 0-100.
Preferably, the regulator device of the present embodiment also includes：Second distribute module 1006, if for comparison module
1002 comparative result is that the occupied length of memory headroom of current QOS dequeues is less than first threshold, then arrange current QOS and go out
Queue is to borrow queue, it is allowed to confesses which to other QOS dequeues and borrows relief area；And/or, the 3rd distribute module 1008,
If for comparison module 1002 comparative result be current QOS dequeues the occupied length of memory headroom be less than or equal to second
Threshold value and it is more than or equal to first threshold, then current QOS dequeues judge which borrows whether relief area lends；If so, then withdraw
Which borrows relief area, and redistributes borrow relief area for borrowing its QOS dequeue for borrowing relief area；If it is not, then arrange working as
Front QOS dequeues are refused to lend its borrow relief area to other QOS queues for borrowing queue.
Preferably, the regulator device of the present embodiment also includes：Module is given back, for when in current QOS dequeues
Deposit the occupied length in space and Second Threshold is reduced to less than or equal to by more than Second Threshold, then current QOS dequeues are given back and borrowed
Borrow relief area.
Preferably, the regulator device of the present embodiment also includes：Preserving module, preserves current QOS for system and goes out team
The corresponding relation for borrowing the QOS dequeues belonging to relief area of row and its borrow, and real-time update.
The method that the present embodiment is realized is referred to the associated description of aforementioned correlation method embodiment, and has above-mentioned enforcement
The beneficial effect of example, will not be described here.
As can be seen from the above description, queue of the QOS dequeues that the present invention is provided on distribution switch is adjusted
Adjusting technique scheme, is dynamically adjusted according to the queue service condition of QOS dequeues, by the part resource of shared queue,
Queue resource is made using more abundant.It is especially when certain queue is when using more than maximum threshold values, other are little using queue length
Increase to the queue in the buffer queue of minimum threshold values so that the length of the queue works as team more than the length of original original allocation
When message in row exceedes queue initial length, message can be still cached in the queue, without being dropped, so as to improve
Service quality.
Obviously, those skilled in the art should be understood that each module or each step of the above-mentioned present invention can be with general
Computing device realizing, they can be concentrated on single computing device, or are distributed in multiple computing devices and are constituted
Network on, alternatively, they can be realized with the executable program code of computing device, it is thus possible to they are stored
In the storage device by computing device performing, and in some cases, can be with shown in performing different from order herein
The step of going out or describe, or they are fabricated to each integrated circuit modules respectively, or by the multiple modules in them or
Step is fabricated to single integrated circuit module to realize.So, the present invention is not restricted to any specific hardware and software combination.
The preferred embodiments of the present invention are the foregoing is only, the present invention is not limited to, for the skill of this area
For art personnel, the present invention can have various modifications and variations.It is all within the spirit and principles in the present invention, made any repair
Change, equivalent, improvement etc., should be included within the scope of the present invention.
Claims (10)
1. a kind of queue regulation method, it is characterised in that include：
The occupied length of memory headroom for comparing current service quality Q OS dequeue and the first threshold and Second Threshold for setting
Relation, wherein, the memory headroom length of the first threshold and the Second Threshold respectively less than current QOS dequeues, institute
First threshold is stated less than the Second Threshold；
It is if the occupied length of memory headroom of the current QOS dequeues is more than the Second Threshold, empty from other internal memories
Between occupied length less than the first threshold QOS dequeues in select a queue, wherein, the current QOS dequeues
Relief area is borrowed with being provided with other QOS dequeues；
The borrow relief area of the QOS dequeues for selecting is borrowed into the tail of the queue of the current QOS dequeues.
2. method according to claim 1, it is characterised in that the length of the borrow relief area is：
The length of the borrow relief area=QOS dequeue length-Second Threshold+(Second Threshold-first threshold
Value) × N%, wherein, N is the integer between 0-100.
3. method according to claim 1 and 2, it is characterised in that also include：
If the occupied length of memory headroom of the current QOS dequeues is less than the first threshold, the current QOS is set
Dequeue is to borrow queue, it is allowed to confesses which to other QOS dequeues and borrows relief area.
4. method according to claim 1 and 2, it is characterised in that also include：
If the occupied length of memory headroom of the current QOS dequeues is less than or equal to the Second Threshold and is more than or equal to
The first threshold, then the current QOS dequeues judge its borrow relief area whether lend；If so, then withdraw which and borrow slow
Area is rushed, and borrow relief area is redistributed for borrowing its QOS dequeue for borrowing relief area；If it is not, then arranging the current QOS
Dequeue is refused to lend its borrow relief area to described other QOS queues for borrowing queue.
5. method according to claim 1, it is characterised in that also include：
When the occupied length of memory headroom of the current QOS dequeues is by being reduced to less than or equal to more than the Second Threshold
The Second Threshold, then the current QOS dequeues give back the borrow relief area of borrow.
6. method according to claim 1, it is characterised in that also include：
System preserves the correspondence pass of the QOS dequeues belonging to the borrow relief area of the current QOS dequeues and its borrow
System, and real-time update.
7. a kind of regulator device, it is characterised in that include：
Comparison module, for comparing the first threshold of the occupied length of memory headroom of current service quality Q OS dequeue and setting
The relation of value and Second Threshold, wherein, the first threshold and the Second Threshold are respectively less than the interior of the current QOS dequeues
Space length is deposited, the first threshold is less than the Second Threshold；
First distribute module, for when the memory headroom quilt that the comparative result of the comparison module is the current QOS dequeues
When taking length more than the Second Threshold, go out from QOS of the occupied length of other memory headrooms less than the first threshold
A queue is selected in queue, wherein, is provided with borrow relief area in the current QOS dequeues and other QOS dequeues；
The borrow relief area of the QOS dequeues for selecting is borrowed into the tail of the queue of the current QOS dequeues.
8. device according to claim 7, it is characterised in that the length of the borrow relief area is：
The length of the borrow relief area=QOS dequeue length-Second Threshold+(Second Threshold-first threshold
Value) × N%, wherein, N is the integer between 0-100.
9. the device according to claim 7 or 8, it is characterised in that described device also includes：
Second distribute module, if for the comparison module comparative result be the current QOS dequeues memory headroom quilt
Take length and be less than the first threshold, then the current QOS dequeues are set for borrowing queue, it is allowed to go out to other QOS
Queue is confessed which and borrows relief area；
And/or,
3rd distribute module, if for the comparison module comparative result be the current QOS dequeues memory headroom quilt
Take length to be less than or equal to the Second Threshold and be more than or equal to the first threshold, then the current QOS dequeues are sentenced
Breaking, which borrows whether relief area lends；If so, then withdraw which and borrow relief area, and go out team for borrowing its QOS for borrowing relief area
Row redistribute borrow relief area；If it is not, then arrange the current QOS dequeues for queue can not be borrowed, refuse to it is described its
Its QOS queue is lent which and borrows relief area.
10. device according to claim 7, it is characterised in that described device also includes：
Module is given back, for the occupied length of memory headroom when the current QOS dequeues by subtracting more than the Second Threshold
It is little to less than or equal to the Second Threshold, then the current QOS dequeues give back the borrow relief area of borrow.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201110240078.7A CN102263701B (en) | 2011-08-19 | 2011-08-19 | Queue regulation method and device |
| PCT/CN2012/077762 WO2013026324A1 (en) | 2011-08-19 | 2012-06-28 | Queue adjustment method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201110240078.7A CN102263701B (en) | 2011-08-19 | 2011-08-19 | Queue regulation method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102263701A CN102263701A (en) | 2011-11-30 |
| CN102263701B true CN102263701B (en) | 2017-03-22 |
Family
ID=45010167
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201110240078.7A Active CN102263701B (en) | 2011-08-19 | 2011-08-19 | Queue regulation method and device |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN102263701B (en) |
| WO (1) | WO2013026324A1 (en) |
Families Citing this family (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102263701B (en) * | 2011-08-19 | 2017-03-22 | 中兴通讯股份有限公司 | Queue regulation method and device |
| CN103248583B (en) * | 2012-02-09 | 2016-04-13 | 迈普通信技术股份有限公司 | A kind of first in first out message queue management method based on finite depth and device |
| CN102984083B (en) * | 2012-11-19 | 2018-07-24 | 南京中兴新软件有限责任公司 | queue management method and device |
| CN104022971B (en) * | 2014-05-26 | 2018-04-06 | 上海斐讯数据通信技术有限公司 | Optimal change machine message up sending method and its system |
| CN107346265B (en) * | 2016-05-05 | 2021-05-18 | 北京忆恒创源科技有限公司 | Method and device for realizing QoS |
| CN106970827A (en) * | 2017-03-27 | 2017-07-21 | 联想(北京)有限公司 | Information processing method, information processor, electronic equipment |
| CN109947798A (en) * | 2017-09-18 | 2019-06-28 | 中国移动通信有限公司研究院 | A kind of processing method and processing device of stream event |
| CN108040019B (en) * | 2017-12-21 | 2021-12-24 | 杭州迪普科技股份有限公司 | Message forwarding method and device |
| CN109495401B (en) * | 2018-12-13 | 2022-06-24 | 迈普通信技术股份有限公司 | Cache management method and device |
| CN114995748A (en) * | 2022-05-09 | 2022-09-02 | 阿里巴巴（中国）有限公司 | Request processing method and device |
Family Cites Families (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN100359489C (en) * | 2004-07-13 | 2008-01-02 | 中兴通讯股份有限公司 | Method for internal memory allocation in the embedded real-time operation system |
| CN101094181B (en) * | 2007-07-25 | 2010-12-08 | 华为技术有限公司 | Dispatch device and method of enqueuing and dequeuing message |
| CN101141399A (en) * | 2007-10-16 | 2008-03-12 | 武汉烽火网络有限责任公司 | Dual-threshold port queue control method for sharing storage exchange supporting multicast |
| CN101521624B (en) * | 2008-02-26 | 2012-01-11 | 大唐移动通信设备有限公司 | Method and device for distributing downlink bandwidth of lub port |
| US7864676B2 (en) * | 2008-07-14 | 2011-01-04 | The Mitre Corporation | Network cross-domain precedence and service quality conflict mitigation |
| CN101800699A (en) * | 2010-02-09 | 2010-08-11 | 上海华为技术有限公司 | Method and device for dropping packets |
| CN102263701B (en) * | 2011-08-19 | 2017-03-22 | 中兴通讯股份有限公司 | Queue regulation method and device |
-
2011
- 2011-08-19 CN CN201110240078.7A patent/CN102263701B/en active Active
-
2012
- 2012-06-28 WO PCT/CN2012/077762 patent/WO2013026324A1/en active Application Filing
Also Published As
| Publication number | Publication date |
|---|---|
| CN102263701A (en) | 2011-11-30 |
| WO2013026324A1 (en) | 2013-02-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102263701B (en) | Queue regulation method and device | |
| Kumar et al. | Peacock hashing: Deterministic and updatable hashing for high performance networking | |
| EP3334123A1 (en) | Content distribution method, virtual server management method, cloud platform and system | |
| US20160132541A1 (en) | Efficient implementations for mapreduce systems | |
| CN104731799B (en) | Main memory DBM device | |
| CN110336891A (en) | Data cached location mode, equipment, storage medium and device | |
| CN102971732A (en) | System architecture for integrated hierarchical query processing for key/value stores | |
| CN107704401A (en) | Data cached method of replacing, system and storage system in a kind of storage system | |
| WO2005086435A1 (en) | A method for congestion management of a network, a signalling protocol, a switch, an end station and a network | |
| CN102279810A (en) | Network storage server and method for caching data | |
| CN101499956B (en) | Hierarchical buffer zone management system and method | |
| US20130322271A1 (en) | System for performing data cut-through | |
| CN101005461A (en) | IPv6 route list checking and repeating method | |
| CN102185725A (en) | Cache management method and device as well as network switching equipment | |
| CN104572498B (en) | The buffer memory management method and device of message | |
| CN106940696B (en) | Information query method and system for SDN multi-layer controller | |
| CN105791455A (en) | Ternary content addressable memory TCAM space processing method and device | |
| CN105243031B (en) | A kind of method and device of cache partitions distribution free page | |
| CN109002260A (en) | A kind of data cached processing method and processing system | |
| CN103226520A (en) | Self-adaptive cluster memory management method and server clustering system | |
| Escudero-Sahuquillo et al. | Cost-effective congestion management for interconnection networks using distributed deterministic routing | |
| CN106228074A (en) | storage object aging method and device | |
| CN106294191A (en) | The method processing table, the method and apparatus accessing table | |
| CN104935517B (en) | Entry management method and device | |
| CN1812377A (en) | Management method for two-stage route table in data communication equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |