CN103685062A - Cache management method and device - Google Patents

Abstract

An embodiment of the invention provides a cache management method and device. The cache management method includes: judging whether a stream queue with queue length exceeding a queue cache threshold exists in first stream queues or not; if so, executing dequeuing operation to the stream queue exceeding the queue cache threshold. By the method and device, utilization rate of caches can be increased.

Description

Cache management method and device

technical field

The embodiments of the present invention relate to network technologies, and in particular to a buffer management method and device.

Background technique

With the rapid growth of multimedia data flow requirements, the buffer management of the switch becomes extremely important. Shared cache is a commonly used cache structure in switches, and all input and output ports share a cache space. Commonly used shared cache management methods in the prior art mainly include a static threshold method and a dynamic threshold method.

The static threshold method allocates a certain cache for each queue and/or port, and other caches or ports cannot be preempted. Although the fairness of cache sharing can be guaranteed, the utilization rate of the cache is low. The dynamic threshold method configures a certain buffer occupation ratio factor for each queue and/or port, and dynamically allocates available buffers for each queue according to the remaining shared buffers during the application process. Due to the complex configuration process of this method, the improved cache utilization is limited to a certain extent.

In the prior art, the cache utilization rate of the cache management method is low.

Contents of the invention

Embodiments of the present invention provide a cache management method and device to solve the problem of low cache utilization in the prior art.

In a first aspect, an embodiment of the present invention provides a cache management method, including:

Judging whether there is a flow queue whose queue length exceeds the queue buffer threshold in the first flow queue;

If there is a flow queue whose queue length exceeds the queue buffer threshold in the first flow queue, perform a dequeue operation on the flow queue whose queue length exceeds the queue buffer threshold.

According to the first aspect, in the first possible implementation manner of the first aspect, before the judging whether the queue length of the first flow queue exceeds the queue cache threshold, further includes:

Determine whether the current system has a virtual port whose port length exceeds the port cache threshold;

If the current system does not have a virtual port whose port length exceeds the port cache threshold, determine the flow queue owned by the current system as the first flow queue.

According to the first possible implementation manner of the first aspect, in the second possible implementation manner, the queue cache threshold includes at least two cache thresholds with different priorities;

The judging whether there is a flow queue whose queue length exceeds the queue buffer threshold in the first flow queue includes:

Comparing the queue length of the flow queue in the first flow queue with the highest priority cache threshold in the queue cache threshold, and determining whether there is a flow queue in the first flow queue whose queue length exceeds the highest priority cache threshold;

If there is a flow queue whose queue length exceeds the queue cache threshold in the first flow queue, performing a dequeue operation on the flow queue that exceeds the queue cache threshold includes:

If there is a flow queue whose queue length exceeds the highest-priority cache threshold in the first flow queue, perform a dequeue operation on the flow queue that exceeds the highest-priority cache threshold.

According to the second possible implementation manner of the first aspect, in the third possible implementation manner, the method further includes:

If there is no flow queue whose queue length exceeds the cache threshold of the highest priority in the first flow queue, determine the virtual port with the smallest difference between the port cache threshold and the port length in the current system, and store the port cache threshold The flow queue of the virtual port having the smallest difference with the port length is determined as the second flow queue;

According to the priority order of each cache threshold in the queue cache threshold, sequentially determine whether there is a flow queue in the second flow queue whose queue length exceeds the cache threshold in the queue cache threshold;

If there is a flow queue whose queue length exceeds a cache threshold in the second flow queue, a dequeue operation is performed on the flow queue that exceeds the cache threshold.

According to the third possible implementation manner of the first aspect, in the fourth possible implementation manner, the method further includes:

If there is no flow queue whose queue length exceeds the cache threshold with the lowest priority among the queue cache thresholds in the second flow queue, then send first warning indication information; the first warning indication information includes configuration error prompt information.

According to the second possible implementation of the first aspect, in the fifth possible implementation, if the current system has a virtual port whose port length exceeds the port cache threshold, determine the port length and port cache threshold in the current system. the virtual port with the largest threshold difference, and determine the flow queue of the virtual port with the largest difference between the port length and the port cache threshold as the first flow queue.

According to the fifth possible implementation manner of the first aspect, in the sixth possible implementation manner, the judging whether there is a flow queue whose queue length exceeds the queue buffer threshold in the first flow queue further includes:

If there is no flow queue in the first flow queue whose queue length exceeds the cache threshold of the highest priority, according to the priority order of each cache threshold in the queue cache threshold, sequentially determine whether there is a queue in the first flow queue a flow queue whose length exceeds a cache threshold in said queue cache threshold;

If there is a flow queue whose queue length exceeds the queue cache threshold in the first flow queue, performing a dequeue operation on the flow queue that exceeds the queue cache threshold further includes:

If there is a flow queue whose queue length exceeds a cache threshold in the first flow queue, a dequeue operation is performed on the flow queue that exceeds the cache threshold.

According to the sixth possible implementation manner of the first aspect, in the seventh possible implementation manner, the method further includes:

If there is no flow queue in the first flow queue whose queue length exceeds the cache threshold with the lowest priority among the queue cache thresholds, send a second warning message; the second warning message includes configuration error message.

According to the fourth or seventh possible implementation manner of the first aspect, in the eighth possible implementation manner, before the judging whether there is a flow queue whose queue length exceeds the queue buffer threshold in the first flow queue, further includes :

Set the queue cache threshold according to the priority of the first flow queue.

According to the first possible implementation manner of the first aspect, in the ninth possible implementation manner, the sum of the port cache thresholds of the flow queues in the current system is less than or equal to N times the shared cache; the current system The total cache includes the shared cache;

The N is the ratio of the sum of the port cache thresholds of the flow queues in the current system to the actual cache of the current system.

According to the ninth possible implementation manner of the first aspect, in the tenth possible implementation manner, before the judging whether there is a virtual port whose port length exceeds the port cache threshold in the current system, it further includes:

Judging whether the free cache of the current system is smaller than the threshold of the protection cache; the total cache of the current system also includes the protection cache;

When the free cache of the current system is smaller than the threshold of the protection cache, an operation of judging whether the current system has a virtual port whose port length exceeds the threshold of the port cache is performed.

According to the tenth possible implementation manner of the first aspect, in the eleventh possible implementation manner, before the judging whether the free cache of the current system is smaller than the protection cache threshold, the method further includes:

The protection cache is configured according to the time delay of processing information in the current system.

According to the eighth possible implementation manner of the first aspect, in the twelfth possible implementation manner, the performing the dequeue operation on the flow queue exceeding the queue cache threshold includes:

selecting a flow queue from the flow queues exceeding the queue cache threshold;

The information corresponding to the selected flow queue is delivered to the queue manager, so that the queue manager performs a dequeue operation on the information of one stored packet in the information corresponding to the selected flow queue.

According to the twelfth possible implementation manner of the first aspect, in the thirteenth possible implementation manner, the method further includes:

According to the aging indication message corresponding to the selected flow queue, determine whether the selected flow queue is a flow queue that needs to be aged;

Perform an aging operation on the flow queue that needs to be aged.

According to the twelfth possible implementation manner of the first aspect, in the fourteenth possible implementation manner, after transmitting the information corresponding to the selected flow queue to the queue manager, the method further includes:

According to the queue manager, the dequeue operation is performed on the information of a stored packet in the information corresponding to the selected flow queue, and then the dequeue instruction message generated will set the dequeue counter corresponding to the selected flow queue value plus 1.

According to the fourteenth possible implementation manner of the first aspect, in the fifteenth possible implementation manner, according to the aging indication message corresponding to the selected flow queue, it is judged whether the selected flow queue is needed Aged flow queues, including:

If the value of the dequeue counter is greater than the dequeue times threshold, configure the aging indication message corresponding to the selected flow queue as 1; if the value of the dequeue counter is less than or equal to the dequeue times threshold, set the The aging indication message corresponding to the selected flow queue is configured as 0;

Within a predetermined period, if the aging indication message corresponding to the selected flow queue continues to be 1, it is determined that the selected flow queue is the flow queue that needs to be aged.

According to the fifteenth possible implementation manner of the first aspect, in the sixteenth possible implementation manner, after determining that the selected flow queue is the flow queue that needs to be aged, the method further includes:

reducing the configuration of the queue cache threshold corresponding to the selected flow queue;

Within the predetermined period, if the aging indication message corresponding to the selected flow queue continues to be 0, restore the configuration of the queue cache threshold corresponding to the selected flow queue.

In the second aspect, the embodiment of the present invention also provides a cache management device, including:

A judging module, configured to judge whether there is a flow queue whose queue length exceeds the queue buffer threshold in the first flow queue;

An executing module, configured to, if there is a flow queue whose queue length exceeds the queue buffer threshold in the first flow queue, perform a dequeue operation on the flow queue whose queue length exceeds the queue buffer threshold.

According to the second aspect, in the first possible implementation manner of the second aspect, the judging module is further configured to judge whether there is a port in the current system before judging whether the queue length of the first flow queue exceeds the queue cache threshold Virtual ports whose length exceeds the port cache threshold;

The device also includes:

A determining module, configured to determine the flow queue of the current system as the first flow queue if the current system does not have a virtual port whose port length exceeds a port cache threshold.

According to the first possible implementation manner of the second aspect, in the second possible implementation manner, the queue cache threshold includes at least two cache thresholds with different priorities;

The judging module is configured to compare the queue length of the flow queue in the first flow queue with the cache threshold of the highest priority in the queue cache threshold, and judge whether there is a queue length exceeding the highest priority in the first flow queue flow queue for cache threshold;

The execution module is configured to, if there is a flow queue whose queue length exceeds the highest-priority cache threshold in the first flow queue, perform a dequeue operation on the flow queue that exceeds the highest-priority cache threshold.

According to the second possible implementation manner of the second aspect, in the third possible implementation manner, the determining module is further configured to: if there is no buffer whose queue length exceeds the highest priority in the first flow queue Threshold flow queue, determine the virtual port with the smallest difference between the port cache threshold and port length in the current system, and determine the flow queue of the virtual port with the smallest difference between the port cache threshold and port length as the second flow queue;

The judging module is further configured to sequentially judge whether there is a flow queue whose queue length exceeds the cache threshold in the queue cache threshold in the second flow queue according to the priority order of each cache threshold in the queue cache threshold;

The execution module is further configured to, if there is a flow queue whose queue length exceeds a cache threshold in the second flow queue, perform a dequeue operation on the flow queue that exceeds the cache threshold.

According to the third possible implementation manner of the second aspect, in the fourth possible implementation manner, the device further includes:

The first sending module is configured to send first alarm prompt information if there is no flow queue whose queue length exceeds the lowest priority cache threshold among the queue cache thresholds in the second flow queue; the first alarm indication information Contains information about configuration errors.

According to the second possible implementation manner of the second aspect, in the fifth possible implementation manner, the determining module is further configured to determine the The virtual port with the largest difference between the port length and the port cache threshold in the current system, and determining the flow queue of the virtual port with the largest difference between the port length and the port cache threshold as the first flow queue.

According to the fifth possible implementation manner of the second aspect, in the sixth possible implementation manner, the judging module is further configured to: if there is no buffer whose queue length exceeds the highest priority in the first flow queue Threshold flow queues, according to the priority order of each cache threshold in the queue cache threshold, sequentially determine whether there is a flow queue in the first flow queue whose queue length exceeds the cache threshold in the queue cache threshold;

The execution module is further configured to, if there is a flow queue whose queue length exceeds a cache threshold in the first flow queue, perform a dequeue operation on the flow queue exceeding the cache threshold.

According to the sixth possible implementation manner of the second aspect, in the seventh possible implementation manner, the device further includes:

The second sending module is configured to send second warning prompt information if there is no flow queue whose queue length exceeds the lowest priority cache threshold among the queue cache thresholds in the first flow queue; the second warning indication information includes Configuration error message.

According to the fourth or seventh possible implementation manner of the second aspect, in the eighth possible implementation manner, the device further includes:

The configuration module is configured to set the queue buffer threshold according to the priority of the first flow queue before determining whether there is a flow queue whose queue length exceeds the queue buffer threshold in the first flow queue.

According to the first possible implementation manner of the second aspect, in the ninth possible implementation manner, the sum of the port cache thresholds of the flow queues in the current system is less than or equal to N times the shared cache; the current The total cache of the system includes the shared cache;

The N is the ratio of the sum of the port cache thresholds of the flow queues in the current system to the actual cache of the current system.

According to the ninth possible implementation manner of the second aspect, in the tenth possible implementation manner, the judging module is further configured to, before judging whether there is a virtual port whose port length exceeds the port cache threshold in the current system, judge Whether the free cache of the current system is less than the threshold of the protection cache; the total cache of the current system also includes the protection cache; when the free cache of the current system is less than the threshold of the protection cache, execute the judgment Whether the current system has a virtual port whose port length exceeds the port cache threshold.

According to the tenth possible implementation manner of the second aspect, in the eleventh possible implementation manner, the configuration module is further configured to, before determining whether the free cache of the current system is smaller than the protection cache threshold, The protection cache is configured according to the time delay of processing information in the current system.

According to the eighth possible implementation manner of the second aspect, in the twelfth possible implementation manner, the executing module is specifically configured to select a flow queue from the flow queues exceeding the queue cache threshold;

The sending module is further configured to transfer the information corresponding to the selected flow queue to the queue manager, so that the queue manager performs dequeuing of information of a storage packet in the information corresponding to the selected flow queue operate.

According to the twelfth possible implementation manner of the second aspect, in the thirteenth possible implementation manner, the judging module is further configured to judge the selected flow queue according to the aging indication message corresponding to the selected flow queue. Whether the flow queue is a flow queue that needs to be aged;

The execution module is further configured to perform an aging operation on the flow queue that needs to be aged.

According to the twelfth possible implementation manner of the second aspect, in the fourteenth possible implementation manner, the device further includes:

The counting module is configured to, after the information corresponding to the selected flow queue is passed to the queue manager, according to the queue manager, execute the information of one stored packet in the information corresponding to the selected flow queue In the dequeue operation, the value of the dequeue counter corresponding to the selected flow queue is increased by 1 in the dequeue indication message generated later.

According to the fourteenth possible implementation manner of the second aspect, in the fifteenth possible implementation manner, the configuration module is further configured to: The aging indication message corresponding to the selected flow queue is configured as 1; if the value of the dequeuing counter is less than or equal to the dequeuing times threshold, the aging indication message corresponding to the selected flow queue is configured as 0;

The determining module is further configured to determine that the selected flow queue is the flow queue that needs to be aged if the aging indication message corresponding to the selected flow queue continues to be 1 within a predetermined period.

According to the fifteenth possible implementation manner of the second aspect, in the sixteenth possible implementation manner, the configuration module is further configured to determine that the selected flow queue is the flow that needs to be aged After queuing, reduce the configuration of the queue cache threshold corresponding to the selected flow queue; within the predetermined period, if the aging indication message corresponding to the selected flow queue continues to be 0, restore the corresponding The configuration of the queue cache threshold.

The cache management method and device provided by the embodiments of the present invention perform dequeue operations on the stream queues whose queue length exceeds the queue cache threshold in the first stream queue, so that the cache space occupied by the stream queue is adapted to its current cache information, and the cache memory is improved. utilization rate.

Description of drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings that need to be used in the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description These are some embodiments of the present invention. For those skilled in the art, other drawings can also be obtained according to these drawings without any creative effort.

FIG. 1 is a flowchart of a cache management method provided by Embodiment 1 of the present invention;

FIG. 2 is a flowchart of a cache management method provided in Embodiment 2 of the present invention;

FIG. 3 is a flow chart of a cache management method provided in Embodiment 3 of the present invention;

FIG. 4 is a flowchart of a cache management method provided in Embodiment 4 of the present invention;

FIG. 5 is a flowchart of a cache management method provided in Embodiment 5 of the present invention;

FIG. 6 is a flowchart of a cache management method provided in Embodiment 6 of the present invention;

FIG. 7 is a flowchart of a cache management method provided by Embodiment 7 of the present invention;

FIG. 8 is a schematic structural diagram of a cache management device provided by Embodiment 8 of the present invention.

Detailed ways

In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments It is a part of embodiments of the present invention, but not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without making creative efforts belong to the protection scope of the present invention.

It should be noted that the cache thresholds involved in each embodiment of the present invention are pushout thresholds (PushOut with Threshold, referred to as POT), and the dequeue operation involved refers to Push Out dequeue rather than normal dequeue. Normal dequeuing refers to the process of moving the cached information in the flow queue that is executed to the current system memory or hard disk, but Push Out refers to the process of moving the cached information in the flow queue that is executed to be dequeued Push out and discard, so as to achieve the process of cache release.

Embodiment one

FIG. 1 is a flowchart of a cache management method provided by Embodiment 1 of the present invention. The method is executed by a cache management device, which is usually implemented in hardware and/or software and integrated in a network switch or routing device. The method of the present embodiment comprises the steps:

Step 101. Determine whether there is a flow queue whose queue length exceeds the queue cache threshold in the first flow queue.

Specifically, according to different service requirements of users, different flow queues (Flow Queues, FQ for short) correspond to different service flows, and the different flow queues can be used to store data information of the different service flows. The queue length of the flow queue may be, for example, the number of bits of data information stored in the flow queue. The data information stored in the flow queue changes with the change of the user's current processing business, so the queue length of the flow queue also changes with the change of the user's current processing business. The data information of the service flow can be multimedia network data. According to the different service requirements of users, it can be Internet network data, such as voice, video, etc., streaming data information such as Internet TV and Internet phone, and related protocol messages or configuration information. .

The queue cache thresholds of different flow queues can be the same or different. When the types of service flows corresponding to the data information stored in different flow queues are the same, the queue cache thresholds of the different flow queues are the same; when the types of service flows corresponding to the data information stored in different flow queues are different, the The queue cache thresholds are different. Wherein, the queue cache threshold can be represented by fq_thld, and the queue length can be represented by fq_qlen.

The first flow queue includes multiple flow queues. Judging whether there is a flow queue whose queue length exceeds the queue buffer threshold in the first flow queue can be determined by comparing the queue lengths of the multiple flow queues with their corresponding queue buffer thresholds, or according to the queue length of the flow queues. Judgment by limit instructions. For example, the queue overrun indication can use flag bit 1 to indicate that the queue length of the flow queue is greater than its corresponding queue buffer threshold. Correspondingly, flag bit 0 can be used to indicate that the queue length of the flow queue is less than or equal to its corresponding Flow queue for queue cache threshold.

Step 102, if there is a flow queue whose queue length exceeds the queue buffer threshold in the first flow queue, perform a dequeue operation on the flow queue whose queue length exceeds the queue buffer threshold.

Specifically, the queue length of the flow queue does not exceed the queue cache threshold, that is, the buffer space currently allocated for the flow queue is sufficient, and the information cached by the flow queue can be stored and retrieved normally, thereby improving the processing speed of the system and reducing delay.

Correspondingly, the queue length of the flow queue exceeds the queue cache threshold, and the cache space currently occupied by the flow queue has exceeded the configured cache space. Since the total cache space of the current system is certain, that is to say, the flow queue occupies the original By configuring the cache space allocated to other flow queues, it may affect other flow queues to store cache information, and the cache utilization is low. Therefore, it is necessary to dequeue the flow queue that exceeds the queue cache threshold to free up the cache space of other flow queues originally occupied by the flow queue that exceeds the queue cache threshold. Under the condition that the total cache space of the current system is certain , to better ensure the utilization of the cache.

Executing the dequeue operation on the flow queues in the first flow queue that exceed the queue cache threshold may be to perform dequeue operations on all flow queues in the first flow queue whose queue length exceeds the queue cache threshold, or, in the first flow queue Select a predetermined number of flow queues from the flow queues whose queue length exceeds the queue cache threshold to perform the dequeue operation.

Executing the dequeue operation for the flow queue exceeding the queue cache threshold refers to pushing the data information stored in the flow queue exceeding the queue cache threshold out of the corresponding flow queue, which can be discarded to replace the flow queue Perform cache release to improve cache utilization.

The cache management method provided in this embodiment, by dequeuing the flow queue whose queue length exceeds the queue cache threshold in the first flow queue, makes the cache space occupied by the flow queue adapt to the service flow data information currently stored, and improves cache utilization.

At the same time, since the utilization rate of the buffer is improved, the problem of buffer data discarding caused by the full buffer is avoided, so the solution of this embodiment also reduces the packet loss rate of the system, which can meet the quality of service requirements of multimedia network communication.

Embodiment two

This embodiment also provides a cache management method. FIG. 2 is a flowchart of a cache management method provided by Embodiment 2 of the present invention. In the above embodiment, in step 101, it is judged whether there is a flow queue whose queue length exceeds the queue buffer threshold in the first flow queue. Before that, the method also includes:

Step 201. Determine whether the current system has a virtual port whose port length exceeds the port cache threshold.

In the cache management solution of this embodiment, the cache space of the current system is virtually divided into N parts, and the N parts are N virtual ports. Each virtual port corresponds to a plurality of flow queues, and the corresponding relationship between the virtual port and the flow queues may be, for example, configuring the corresponding flow queues for each virtual port by configuring a mapping relationship table. Configuring the corresponding flow queues for each virtual port may be based on the priority of the flow queues, mapping the flow queues with the same priority or close to the same priority to the same virtual port. The priority of the flow queue can be determined according to the type of service flow data information stored in the flow queue. For example, the priority of the flow queue storing protocol packets or configuration information is higher, while the priority of the flow queue storing video data is lower. The priority of the flow queue for storing voice data is lower than that of the flow queue for video data.

The port length of the virtual port can be represented by vi_qlen, which can be the sum of the queue lengths of the current flow queues corresponding to the virtual port. Since each virtual port corresponds to different flow queues, and the priorities of the different flow queues are different, the port cache thresholds corresponding to each virtual port are different. Wherein, the port cache threshold may be represented by vi_thld, and the port cache threshold corresponding to each virtual port may be the sum of the queue cache thresholds of the flow queues corresponding to each virtual port.

According to different cache management strategies, the cache space of the current system is virtually divided into different numbers of virtual ports.

To judge whether there is a virtual port whose port length exceeds the port cache threshold in the current system, it can be judged by comparing the queue lengths of multiple divided virtual ports with their corresponding port cache thresholds, or it can be judged based on the port limit of the virtual port Instructions to judge. The port overrun indication may, for example, use a flag bit 1 to indicate a virtual port whose port length is greater than the port cache threshold, and correspondingly, a flag bit 0 may indicate a virtual port whose port length is less than or equal to the port cache threshold.

Step 202: If the current system does not have a virtual port whose port length exceeds the port cache threshold, determine the flow queue owned by the current system as the first flow queue.

The virtual port whose port length does not exceed the port cache threshold refers to the sum of the queue lengths of the multiple flow queues corresponding to the virtual port, that is, the port length, which does not exceed the port cache threshold of the virtual port. There may be flow queues whose queue length exceeds the queue cache threshold among the corresponding multiple flow queues. Therefore, if it is determined through the above step 201 that there is no virtual port whose port length exceeds the port cache threshold in the current system, it is necessary to use all flow queues in the current system as the first flow queue to implement the solution of the above embodiment and determine the current system. Whether there is a flow queue whose queue length exceeds the queue cache threshold in the flow queues, and then dequeue the flow queue whose queue length exceeds the queue cache threshold.

In the solution of this embodiment, if the current system does not have a virtual port whose port length exceeds the port cache threshold, the flow queue owned by the current system is used as the first flow queue, and then the queue overrun is judged. On the basis of the above-mentioned embodiment solution , to determine the scope of the flow queue more accurately, so as to better ensure the utilization of the cache.

Further, in the solution of this embodiment, the queue cache threshold includes at least two cache thresholds of different levels.

The higher the priority of the at least two caching thresholds of different levels, the larger the caching threshold of the level, and correspondingly, the smaller the caching threshold of the lower priority.

In the above solution, in step 101, it is judged whether there is a flow queue whose queue length exceeds the queue buffer threshold in the first flow queue, specifically including:

Step 203: Compare the queue length of the flow queues in the first flow queue with the highest priority cache threshold in the queue cache threshold, and determine whether there is a flow queue in the first flow queue whose queue length exceeds the highest priority cache threshold.

In step 100 of the above solution, if there is a flow queue whose queue length exceeds the queue cache threshold in the first flow queue, the dequeue operation is performed on the flow queue that exceeds the queue cache threshold, including:

Step 204: If there is a flow queue whose queue length exceeds the highest priority cache threshold in the first flow queue, perform a dequeue operation on the flow queue that exceeds the highest priority cache threshold.

For the judgment result of the above step 203, there is another situation other than step 204, including:

Step 205: If there is no flow queue whose queue length exceeds the highest priority cache threshold in the first flow queue, determine the virtual port with the smallest difference between the port cache threshold and the port length in the current system, and compare the port cache threshold with the The flow queue of the virtual port with the smallest port length difference is determined as the second flow queue.

If it is determined through step 201 that the length of the virtual ports owned by the current system does not exceed the port cache threshold, and it is judged through step 203 in the above scheme that the length of the flow queues owned by the current system does not exceed the highest priority. The cache threshold can only indicate that the system's cache is in a safe state at the current moment. Therefore, it is also necessary to use the flow queue of the virtual port with the smallest difference between the port cache threshold and the port length in the current system as the second flow queue, that is, the subsequent virtual port whose port length is most likely to exceed the port cache threshold. It is called a potentially dangerous virtual port and corresponds to a flow queue.

Step 206 : According to the priority order of each cache threshold in the queue cache threshold, sequentially determine whether there is a flow queue in the second flow queue whose queue length exceeds the cache threshold in the queue cache threshold.

If yes, execute step 207; if not, continue to execute step 206. Judgments are made sequentially according to the priority order of the cache thresholds in the queue cache thresholds. If there is a flow queue whose queue length exceeds a cache threshold in the second flow queue, the cache thresholds behind the cache threshold do not need to be judged. When it is judged that the cache threshold of the last level is reached, if yes, continue to execute step 207 ; if not, execute step 208 . Since the judgment is performed in sequence according to the priority order of the cache thresholds in the queue cache thresholds, the cache threshold of the last level is the cache threshold with the lowest priority among the queue cache thresholds.

Step 207, if there is a flow queue whose queue length exceeds a cache threshold in the second flow queue, perform a dequeue operation on the flow queue that exceeds the cache threshold.

Step 208: If there is no flow queue in the second flow queue whose queue length exceeds the cache threshold with the lowest priority among the queue cache thresholds, send the first warning prompt information. The first warning indication information includes configuration error prompt information.

If there is no flow queue whose queue length exceeds the lowest priority cache threshold among the queue cache thresholds in the second flow queue, it means that the preconfigured queue cache threshold is too large. Therefore, the configuration error prompt message may be a prompt message for an excessively large threshold configuration.

Steps 206 to 208, through hierarchical comparison and judgment, determine the flow queue whose queue length is most likely to exceed the queue cache threshold in the second flow queue, and execute the dequeue operation, so as to provide a more complete cache management solution. It can not only dequeue the flow queues in the "caching danger period", that is, the flow queues whose queue length exceeds the queue cache threshold, but also judge the classification of the flow queues in the "potential cache danger period", making the cache management scheme more efficient. Refined to better ensure the utilization of the cache.

Further, on the basis of the above solution, before judging in step 101 whether there is a flow queue whose queue length exceeds the queue cache threshold in the first flow queue, it also includes:

Configure the queue buffer threshold according to the priority of the first flow queue.

The higher the priority of the first flow queue, the greater the configured queue cache threshold. Correspondingly, the lower the priority of the first flow queue, the smaller the configured queue cache threshold.

Further, the sum of the port cache thresholds of the virtual ports in the current system is less than or equal to N times the shared cache; the total cache of the current system includes the shared cache;

The N is a ratio of the cache threshold with the highest priority to the cache threshold with the lowest priority.

Before the shared cache is exhausted, flow queues and/or packets of any priority will occupy the space of the shared cache first, and different flow queues and/or packets will occupy a certain amount of cache space in a first-come-first-served manner. .

The sum of the port cache thresholds of the virtual ports in the current system refers to the cache threshold of the current system. The cache threshold configuration of the current system is too small, and its cache is difficult to be fully utilized. Therefore, the cache threshold should be at least larger than the shared cache; if the configuration is too large, it is difficult to dequeue the flow queue through this cache management method, and configuration error prompts may easily appear.

The cache threshold of the current system is also equal to the sum of the highest cache thresholds among the queue cache thresholds of the flow queues in the current system. The queue cache threshold of the flow queue may include, for example, four levels of cache thresholds, which are fq_thld, fq_thld*1/2, fq_thld*1/4, and fq_thld*1/8 in order of priority.

If the ratio of the cache threshold with the highest priority to the cache threshold with the lowest priority is 8, then the sum of the port cache thresholds of the virtual ports in the current system is less than or equal to 8 times of the shared cache. If the cache threshold of the lowest priority is fq_thld*1/16, the sum of the port cache thresholds of the virtual ports in the current system is less than or equal to 16 times of the shared cache.

This embodiment scheme, on the basis of the above-mentioned embodiment scheme, provides a more complete cache management scheme, which can not only dequeue the flow queues in the "cache danger period", that is, the flow queues whose queue length exceeds the queue cache threshold It also judges the classification of flow queues in the "potential cache danger period", making the cache management plan more refined and better ensuring the utilization of the cache.

Embodiment Three

This embodiment also provides a cache management method. FIG. 3 is a flowchart of a cache management method provided by Embodiment 3 of the present invention. In the above embodiment, the judgment result of step 201, in addition to the above step 202, there is another situation, specifically including:

Step 301, if the current system has a virtual port whose port length exceeds the port cache threshold, determine the virtual port with the largest difference between the port length and the port cache threshold in the current system, and calculate the virtual port with the largest difference between the port length and the port cache threshold The owned flow queue is determined as the first flow queue.

Specifically, on the basis of the solution in the above-mentioned embodiment 2, this embodiment also provides a cache management method in the case that the current system has a virtual port whose port length exceeds the port cache threshold. In this method, the flow queue of the virtual port with the largest difference between the port length and the port cache threshold in the current system is used as the first flow queue to perform the operation of judging the queue overrun.

On the basis of the above scheme, in step 101 of the above scheme, it is judged whether there is a flow queue whose queue length exceeds the queue buffer threshold in the first flow queue, and further includes:

Step 302: If there is no flow queue whose queue length exceeds the highest priority cache threshold in the first flow queue, determine whether there is a queue in the first flow queue according to the priority order of each cache threshold in the queue cache threshold Flow queues whose length exceeds the cache threshold in Cache Thresholds for this queue.

Correspondingly, in step 102 of the above solution, if there is a flow queue whose queue length exceeds the queue cache threshold in the first flow queue, the dequeue operation is performed on the flow queue exceeding the queue cache threshold, which also includes:

Step 303, if there is a flow queue whose queue length exceeds a buffer threshold in the first flow queue, perform a dequeue operation on the flow queue whose queue length exceeds the buffer threshold.

Further, the method also includes:

Step 304: If there is no flow queue in the first flow queue whose queue length exceeds the cache threshold with the lowest priority among the queue cache thresholds, send a second warning indication message; the second warning indication information includes configuration error prompt information.

Further, before judging in step 101 of the above solution whether there is a flow queue whose queue length exceeds the queue cache threshold in the first flow queue, it also includes:

Set the queue cache threshold according to the priority of the first flow queue.

The sum of the port cache thresholds of the flow queues in the current system is less than or equal to N times the shared cache; the total cache of the current system includes the shared cache;

N is the ratio of the sum of the port cache thresholds of the flow queues in the current system to the actual cache of the current system.

The solution of this embodiment and the solution of the above-mentioned embodiments are respectively applicable to different scenarios. It can not only dequeue the flow queues in the "caching danger period", that is, the flow queues whose queue length exceeds the queue cache threshold, but also dequeue the flow queues in the The "potential cache danger period" is judged by the flow queue classification, which makes the cache management plan more refined and better guarantees the utilization rate of the cache.

Embodiment Four

This embodiment also provides a cache management method. FIG. 4 is a flowchart of a cache management method provided by Embodiment 4 of the present invention. Before step 201 in the solution of the above embodiment determines whether there is a virtual port whose port length exceeds the port cache threshold in the current system, the method further includes:

Step 401, judging whether the free cache of the current system is smaller than the threshold of the protected cache.

In order to make the cache management solution run more stably, on the basis of the above solution, the total cache space of the current system includes the protection cache in addition to the shared cache. The protection cache is used to absorb and process the packets with path delay during the process of dequeuing the flow queue. Among them, the idle cache can be represented by idle_buf, and the protection cache threshold can be represented by glb_buf.

The free cache of the current system refers to the total cache space minus the sum of the port lengths corresponding to all virtual ports in the current system, that is, the remaining cache space of the actual cache of the current system. Judging whether the free cache of the current system is smaller than the protection cache threshold is actually judging whether the shared cache of the current system is exhausted.

Step 402: If the free cache of the current system is smaller than the threshold of the protection cache, perform an operation of judging whether the current system has a virtual port whose port length exceeds the threshold of the port cache.

The space cache of the previous system is less than the protection cache threshold, which means that the shared cache of the current system has been exhausted, and it is necessary to implement a cache management strategy to dequeue the scheduled flow queues to free up the corresponding cache space, so as to ensure that the current system total cache utilization.

If the free cache of the current system is greater than or equal to the protection cache threshold, it means that there is still a shared cache in the current system, so there is no need to dequeue the flow queue through the cache management strategy.

Further, before step 401 is judged whether the free cache of the current system is smaller than the protection cache threshold, it also includes:

According to the time delay of processing information in the current system, the threshold for protecting the cache is set.

If the delay in processing information in the current system is large, the queue length corresponding to the flow queue may not be updated in time after the flow queue is dequeued, resulting in the idle cache of the current system not being updated in time. If the protection cache threshold is set too small, the current system will leave the queue too frequently, which will easily cause the overflow of cache information, thus affecting the stability of the system. Therefore, if the delay in processing information in the current system is relatively large, a larger protection cache threshold can be configured to avoid overflow of cached information. Correspondingly, if the delay in processing information in the current system is small, the protection cache threshold can be configured to be smaller.

The scheme of this embodiment can avoid the overflow of cached information, thereby ensuring the stability of the current system.

Embodiment five

This embodiment also provides a buffer management method. This embodiment explains the solutions of the above embodiments through specific examples, wherein the queue buffer thresholds include: fq_thld, fq_thld*1/2, fq_thld*1/4, fq_thld*1/ 8. fq_thld has the highest priority, and fq_thld*1/8 has the lowest priority. FIG. 5 is a flowchart of a cache management method provided by Embodiment 5 of the present invention. As shown in Figure 5, the method includes:

Step 501, judge that idle_buf<glb_buf of the current system. If yes, execute step 502; if not, the cache management solution ends.

Step 502, judging whether there is a virtual port with vi_qlen>vi_thld in the current system. If yes, execute step 503; if not, execute step 504.

Step 503: Determine the virtual port with the largest difference between vi_qlen-vi_thld in the current system, and determine the flow queue of the virtual port as the first flow queue.

Step 505 is continued after step 503 .

Step 504: Determine the flow queue of the current system as the first flow queue.

After step 504, continue to execute step 514.

Step 505. Determine whether there is a flow queue with fq_qlen>fq_thld in the first flow queue. If yes, go to step 506; if not, go to step 507.

Step 506: Select a flow queue from the flow queues exceeding the queue cache threshold to perform a dequeue operation.

Step 507. Determine whether there is a flow queue with fq_qlen>fq_thld*1/2 in the first flow queue. If yes, go to step 508; if not, go to step 509.

Step 508 : Select a flow queue from the flow queues that exceed 1/2 of the queue cache threshold to perform a dequeue operation.

Step 509, judging whether there is a flow queue with fq_qlen>fq_thld*1/4 in the first flow queue. If yes, execute step 510; if not, execute step 511.

Step 510: Select a flow queue from the flow queues exceeding 1/4 of the queue cache threshold to perform a dequeue operation.

Step 511. Determine whether there is a flow queue with fq_qlen>fq_thld*1/8 in the first flow queue. If yes, execute step 512; if not, execute step 513.

Step 512: Select a flow queue from the flow queues exceeding 1/8 of the queue cache threshold to perform a dequeue operation.

Step 513: Send second warning indication information, where the second warning indication information includes configuration error prompt information.

Step 514, determine whether there is a flow queue with fq_qlen>fq_thld in the first flow queue. If yes, go to step 515; if not, go to step 516.

The difference between step 514 and step 505 is that the objects contained in the first flow queue are different.

Step 515 , select a flow queue from the flow queues exceeding the queue cache threshold to perform a dequeue operation.

Step 516: Determine the virtual port with the smallest vi_thld-vi_qlen difference in the current system, and determine the flow queue of the virtual port as the second flow queue.

Step 517. Determine whether there is a flow queue with fq_qlen>fq_thld*1/2 in the second flow queue. If yes, go to step 518; if not, go to step 519.

Step 518: Select a flow queue from the flow queues that exceed 1/2 of the queue cache threshold to perform a dequeue operation.

Step 519. Determine whether there is a flow queue with fq_qlen>fq_thld*1/4 in the second flow queue. If yes, go to step 520; if not, go to step 521.

Step 520 , select a flow queue from the flow queues exceeding 1/4 of the queue cache threshold to perform a dequeue operation.

Step 521. Determine whether there is a flow queue with fq_qlen>fq_thld*1/8 in the second flow queue. If yes, execute step 522; if not, execute step 523.

Step 522: Select a flow queue from the flow queues exceeding 1/8 of the queue cache threshold to perform a dequeue operation.

Step 523: Send first warning indication information, where the first warning indication information includes configuration error prompt information.

The solution of this embodiment is based on the solutions of the above embodiments, and is explained through specific examples. Its specific implementation process and beneficial effects are similar to those of the above embodiments, and will not be repeated here.

Embodiment six

Embodiment 6 of the present invention provides a cache management method. FIG. 6 is a flowchart of a cache management method provided in Embodiment 6 of the present invention.

As shown in FIG. 6 , further, in step 102 of the above embodiment solution, the dequeue operation is performed on the flow queue exceeding the queue cache threshold, which specifically includes the following steps:

Step 601. Select a flow queue from the flow queues exceeding the queue cache threshold.

Specifically, the flow queue exceeding the queue cache threshold may be multiple flow queues, or may be one flow queue. In order to avoid the impact on the system caused by the dequeue of the flow queue as much as possible, the solution of this embodiment preferably selects a flow queue from the flow queues exceeding the queue cache threshold to dequeue.

Step 602: Send the information corresponding to the selected flow queue to the queue manager, so that the queue manager performs a dequeue operation on the information of one stored packet in the information corresponding to the selected flow queue.

Executing the dequeue operation on the stream queue refers to performing the dequeue operation on the data information cached in the stream queue. There are multiple storage packets in each flow queue, and the flow queue caches information through its buffer unit. Preferably, in the solution of this embodiment, among the information cached in the flow queue to be dequeued selected in step 601 , select information of a stored packet to be dequeued. Correspondingly, after the dequeue operation is performed, the queue length of the selected flow queue is reduced by the information length of one storage packet.

In this embodiment scheme, on the basis of the above-mentioned embodiment scheme, specifically, a flow queue is selected in the flow queue exceeding the cache threshold of the queue, and then the information corresponding to the flow queue is dequeued for a stored packet information , so as to avoid the impact of dequeue operation on the current system processing business as much as possible.

Embodiment seven

Embodiment 7 of the present invention provides a cache management method. FIG. 7 is a flowchart of a cache management method provided by Embodiment 7 of the present invention.

As shown in Figure 7, further, on the basis of the above solution, the method also includes:

Step 701. According to the aging indication message corresponding to the selected flow queue, determine whether the selected flow queue is a flow queue that needs to be aged.

Step 702, perform an aging operation on the flow queue that needs to be aged.

Wherein, the aging indication message may be represented by flush_ind, for example. The information carried by the flush_ind is used to identify whether the flow queue is a flow queue to be aged.

In the prior art, queue aging is performed according to the timestamp printed by the storage packets in the flow queue when data is enqueued. Since each flow queue has multiple storage packages, printing a time stamp for each storage package occupies a large amount of time stamp resources, resulting in a waste of resources. In the solution of this embodiment, the aging of the flow queue can be realized without printing complicated time stamps, and the waste of system resources can be reduced.

In step 602 of the above solution, after transferring the information corresponding to the selected flow queue to the queue manager, it also includes:

According to the queue manager, the dequeue operation is performed on the information of a stored packet in the information corresponding to the selected flow queue, and then the dequeue indication message generated will add 1 to the value of the dequeue counter corresponding to the selected flow queue.

The dequeue indication message may be indicated by fq_dep_ind, for example. The dequeue counter, for example, can be represented by fq_pot_cnt. After the selected flow queue executes the dequeue operation, its corresponding dequeue indication message, fq_dep_ind can be configured as 1. If it is detected that fq_dep_ind=1, the value of fq_pot_cnt, its corresponding out-of-queue counter, can be increased by one.

Each flow queue has its own corresponding dequeue timer. The dequeue counter is used to record the number of dequeue times of its corresponding flow queue. Every time a stream queue executes a dequeue operation, the value of its dequeue counter is incremented by one.

Further, the above step 701 judges whether the selected flow queue is a flow queue that needs to be aged according to the aging indication message corresponding to the selected flow queue, including:

If the value of the dequeue counter is greater than the dequeue times threshold, configure the aging indication message corresponding to the selected flow queue as 1; if the value of the dequeue counter is less than or equal to the dequeue times threshold, the selected flow queue corresponds The aging indication message is configured as 0;

Within a predetermined period, if the aging indication message corresponding to the selected flow queue continues to be 1, it is determined that the selected flow queue is a flow queue that needs to be aged.

Specifically, in the solution of this embodiment, in the case that the flow queues all have dequeue counters, they also have at least a dequeue times threshold. Different flow queues may have different dequeue times thresholds according to the priority of their corresponding virtual ports. The threshold of dequeuing times of the flow queue, for example, can be passed through fq_pot_thld.

When the value of the dequeue counter is greater than the threshold of dequeue times, that is, fq_pot_cnt>fq_pot_thld, it means that the flow queue is dequeued too frequently, indicating that the flow queue is congested, and the cached information stored in it cannot be dequeued naturally for a long time. Therefore, the corresponding aging indication message, flush_ind, can be configured as 1, and then the queue aging operation is performed on the flow queue according to the queue aging mode. In the predetermined period, if the aging indication message corresponding to the selected flow queue, flush_ind continues to be 1, that is to say, the selected flow queue continues to be in the queue length exceeding state, and the queue needs to be repeatedly dequeued, then the surface The selected flow queue belongs to the long-term congestion queue. Therefore, the long-term congestion queue can be determined as a flow queue that needs to be aged, and then the queue operation is performed to transfer the cache information currently stored in the flow queue to other queues or ports.

Further, after determining that the selected flow queue is a flow queue that needs to be aged, it also includes:

Reduce the configuration of the queue cache threshold corresponding to the selected flow queue.

Within a predetermined period, if the aging indication message corresponding to the selected flow queue continues to be 0, the configuration of the queue cache threshold corresponding to the selected flow queue is restored.

Specifically, if the selected flow queue is a flow queue that needs to be aged, in order to avoid queue congestion that is likely to be caused when information is cached through the flow queue, the pre-configured queue cache threshold corresponding to the flow queue can be reduced by Small, so as to increase the possibility that the flow queue may be dequeued and reduce the possibility of queue aging. The pre-configured queue cache threshold corresponding to the flow queue can be represented by fq_thld, for example. Reduce the configuration of the queue cache threshold corresponding to the selected flow queue, that is, the configuration of fq_thld.

If the aging indication message corresponding to the selected flow queue, flush_ind remains 0, that is to say, the flow queue is not a congested queue, does not need to perform queue aging, and the state is stable, so restore the queue cache threshold value corresponding to the selected flow queue configuration. Restore the configuration of the queue cache threshold corresponding to the selected flow queue, reduce the possibility that the flow queue may be dequeued, and avoid resource waste caused by multiple executions of the dequeue operation.

Furthermore, in addition to the long-term congestion queues, the flow queues that need to be aged in the system also include flow queues that only go out of the queue and flow queues that only enter the queue. During the predetermined period, if it is detected that the fq_dep_ind corresponding to the flow queue is continuously 1, then the flow queue is the flow queue that only dequeues, and the queue aging operation can be performed on the flow queue. After the queue aging operation is performed, the fq_dep_ind is reset to 0. For an empty flow queue, that is, when it is detected that the queue length of the flow queue is 0, the fq_dep_ind corresponding to the flow queue is also updated to 1. Therefore, within a predetermined period, if it is detected that the fq_dep_ind corresponding to the flow queue is continuously 0, the flow queue has no dequeue operation and is a flow queue that only enters the queue, and the queue aging operation can be performed on the flow queue.

In the solution of this embodiment, on the basis that the solution in the above embodiments can improve the utilization rate of the system cache, system resources can also be saved.

Embodiment Eight

Embodiment 8 of the present invention provides a cache management device. FIG. 8 is a schematic structural diagram of a cache management device provided by Embodiment 8 of the present invention. As shown in Figure 8, the cache management device 801 includes:

A judging module 802, configured to judge whether there is a flow queue whose queue length exceeds a queue cache threshold in the first flow queue.

The execution module 803 is configured to, if there is a flow queue whose queue length exceeds the queue cache threshold in the first flow queue, perform a dequeue operation on the flow queue exceeding the queue cache threshold.

On the basis of the above solution, the judging module 802 is further configured to judge whether there is a virtual port whose port length exceeds the port caching threshold in the current system before judging whether the queue length of the first flow queue exceeds the queue caching threshold.

The cache management device 801 also includes:

The determining module is configured to determine the flow queue of the current system as the first flow queue if there is no virtual port whose port length exceeds the port cache threshold in the current system.

In the solutions of the foregoing embodiments, the queue cache threshold includes at least two cache thresholds with different priorities.

Correspondingly, the judging module 802 is configured to compare the queue length of the flow queue in the first flow queue with the highest priority cache threshold in the queue cache threshold, and judge whether there is a cache whose queue length exceeds the highest priority in the first flow queue Threshold for flow queues.

The execution module 803 is configured to, if there is a flow queue whose queue length exceeds the highest-priority cache threshold in the first flow queue, perform a dequeue operation on the flow queue that exceeds the highest-priority cache threshold.

Further, the determination module in the above solution is also used to determine the virtual flow queue with the smallest difference between the port buffer threshold and the port length in the current system if there is no flow queue whose queue length exceeds the highest priority buffer threshold in the first flow queue. port, and determine the flow queue of the virtual port with the smallest difference between the port cache threshold and the port length as the second flow queue.

Correspondingly, the judging module 802 is further configured to sequentially judge whether there is a flow queue in the second flow queue whose queue length exceeds the cache threshold in the queue cache threshold according to the priority sequence of each cache threshold in the queue cache threshold.

The executing module 803 is further configured to, if there is a flow queue whose queue length exceeds a cache threshold in the second flow queue, perform a dequeue operation on the flow queue exceeding the cache threshold.

Further, the cache management device 801 also includes:

The first sending module is configured to send first warning prompt information if there is no flow queue whose queue length exceeds the lowest priority cache threshold among the queue cache thresholds in the second flow queue; the first warning indication information includes configuration Error message.

Wherein, the determination module is also used to determine the virtual port with the largest difference between the port length and the port cache threshold in the current system if the current system has a virtual port whose port length exceeds the port cache threshold, and compare the port length with the port cache threshold The flow queue of the virtual port with the largest difference is determined as the first flow queue.

On the basis of the solutions in the above embodiments, the judging module 802 is further configured to: if there is no flow queue whose queue length exceeds the highest priority cache threshold in the first flow queue, according to the priority of each cache threshold in the queue cache threshold In order, determine whether there is a flow queue in the first flow queue whose queue length exceeds the cache threshold in the queue cache threshold;

The execution module 803 is further configured to, if there is a flow queue whose queue length exceeds a cache threshold in the first flow queue, perform a dequeue operation on the flow queue that exceeds the cache threshold.

Further, on the basis of the solutions in the above embodiments, the cache management device 801 further includes:

The second sending module is configured to send a second warning message if there is no flow queue in the first flow queue whose queue length exceeds the cache threshold with the lowest priority among the queue cache thresholds; the second warning message includes a configuration error prompt information.

Further, the cache management device 801 also includes:

The configuration module is configured to set the queue cache threshold according to the priority of the first flow queue before judging whether there is a flow queue whose queue length exceeds the queue cache threshold in the first flow queue.

In the solution of the above embodiment, the sum of the port cache thresholds of the flow queues in the current system is less than or equal to N times the shared cache; the total cache of the current system includes the shared cache;

Wherein, N is the ratio of the sum of the port cache thresholds of the flow queues in the current system to the actual cache of the current system.

Further, the judging module 802 is also used to judge whether the current system has a virtual port whose port length exceeds the port cache threshold, and judge whether the current system’s free cache is smaller than the protection cache threshold; the current system’s total cache also includes protection Cache: when the free cache of the current system is less than the threshold of the protection cache, perform the operation of judging whether the current system has a virtual port whose port length exceeds the threshold of the port cache.

On the basis of the above solution, further, the configuration module is further configured to set the threshold of the protection cache according to the time delay of processing information in the current system before judging whether the free cache of the current system is smaller than the threshold of the protection cache.

Further, the executing module 803 is specifically configured to select a flow queue from the flow queues exceeding the queue cache threshold;

The sending module is further configured to transmit the information corresponding to the selected flow queue to the queue manager, so that the queue manager performs a dequeue operation on the information of one stored packet in the information corresponding to the selected flow queue.

On the basis of the above solution, the judging module 802 is further configured to judge whether the selected flow queue is a flow queue that needs to be aged according to the aging indication message corresponding to the selected flow queue;

The execution module 803 is further configured to perform an aging operation on the flow queue that needs to be aged.

Further, the cache management device 801 also includes:

The counting module is configured to, after passing the information corresponding to the selected flow queue to the queue manager, according to the queue manager, perform a dequeue operation on the information of a storage package in the information corresponding to the selected flow queue, and then The generated dequeue indication message adds 1 to the value of the dequeue counter corresponding to the selected flow queue.

Further, the configuration module is also used to configure the aging indication message corresponding to the selected flow queue as 1 if the value of the dequeue counter is greater than the dequeue times threshold; if the value of the dequeue counter is less than or equal to the dequeue times Threshold, the aging indication message corresponding to the selected flow queue is configured as 0; the determination module is also used to determine the selected flow queue if the aging indication message corresponding to the selected flow queue continues to be 1 within a predetermined period For the flow queue that needs to be aged.

On the basis of the solutions of the above embodiments, the configuration module is further configured to reduce the configuration of the queue cache threshold corresponding to the selected flow queue after determining that the selected flow queue is a flow queue that needs to be aged;

Within the predetermined period, if the aging indication message corresponding to the selected flow queue continues to be 0, the configuration of the queue cache threshold corresponding to the selected flow queue is restored.

The cache management device provided in this embodiment can execute the cache management method provided in any of the above-mentioned embodiments, and its specific implementation process and explanation are similar to those of the above-mentioned embodiments, and will not be repeated here.

Those of ordinary skill in the art can understand that all or part of the steps for implementing the above method embodiments can be completed by program instructions and related hardware. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps including the above-mentioned method embodiments; and the aforementioned storage medium includes: ROM, RAM, magnetic disk or optical disk and other various media that can store program codes.

Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present invention, rather than limiting them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: It is still possible to modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some or all of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the technical solutions of the various embodiments of the present invention. scope.

Claims (34)

1. a buffer memory management method, is characterized in that, comprising:

Whether judge in first-class queue exists queue length to exceed the flow queue of queue buffer memory threshold value;

If exist queue length to exceed the flow queue of described queue buffer memory threshold value in described first-class queue, the described flow queue that exceeds described queue buffer memory threshold value carried out out to team's operation.

2. method according to claim 1, is characterized in that, the described queue length that judges first-class queue also comprises before whether exceeding queue buffer memory threshold value:

Judge whether current system exists port length to exceed the virtual port of buffer to ports threshold value;

If described current system does not exist port length to exceed the virtual port of buffer to ports threshold value, the flow queue that described current system is had is defined as described first-class queue.

3. method according to claim 2, is characterized in that, described queue buffer memory threshold value comprises the buffer memory threshold value of at least two different priorities;

Describedly judge in first-class queue whether exist queue length to exceed the flow queue of queue buffer memory threshold value, to comprise:

The buffer memory threshold value of limit priority in the queue length of flow queue in described first-class queue and described queue buffer memory threshold value is compared, judge in first-class queue and whether exist queue length to exceed the flow queue of the buffer memory threshold value of described limit priority;

If exist queue length to exceed the flow queue of described queue buffer memory threshold value in described first-class queue, the described flow queue that exceeds described queue buffer memory threshold value is carried out out to team's operation, comprising:

If exist queue length to exceed the flow queue of the buffer memory threshold value of described limit priority in described first-class queue, the described flow queue that exceeds the buffer memory threshold value of described limit priority carried out out to team's operation.

4. method according to claim 3, is characterized in that, also comprises:

If do not exist queue length to exceed the flow queue of the buffer memory threshold value of described limit priority in described first-class queue, determine the virtual port of described current system middle port buffer memory threshold value and port length difference minimum, and the flow queue that the virtual port of described buffer to ports threshold value and port length difference minimum is had is defined as second queue;

According to the priority orders of each buffer memory threshold value in described queue buffer memory threshold value, judge successively in described second queue and whether exist queue length to exceed the flow queue of the buffer memory threshold value in described queue buffer memory threshold value;

If exist queue length to exceed the flow queue of a buffer memory threshold value in described second queue, carry out out team's operation to exceeding the flow queue of this buffer memory threshold value.

5. method according to claim 4, is characterized in that, also comprises:

If do not exist queue length to exceed the flow queue of the buffer memory threshold value of lowest priority in described queue buffer memory threshold value in described second queue, send the first alarm prompt; Described the first alarm indication information comprises configuration error information.

6. method according to claim 3, is characterized in that,

If described current system exists port length to exceed the virtual port of buffer to ports threshold value, determine the virtual port of described current system middle port length and buffer to ports threshold difference maximum, and the flow queue that the virtual port of described port length and buffer to ports threshold difference maximum is had is defined as described first-class queue.

7. method according to claim 6, is characterized in that, describedly judges in first-class queue whether exist queue length to exceed the flow queue of queue buffer memory threshold value, also to comprise:

If do not exist queue length to exceed the flow queue of the buffer memory threshold value of described limit priority in described first-class queue, according to the priority orders of each buffer memory threshold value in described queue buffer memory threshold value, judge successively in described first-class queue and whether exist queue length to exceed the flow queue of the buffer memory threshold value in described queue buffer memory threshold value;

If exist queue length to exceed the flow queue of described queue buffer memory threshold value in described first-class queue, the described flow queue that exceeds described queue buffer memory threshold value is carried out out to team's operation, also comprise:

If exist queue length to exceed the flow queue of a buffer memory threshold value in described first-class queue, carry out out team's operation to exceeding the flow queue of this buffer memory threshold value.

8. method according to claim 7, is characterized in that, also comprises:

If do not exist queue length to exceed the flow queue of the buffer memory threshold value of lowest priority in described queue buffer memory threshold value in described first-class queue, send the second alarm prompt; Described the second alarm indication information comprises configuration error information.

9. according to the method described in claim 5 or 8, it is characterized in that whether existing in judging first-class queue before queue length exceeds the flow queue of queue buffer memory threshold value described, also comprise:

According to the priority of described first-class queue, described queue buffer memory threshold value is set.

10. method according to claim 2, is characterized in that, the buffer to ports threshold value sum of the flow queue having in described current system is less than or equal to N times of shared buffer memory; Total buffer memory of described current system comprises described shared buffer memory;

Described N is the buffer to ports threshold value sum of flow queue that has in described current system and the ratio of the real cache of described current system.

11. methods according to claim 10, is characterized in that, describedly judge that whether current system exists before port length exceeds the virtual port of buffer to ports threshold value, also comprises:

Judge whether the free buffer of described current system is less than the threshold value of protection buffer memory; Total buffer memory of described current system, also comprises described protection buffer memory;

When the free buffer of described current system is less than the threshold value of described protection buffer memory, carries out and judge whether described current system exists port length to exceed the operation of the virtual port of buffer to ports threshold value.

12. methods according to claim 11, is characterized in that, before whether the described free buffer that judges described current system is less than protection buffer memory threshold value, also comprise:

According to the time delay of process information in described current system, described protection buffer memory is configured.

13. methods according to claim 9, is characterized in that, described the described flow queue that exceeds queue buffer memory threshold value are carried out out to team operation, comprising:

In the described flow queue that exceeds queue buffer memory threshold value, select a flow queue;

The information that the flow queue of described selection is corresponding passes to queue management device, so that described queue management device goes out team's operation to the information and executing of a storage bag in information corresponding to the flow queue of described selection.

14. methods according to claim 13, is characterized in that, also comprise:

The aging Indication message corresponding according to the flow queue of described selection, judges whether the flow queue of described selection is to need aging flow queue;

The aging flow queue of described needs is carried out to aging operation.

15. methods according to claim 13, is characterized in that, described information corresponding to the flow queue of described selection is passed to queue management device after, also comprise:

According to described queue management device, in to information corresponding to the flow queue of described selection, the information and executing of a storage bag goes out team's operation, and what produce afterwards goes out group Indication message, and the value that goes out group counter that the flow queue of described selection is corresponding adds 1.

16. methods according to claim 15, is characterized in that, the described aging Indication message corresponding according to the flow queue of described selection judges whether the flow queue of described selection is to need aging flow queue, comprising:

If described in go out group counter value be greater than out group frequency threshold value, the aging Indication message that the flow queue of described selection is corresponding is configured to 1; If described in go out group counter value go out group frequency threshold value described in being less than or equal to, the aging Indication message that the flow queue of described selection is corresponding is configured to 0;

In predetermined period, if aging Indication message corresponding to the flow queue of described selection is continuously 1, determine that the flow queue of described selection is the flow queue that described needs are aging.

17. methods according to claim 16, is characterized in that, after the flow queue of described definite described selection is the flow queue that described needs are aging, also comprise:

Reduce the configuration of the queue buffer memory threshold value that the flow queue of described selection is corresponding;

In described predetermined period, if aging Indication message corresponding to the flow queue of described selection is continuously 0, recover the configuration of the queue buffer memory threshold value that the flow queue of described selection is corresponding.

18. 1 kinds of cache management devices, is characterized in that, comprising:

Judge module, for judging whether first-class queue exists queue length to exceed the flow queue of queue buffer memory threshold value;

Executive Module, if exist queue length to exceed the flow queue of described queue buffer memory threshold value for described first-class queue, carries out out team's operation to the described flow queue that exceeds described queue buffer memory threshold value.

19. devices according to claim 18, is characterized in that,

Described judge module, also, for before whether exceeding queue buffer memory threshold value in the described queue length that judges first-class queue, judges whether current system exists port length to exceed the virtual port of buffer to ports threshold value;

Described device, also comprises:

Determination module, if do not exist port length to exceed the virtual port of buffer to ports threshold value for described current system, the flow queue that described current system is had is defined as described first-class queue.

20. devices according to claim 19, is characterized in that, described queue buffer memory threshold value comprises the buffer memory threshold value of at least two different priorities;

Described judge module, for the buffer memory threshold value of limit priority in the queue length of described first-class queue flow queue and described queue buffer memory threshold value is compared, judge in first-class queue and whether exist queue length to exceed the flow queue of the buffer memory threshold value of described limit priority;

Described Executive Module, if exist queue length to exceed the flow queue of the buffer memory threshold value of described limit priority for described first-class queue, carries out out team's operation to the described flow queue that exceeds the buffer memory threshold value of described limit priority.

21. devices according to claim 20, is characterized in that,

Described determination module, if also do not exist queue length to exceed the flow queue of the buffer memory threshold value of described limit priority for described first-class queue, determine the virtual port of described current system middle port buffer memory threshold value and port length difference minimum, and the flow queue that the virtual port of described buffer to ports threshold value and port length difference minimum is had is defined as second queue;

Whether described judge module, also, for according to the priority orders of described each buffer memory threshold value of queue buffer memory threshold value, judge successively in described second queue and exist queue length to exceed the flow queue of the buffer memory threshold value in described queue buffer memory threshold value;

Described Executive Module, if also exist queue length to exceed the flow queue of a buffer memory threshold value for described second queue, carries out out team's operation to exceeding the flow queue of this buffer memory threshold value.

22. devices according to claim 21, is characterized in that, described device, also comprises:

The first sending module, if do not exist queue length to exceed the flow queue of the buffer memory threshold value of lowest priority in described queue buffer memory threshold value for described second queue, sends the first alarm prompt; Described the first alarm indication information comprises configuration error information.

23. devices according to claim 20, is characterized in that,

Described determination module, if also exist port length to exceed the virtual port of buffer to ports threshold value for described current system, determine the virtual port of described current system middle port length and buffer to ports threshold difference maximum, and the flow queue that the virtual port of described port length and buffer to ports threshold difference maximum is had is defined as described first-class queue.

24. devices according to claim 23, is characterized in that,

Described judge module, if also do not exist queue length to exceed the flow queue of the buffer memory threshold value of described limit priority for described first-class queue, according to the priority orders of each buffer memory threshold value in described queue buffer memory threshold value, judge successively in described first-class queue and whether exist queue length to exceed the flow queue of the buffer memory threshold value in described queue buffer memory threshold value;

Described Executive Module, if also exist queue length to exceed the flow queue of a buffer memory threshold value for described first-class queue, carries out out team's operation to exceeding the flow queue of this buffer memory threshold value.

25. devices according to claim 24, is characterized in that, described device, also comprises:

The second sending module, if do not exist queue length to exceed the flow queue of the buffer memory threshold value of lowest priority in described queue buffer memory threshold value for described first-class queue, sends the second alarm prompt; Described the second alarm indication information comprises configuration error information.

26. according to the device described in claim 22 or 25, it is characterized in that, described device, also comprises:

Configuration module, for judging that described whether first-class queue exists before queue length exceeds the flow queue of queue buffer memory threshold value, according to the priority of described first-class queue, arranges described queue buffer memory threshold value.

27. devices according to claim 19, is characterized in that, the buffer to ports threshold value sum of the flow queue having in described current system is less than or equal to N times of shared buffer memory; Total buffer memory of described current system comprises described shared buffer memory;

Described N is the buffer to ports threshold value sum of flow queue that has in described current system and the ratio of the real cache of described current system.

28. devices according to claim 27, is characterized in that,

Described judge module, also judges that for described whether current system exists before port length exceeds the virtual port of buffer to ports threshold value, judges whether the free buffer of described current system is less than the threshold value of protection buffer memory; Total buffer memory of described current system, also comprises described protection buffer memory; When the free buffer of described current system is less than the threshold value of described protection buffer memory, carries out and judge whether described current system exists port length to exceed the operation of the virtual port of buffer to ports threshold value.

29. devices according to claim 28, is characterized in that,

Described configuration module, also, for before whether being less than protection buffer memory threshold value in the described free buffer that judges described current system, according to the time delay of process information in described current system, is configured described protection buffer memory.

30. devices according to claim 26, is characterized in that,

Described Executive Module, specifically for selecting a flow queue in the described flow queue that exceeds queue buffer memory threshold value;

Described sending module, also for information corresponding to the flow queue of described selection is passed to queue management device, so that described queue management device goes out team's operation to the information and executing of a storage bag in information corresponding to the flow queue of described selection.

31. devices according to claim 30, is characterized in that,

Described judge module, also for the aging Indication message corresponding according to the flow queue of described selection, judges whether the flow queue of described selection is to need aging flow queue;

Described Executive Module, also for carrying out aging operation to the aging flow queue of described needs.

32. devices according to claim 30, is characterized in that, described device, also comprises:

Counting module, for described information corresponding to the flow queue of described selection is passed to queue management device after, according to described queue management device, in to information corresponding to the flow queue of described selection, the information and executing of a storage bag goes out team's operation, what produce afterwards goes out group Indication message, and the value that goes out group counter that the flow queue of described selection is corresponding adds 1.

33. devices according to claim 32, is characterized in that,

Described configuration module, if also for described in go out group counter value be greater than out group frequency threshold value, the aging Indication message that the flow queue of described selection is corresponding is configured to 1; If described in go out group counter value go out group frequency threshold value described in being less than or equal to, the aging Indication message that the flow queue of described selection is corresponding is configured to 0;

Described determination module, also in predetermined period, if aging Indication message corresponding to the flow queue of described selection is continuously 1, determines that the flow queue of described selection is the flow queue that described needs are aging.

34. devices according to claim 33, is characterized in that,

Described configuration module, also for after the described flow queue of determining described selection is the flow queue that described needs are aging, reduces the configuration of the queue buffer memory threshold value that the flow queue of described selection is corresponding; In described predetermined period, if aging Indication message corresponding to the flow queue of described selection is continuously 0, recover the configuration of the queue buffer memory threshold value that the flow queue of described selection is corresponding.

Images