WO2017028545A1 - Port queue congestion monitoring method and system - Google Patents

Abstract

Disclosed is a port queue congestion monitoring method, comprising: determining a cache status of a currently polled port queue when it reaches a current monitoring period; determining a first sending status of the port queue when the cache status of the port queue is an abnormal cache state satisfying a preset condition; prolonging the current monitoring period by a preset prolongation duration to determine a second sending status of the port queue when the first sending status of the port queue is an abnormal sending state satisfying a preset condition; and determining that congestion exists in the port queue when the second sending status of the port queue is an abnormal sending state satisfying a preset condition. Also disclosed is a port queue congestion monitoring system. By means of the present invention, congestion monitoring for all port queues on a telecommunications-level router is implemented, and a port queue in which congestion exists is accurately determined in a timely manner, and then the port queue in which congestion exists is processed in a timely manner, thereby reducing negative effects brought by port queue congestion.

Description

Port queue blocking monitoring method and system

The present application claims priority to Chinese Patent Application No. 201510501920.6, filed on Aug. 14, 2015, the entire disclosure of which is hereby incorporated by reference.

Technical field

The present invention relates to the field of network communication technologies, and in particular, to a method and system for monitoring port queue congestion.

Background technique

With the rapid development of network technologies, more and more users are gradually using new services such as IPTV (Internet Protocol Television), VoIP (Voice over Internet Protocol), distance learning, telemedicine, etc. The new business puts higher demands on the service capabilities of the network. When users use a large number of network applications such as IPTV, VoIP, BT (Bit Torrent) download, FTP download, etc., the network traffic of the current network carrier-class router port queue will become very large, which will cause the port-level router to have port queue congestion. risks of. When the port queue is blocked, packets of all user queues under the port cannot be sent, which prevents users from using the network application.

In the current network communication system, the traffic shaping and congestion management functions are implemented by the H-QoS (hierarchical quality of service) queue technology of the traffic management chip hardware (Traffic Management). However, the existing H-QoS queuing technology does not provide a monitoring mechanism for port queue congestion. Therefore, the port queue congestion cannot be monitored in time or even in advance, and the port queue cannot be accurately determined to be blocked, thereby failing to quickly restore the port queue. The message is sent, causing the user to be unable to use the web application for a long time.

Summary of the invention

The main purpose of the present invention is to provide a method and system for monitoring port clogging, which is to solve the problem that the port queue clogging cannot be monitored and accurately determined, and the packet sending of the port queue cannot be quickly restored, so that the user cannot use the network application for a long time. Technical problem.

To achieve the above object, the present invention provides a method for monitoring port queue congestion, which is applied to a carrier-class router, and the router includes a plurality of ports, and the monitoring method for the port queue congestion includes:

When the current monitoring period arrives, it is determined whether the buffer status of the currently polled port queue satisfies a preset abnormal cache state condition;

Determining whether the first sending state of the port queue meets a preset abnormal sending state condition when the buffering state of the port queue meets the abnormal cache state condition;

And when the first sending state of the port queue meets the abnormal sending state, the current monitoring period is extended by a preset delay duration to determine whether the second sending state of the port queue meets the abnormal sending state. condition;

When the second transmission state of the port queue satisfies the abnormal transmission state condition, it is determined that the port queue is blocked.

Preferably, when the current monitoring period arrives, determining whether the buffer status of the currently polled port queue meets a preset abnormal cache state condition includes:

Obtaining a real-time depth value of the port queue when the current monitoring period arrives, where the depth value is used to measure the amount of packet buffered in the port queue;

Determining whether the real-time depth value of the port queue is greater than or equal to a preset depth threshold to determine whether the cache state of the port queue satisfies the abnormal cache state condition;

When the real-time depth value of the port queue is greater than or equal to the depth threshold, determining that the cache status of the port queue is the abnormal cache state.

Preferably, when the buffering state of the port queue satisfies the abnormal cache state condition, determining whether the first sending state of the port queue meets a preset abnormal sending state condition comprises:

When the buffer status of the port queue satisfies the abnormal cache state condition, the first packet count corresponding to the port queue sent in the current monitoring period and in the previous monitoring period is obtained. Two message counts, wherein the message count is counted by using a high and low double counter;

Comparing whether the upper and lower bits of the first packet count are respectively equal to the upper and lower bits of the second packet count to determine whether the first transmission state of the port queue satisfies the abnormal transmission state condition;

When the upper and lower bits of the first packet count are respectively equal to the upper and lower bits of the second packet count, determining that the first transmission state of the port queue is the abnormal transmission state.

Preferably, when the first sending state of the port queue satisfies the abnormal sending state condition, the current monitoring period is extended by a preset delay duration to determine whether the second sending state of the port queue satisfies The abnormal transmission status conditions include:

And when the first sending state of the port queue meets the abnormal sending state condition, extending the current monitoring period by the delay duration;

And acquiring, when the delay duration arrives, a third packet count sent by the port queue in the last monitoring period and the delay duration;

Comparing whether the upper and lower bits of the first packet count are respectively equal to the upper and lower bits of the third packet count to determine that the second transmission state of the port queue satisfies the abnormal transmission state condition;

When the upper and lower bits of the first packet count are respectively equal to the upper and lower bits of the third packet count, determining that the second transmission state of the port queue is the abnormal transmission state.

Preferably, when the second sending state of the port queue satisfies the abnormal sending state condition, determining that the port queue has a jam includes:

When it is detected that there is congestion in the port queue, the port queue is closed;

Clearing the buffered packets in the port queue and retaining the relevant configuration required to reset the port queue parameter;

Resetting the port queue to restore an initial state corresponding to when the port queue does not send a message according to the related configuration parameter that is reserved;

After the port queue is restored to the initial state, the port queue is enabled to enable the port queue to send packets.

Preferably, when the current monitoring period arrives, determining whether the buffer status of the currently polled port queue meets the preset abnormal cache state condition includes:

A monitoring thread of the port queue is created and polling of the port queue jam is performed on all ports on the router in the preset monitoring period.

Further, in order to achieve the above object, the present invention further provides a monitoring system for port queue congestion, which is applied to a carrier-grade router, the router includes a plurality of ports, and the monitoring system for blocking the port queue includes:

a buffer status determining module, configured to determine, when the current monitoring period arrives, whether a buffer status of the currently polled port queue meets a preset abnormal cache status condition;

a first sending state determining module, configured to determine, when the buffering state of the port queue meets the abnormal cache state condition, whether the first sending state of the port queue meets a preset abnormal sending state condition;

a second sending state determining module, configured to: when the first sending state of the port queue meets the abnormal sending state condition, extend the current monitoring period by a preset delay duration to determine a second sending of the port queue Whether the state satisfies the abnormal transmission state condition;

The congestion determination module is configured to determine that the port queue is blocked when the second transmission state of the port queue satisfies the abnormal transmission status condition.

Preferably, the cache status determining module includes:

a depth value obtaining unit, configured to acquire a real-time depth value of the port queue when the current monitoring period arrives, where the depth value is used to measure a buffered amount of packet storage in the port queue;

a depth value judging unit, configured to determine whether the real-time depth value of the port queue is greater than or equal to a preset depth threshold to determine whether the cache state of the port queue satisfies the abnormal cache state condition;

The cache state determining unit is configured to determine, when the real-time depth value of the port queue is greater than or equal to the depth threshold, that the cache state of the port queue is the abnormal cache state.

Preferably, the first sending state determining module includes:

a packet counting first acquiring unit, configured to: when the buffering state of the port queue meets the abnormal buffering state condition, acquire the port queue in the current monitoring period and in the previous monitoring period Corresponding to the first packet count and the second packet count, wherein the packet count is counted by using a high-order and a low-order double counter;

The message counting first comparing unit is configured to compare whether the upper and lower bits of the first packet count are respectively equal to the high and low bits of the second packet count to determine the first sending state of the port queue. Whether the abnormal transmission status condition is satisfied;

a first sending state determining unit, configured to determine, when the upper and lower bits of the first packet count are equal to the high and low bits of the second packet, respectively, determining that the first sending state of the port queue is The abnormal transmission status is described.

Preferably, the second sending state determining module includes:

a delay unit, configured to extend the current monitoring period by the delay duration when the first sending state of the port queue satisfies the abnormal sending state condition;

a second counting unit, configured to acquire, when the delay duration arrives, a third packet count sent by the port queue in the last monitoring period and the delay duration;

The message counting second comparing unit is configured to compare whether the upper and lower bits of the first packet count are respectively equal to the high and low bits of the third packet count to determine that the second sending state of the port queue is satisfied. The abnormal transmission status condition;

a second sending state determining unit, configured to determine, when the upper and lower bits of the first packet count are equal to the high and low bits of the third packet, respectively, determining that the second sending state of the port queue is The abnormal transmission status is described.

Preferably, the monitoring system for blocking the port queue further includes:

a port queue closing module, configured to close the port queue when it is detected that the port queue is blocked;

a buffered message clearing module, configured to clear the buffered message in the port queue and retain relevant configuration parameters required for resetting the port queue;

The port queue reset module is configured to reset the port queue according to the retained related configuration parameter to restore an initial state corresponding to when the port queue does not send a message;

The port queue enabling module is configured to enable the port queue to enable the port queue to send packets after the port queue is restored to the initial state.

Preferably, the monitoring system for blocking the port queue further includes:

The monitoring thread creation module is configured to create a monitoring thread of the port queue and perform polling monitoring of the port queue congestion on all ports on the router in the preset monitoring period.

An embodiment of the present invention further provides a monitoring device for port queue congestion, which is applied to a carrier-grade router including a plurality of ports, including:

processor;

a memory for storing processor executable instructions;

Wherein the processor is configured to:

When the current monitoring period arrives, it is determined whether the buffer status of the currently polled port queue satisfies a preset abnormal cache state condition;

Determining whether the first sending state of the port queue meets a preset abnormal sending state condition when the buffering state of the port queue meets the abnormal cache state condition;

When the first transmission state of the port queue satisfies the abnormal transmission state condition, the current monitoring week will be Extending a preset delay duration to determine whether the second transmission state of the port queue satisfies the abnormal transmission state condition;

When the second transmission state of the port queue satisfies the abnormal transmission state condition, it is determined that the port queue is blocked.

Embodiments of the present invention also provide a non-transitory computer readable storage medium having stored therein instructions that cause the router to implement a port when executed by a processor of a carrier-grade router including a plurality of ports A method of monitoring queue congestion, the method comprising the steps of:

When the current monitoring period arrives, it is determined whether the buffer status of the currently polled port queue satisfies a preset abnormal cache state condition;

Determining whether the first sending state of the port queue meets a preset abnormal sending state condition when the buffering state of the port queue meets the abnormal cache state condition;

And when the first sending state of the port queue meets the abnormal sending state, the current monitoring period is extended by a preset delay duration to determine whether the second sending state of the port queue meets the abnormal sending state. condition;

When the second transmission state of the port queue satisfies the abnormal transmission state condition, it is determined that the port queue is blocked.

The invention monitors all port queue jams by means of periodic polling, so that the port queue jam problem can be locked to one or some port queues more finely, and only the targeted port queues need to be targeted. Perform recovery operations to avoid impact on other normal port queues. In addition, by monitoring the cache status of the port queue, it is possible to determine whether the port queue has a potential congestion risk in time or even in advance, and further monitor the transmission status of the port queue to determine whether there is currently a jam. In addition, when monitoring the transmission status of the port queue, in order to eliminate the misjudgment of the transmission status, the delay is used to further determine whether the current port queue is blocked, thereby making the judgment of the port queue congestion more accurate, thereby enabling Timely and accurately determine the port queues that are blocked, and then timely process the blocked port queues, reducing the negative impact of port queue congestion.

DRAWINGS

1 is a schematic flowchart of a first embodiment of a method for monitoring port port congestion according to the present invention;

2 is a schematic diagram of a refinement process of step S110 in FIG. 1;

3 is a schematic diagram of a refinement process of step S120 in FIG. 1;

4 is a schematic diagram of a refinement process of step S130 in FIG. 1;

5 is a schematic flowchart of a second embodiment of a method for monitoring port port congestion according to the present invention;

FIG. 6 is a schematic flowchart diagram of a third embodiment of a method for monitoring port port congestion according to the present invention; FIG.

7 is a schematic diagram of functional modules of a first embodiment of a monitoring system for port queue congestion according to the present invention;

8 is a schematic diagram of a refinement function module of the cache state determination module in FIG. 7;

9 is a schematic diagram of a refinement function module of the first transmission state determining module in FIG. 7;

10 is a schematic diagram of a refinement function module of the second transmission state determining module in FIG. 7;

11 is a schematic diagram of functional modules of a second embodiment of a monitoring system for port queue congestion according to the present invention;

FIG. 12 is a schematic diagram of functional modules of a third embodiment of a monitoring system for port queue congestion according to the present invention.

The implementation, functional features, and advantages of the present invention will be further described in conjunction with the embodiments.

detailed description

It is understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

Referring to FIG. 1, FIG. 1 is a schematic flowchart diagram of a first embodiment of a method for monitoring port queue congestion according to the present invention. In this embodiment, the monitoring method for the port queue congestion includes:

Step S110: When the current monitoring period arrives, determine whether the buffer status of the currently polled port queue satisfies a preset abnormal cache state condition;

In this embodiment, the monitoring method of the port queue jam is applied to the carrier-class router, wherein the port queue corresponds to the H-QoS in the downlink (Traffic Management) chip hardware H-QoS queue technology of the current network carrier-class router. There are also several layers of middle-tier queues between the root node of the tree, the port queue, and the user queue. User packets are sent to the middle-tier queue through the user queue, and then the intermediate-layer queues are aggregated to the port queue. Finally, the port queue is sent to the corresponding interface daughter card port, and the interface daughter card port can be externally connected to the fiber.

In addition, routers typically include several ports, so all ports on the router need to be monitored. In this embodiment, the polling mode is used to monitor all the ports, and the monitoring period is to complete all rounds for all ports, in order to reduce the efficiency of the routers. The time of the inquiry.

It should be further noted that the port queue monitoring period needs to be set according to actual conditions. For example, set to 100 milliseconds. If the monitoring period is set too large, the port queue will not be quickly monitored when it is blocked, which will cause the traffic interruption time to be too long. If it is set too small, the hardware and software overhead brought by the port queue monitoring will increase. This affects the normal operation of the router line card. In addition, the interval between the two adjacent ports is related to the monitoring period. The interval between the two adjacent ports may be the same or different, and may be set according to actual needs. In addition, for convenience of description, the port queue blocking monitoring of any one of the ports is described in the embodiment, and the descriptions of the other ports are the same, and therefore are not described herein.

When the current monitoring period is reached, that is, when polling monitoring is performed on one of the ports on the router, it is determined whether the buffer status of the currently polled port queue satisfies the preset abnormal cache state condition. In this embodiment, the currently polled port queue specifically refers to the port queue that starts polling when the current monitoring period arrives. The cache status of a port queue refers to the state of the amount of packet buffered in the port queue. For example, the cache status is empty, the cache status is not full (that is, it is not blocked), the cache status is full (the risk of congestion), etc. . In this embodiment, the preset cache state condition is used to determine an abnormal cache state, and if the preset condition is met, the mode is an abnormal cache state; On the other hand, if it is not satisfied, it is in the normal buffer state, and will continue to wait for the arrival of the next monitoring period.

Preferably, in order to facilitate timely or even early monitoring of port queue congestion, the abnormal cache state is determined to be in a state of congestion risk, and is also about to enter a cache state at the time of congestion. Therefore, for the abnormal cache state determined above, the preset cache state condition is determined correspondingly, and the condition is specifically set according to actual needs, for example, the condition may be the cache size of the port queue. Therefore, when the buffer status of the currently polled port queue satisfies the foregoing preset condition, at least the current polled port queue may be determined to have a congestion risk.

Step S120: When the buffering state of the port queue meets the abnormal cache state condition, determine whether the first sending state of the port queue meets a preset abnormal sending state condition;

When the buffered state of the polled port queue satisfies the above-mentioned preset abnormal cache state condition, that is, when at least the current polled port queue has a congestion risk, it is necessary to further determine whether there is a jam, that is, through the port. The determination is made as to whether the first transmission state of the queue satisfies a preset abnormal transmission state condition.

In this embodiment, the sending status of the port queue (the first sending status is only a naming difference) is determined by a preset abnormal sending status condition. If the preset condition is met, the status is abnormal. If it is not satisfied, it is the normal transmission state, and it will continue to wait for the arrival of the next monitoring period.

Preferably, in order to facilitate accurate monitoring of port queue congestion, the abnormal transmission status is determined to be a state in which no message is sent. Therefore, for the abnormal transmission state determined above, a preset transmission state condition is determined correspondingly, and the condition is specifically set according to actual needs, for example, the condition may be a packet count of the port queue. Therefore, when the transmission status of the currently polled port queue satisfies the corresponding preset condition, it is basically determined that there is congestion in the currently polled port queue.

Step S130: When the first sending state of the port queue meets the abnormal sending state condition, the current monitoring period is extended by a preset delay duration to determine whether the second sending state of the port queue satisfies the non- Normal transmission status condition;

Step S140: When the second sending state of the port queue satisfies the abnormal sending state condition, it is determined that the port queue is blocked.

When the first transmission status of the polled port queue meets the preset abnormal transmission status, there may be a misjudgment of the port queue congestion. For example, there may be a packet entering the port at the critical moment when the monitoring period arrives. Queue, but the message has not been sent out from the port queue at this time, so the message is not counted, resulting in a false positive.

Therefore, based on the above-mentioned misidentification of the port clogging, in the embodiment, the current monitoring period is extended by the preset delay duration, thereby further accurately determining the transmission status of the currently polled port queue (the second transmission state is only It is only the naming difference) whether the preset abnormal transmission status condition is met. If the preset abnormal transmission status condition is met again, the transmission status of the currently polled port queue can be accurately determined as not sending a message, that is, it is determined. The current polled port queue is blocked. On the contrary, it is determined that there is no congestion in the current polled port queue. At this time, it will continue to wait for the arrival of the next monitoring period.

In this embodiment, all port queue jams are monitored by means of periodic polling, so that the port queue jam problem can be locked to one or some port queues more finely, and only the targeted lock is needed. Port The queue performs recovery operations to avoid impact on other normal port queues. In addition, by monitoring the cache status of the port queue, it is possible to determine whether the port queue has a potential congestion risk in time or even in advance, and further monitor the transmission status of the port queue to determine whether there is currently a jam. In addition, when monitoring the transmission status of the port queue, in order to eliminate the misjudgment of the transmission status, the delay is used to further determine whether the current port queue is blocked, thereby further improving the accuracy of the port queue clogging judgment, thereby enabling In order to accurately and accurately determine the port queues that are blocked, the port queues that are blocked are processed in time to reduce the negative impact caused by port queue congestion.

Referring to FIG. 2, FIG. 2 is a schematic diagram of the refinement process of step S110 in FIG. Based on the above embodiment, in the embodiment, the foregoing step S110 includes:

In step S1101, when the current monitoring period arrives, the real-time depth value of the port queue is obtained, where the depth value is used to measure the amount of packet buffered in the port queue.

Step S1102: Determine whether the real-time depth value of the port queue is greater than or equal to a preset depth threshold to determine whether the cache state of the port queue satisfies the abnormal cache state condition;

Step S1103: When the real-time depth value of the port queue is greater than or equal to the depth threshold, determine that the cache state of the port queue is the abnormal cache state.

In this embodiment, the depth value of the port queue is used to measure the amount of the buffered packets in the port queue, that is, the real-time depth value of the port queue is obtained, and the real-time depth value of the port queue is compared with the preset depth threshold. Therefore, it is used to determine the buffer status of the port queue, that is, whether a certain number of packets to be sent are buffered in the port queue.

It should be noted that the preset depth threshold must be smaller than the port queue cache. Otherwise, when the port queue is blocked, the real-time depth value of the port queue does not exceed the preset depth threshold. At the same time, the preset depth threshold cannot be set too large. Otherwise, when the port queue is blocked, it cannot be quickly monitored in time, which causes the traffic interruption time to be too long. Therefore, in this embodiment, the preset depth threshold is specifically set according to an actual situation, for example, set to a kilobyte level. In addition, it should be further explained that the port queue cache of all ports on the general router is the default allocation, and the default allocated cache size is the same, therefore, for multiple port queues on the router, only one preset is needed. The depth threshold is OK.

In this embodiment, the abnormal cache state condition is preferably that the real-time depth value of the port queue is greater than or equal to a preset depth threshold. If the real-time depth of the port queue is less than the preset threshold, the port queue is in the normal cache state, that is, the port queue is not blocked. Otherwise, if the real-time depth value is greater than or equal to the preset depth threshold, the port queue cached packets are The storage capacity exceeds the preset depth threshold, that is, the port queue is abnormally cached, and the port queue is at risk of congestion. By setting the preset depth threshold, you can monitor the risk of port queue congestion in time, and then intervene in the port queue with the risk of congestion to minimize the impact of port congestion.

Referring to FIG. 3, FIG. 3 is a schematic diagram of the refinement process of step S120 in FIG. Based on the above embodiment, in this implementation In the example, step S120 includes:

Step S1201: When the buffering status of the port queue meets the abnormal cache state, the first packet sent by the port queue in the current monitoring period and in the previous monitoring period is obtained. Counting and counting the second message, wherein the message count is counted by using a high and low double counter;

When it is determined that the buffer status of the currently polled port queue satisfies the abnormal cache state condition, that is, when the risk of congestion of the currently polled port queue is determined, it is necessary to further determine whether the port queue blockage does exist, that is, further Determine whether the sending status of the currently polled port queue meets the preset abnormal sending status condition.

In this embodiment, the abnormal transmission state is preferably that the number of packets sent in the two adjacent monitoring periods is equal, that is, no packets are sent in the two adjacent monitoring periods. The two adjacent monitoring periods specifically refer to the current monitoring period and the previous monitoring period of the current monitoring period. Specifically, the first packet count sent in the current monitoring period (assumed to be s1) and the second sent in the previous monitoring period are read from the preset dual counter for counting the statistics in a non-reading manner. Message count (assumed to be s0). The non-reading mode means that the counting data is not cleared when the counting data is read.

In this embodiment, for the TM chip that supports the port queue to send the packet count, the double counter used for counting the statistics of the packet directly obtains the count of the packet sent by the port queue provided by the hardware of the TM chip; The TM chip that counts the packets is sent to the corresponding port on the interface sub-card. Therefore, in this embodiment, the dual-counter for counting the statistics is obtained. The number of packets received by the corresponding port on the interface subcard corresponding to the port queue is also counted as the number of packets sent by the port queue.

Optionally, in view of the fact that the existing single counter has the problem that the count of two consecutively transmitted messages overflows and falls over a short period of time, it is preferable to determine whether the counts of the messages sent in the adjacent two monitoring periods are equal. The high and low double counters count the message.

The larger the counter width used by the port queue to send the message count, the smaller the probability that the sent message count overflows and flips during a monitoring period, and the more accurate the comparison is. In order to accurately compare the number of packets sent by the port queue, the present embodiment specifically uses the high-order and low-level dual counters to count the packets sent by the port queue. For example, when using an 8-bit dual counter, the double counter is the upper 8 bits and the lower 8 bits respectively; when using the 16-bit double counter, the double counter is the upper 16 bits and the lower 16 bits respectively; when using the 32-bit double counter, the double The counters are high 32 bits and low 32 bits, respectively. When the low-order counter is full, the high-order counter count is incremented by one, and the low-order counter count is cleared. When the high-order counter is full, the high-order counter is cleared.

Step S1202: Comparing whether the upper and lower bits of the first packet count are respectively equal to the upper and lower bits of the second packet count to determine whether the first transmission state of the port queue satisfies the abnormal transmission. State condition

Step S1203: When the upper and lower bits of the first packet count are respectively equal to the high and low bits of the second packet count, determining that the first sending state of the port queue is the abnormal sending state.

When the first packet count s1 sent in the current monitoring period and the second sent in the previous monitoring period are obtained After the s1 is compared with the s0, the s1 is compared with the s0 to determine the sending status of the port queue, that is, whether the port queue is transmitting packets. If the port queue is transmitting packets, the status is normal. Otherwise, Abnormal transmission status.

In order to avoid the misjudgment caused by the overflow of the message count overflow, the size of s1 and s0 is not compared in this embodiment, but it is preferable to compare whether s1 and s0 are equal. Since the message count in the embodiment is counted by the high-order and low-order double counters, it is necessary to respectively compare whether the upper or lower bits of s1 and s0 are equal: if the upper bits corresponding to s1 and s0 are not equal and/or the low bits are not equal If the s1 and s0 respectively correspond to the high-order bits and the low-order bits are equal, the port queue does not send packets in the two consecutive counting times.

In this embodiment, the abnormal transmission status condition is that the port queue does not send a message within two consecutive counting times, that is, no message is sent in two adjacent monitoring periods. By judging the buffer status of the port queue, the port queue with the risk of blocking is locked; and the status of the port queue is judged to further accurately determine the blocked port queue. The double-layer progressive monitoring mechanism of the buffer state and the transmission state used in this embodiment can lock the blocked port queues in a timely and accurate manner, thereby enabling timely intervention, thereby minimizing the impact caused by port queue congestion. .

In addition, in the embodiment, the high-order and low-level dual counters are used to count the packets sent by the port queue, which solves the problem that the use of the single counter has the overflow of the two consecutively transmitted message counts in a short time, thereby enabling more accurate Determine whether the counts of two consecutively sent messages are equal. At the same time, in this embodiment, the number of received packets of the corresponding port of the interface subcard is obtained, and the count is used as the port queue to send the packet count, thereby solving the problem that the TM chip hardware cannot support the port queue to send the packet count and cannot determine the port queue sending. State problem. In addition, it is preferable to respectively assign s1 and s0 as the message counts sent in the current monitoring period and the parameters of the message counts sent in the previous monitoring period, and then assign the count of s1 to s0 at the end of the current monitoring period. The comparison of the transmitted message counts is performed in the next monitoring cycle.

Referring to FIG. 4, FIG. 4 is a schematic diagram of the refinement process of step S130 in FIG. Based on the above embodiment, in this embodiment, step S130 includes:

Step S1301: When the first sending state of the port queue satisfies the abnormal sending state condition, extending the current monitoring period by the delay duration;

When the transmission status of the polled port queue meets the abnormal transmission status condition, that is, when the port queue does not send packets in the two adjacent monitoring periods, the port queue cannot be determined to be blocked and cannot send packets. For example, when a packet enters the port queue at the critical time when the monitoring period arrives, the packet is not sent out from the port queue at this time (that is, the counter does not count at this time). Therefore, the packet may be sent during the current monitoring period. The count s1 is equal to the s0 sent in the last monitoring period, that is, the port queue is not sent, and if the port queue is not blocked, but the judgment result of the unsent packet is obtained, the adjacent two are generated. No misjudgement of the message was sent during the monitoring period.

In this embodiment, based on the above-mentioned misidentification problem of port queue congestion, the delay time T1 is introduced to advance. After the row counts, the comparison of the counts of the adjacent two monitoring period messages is performed. The delay time T1 introduced must be smaller than the monitoring period T0 of the port queue. Otherwise, after the delay time T1 arrives and the port queue sends a message count, the time has reached the next monitoring period, instead of the current monitoring period. There is no guarantee that the count of sent messages will be compared during the same monitoring period.

In addition, it is preferable that the product of the port queue delay duration T1 and the number of ports polled in the monitoring period T0 must be smaller than the port queue monitoring period T0. Otherwise, if all the ports are delayed by T1, at least one port is acquired. After the port queue sends the packet count, the time has reached the next monitoring period, instead of the current monitoring period. Therefore, it is impossible to compare the number of sent packets in the same monitoring period.

Step S1302: When the delay duration arrives, acquire a third packet count sent by the port queue in the last monitoring period and the delay duration;

Step S1303: Comparing whether the upper and lower bits of the first packet count are respectively equal to the upper and lower bits of the third packet count to determine that the second transmission state of the port queue satisfies the abnormal transmission state. condition;

Step S1304: When the upper and lower bits of the first packet count are respectively equal to the high and low bits of the third packet count, determine that the second transmission state of the port queue is the abnormal transmission state.

When the preset delay duration arrives, the packet received by the corresponding port on the interface subcard corresponding to the port queue is read in the non-reading mode. The third packet count s2 is The sum of the sent packet count s1 and the sent packet count in the last monitoring period, by comparing whether the upper and lower bits of s1 and s2 are respectively equal, that is, determining whether the sent packet count is zero within the delay duration, thereby The previous judgment results are further determined to avoid misjudgment of port queue congestion. If the transmission status of the port queue still meets the preset abnormal transmission status after the delay processing, that is, if the packet is still not sent, the current polled port queue can be accurately determined to be blocked.

In this embodiment, in order to avoid the misjudgment caused by the packet entering the port queue at the critical time when the monitoring period arrives, the delay count is introduced to further judge the sending state of the extended time, thereby achieving accurate locking and blocking. The port queue improves the accuracy of the judgment that the port queue is blocked.

Referring to FIG. 5, FIG. 5 is a schematic flowchart diagram of a second embodiment of a method for monitoring port queue congestion according to the present invention. In this embodiment, after step S140,

Step S210, when it is detected that there is congestion in the port queue, the port queue is closed;

Step S220, clearing the buffered message in the port queue and retaining relevant configuration parameters required for resetting the port queue.

Step S230, resetting the port queue according to the retained related configuration parameter to restore an initial state corresponding to when the port queue does not send a message;

In step S240, after the port queue is restored to the initial state, the port queue is enabled to enable the port queue to send packets.

In this embodiment, when the polling port queue is blocked, the port queue is closed to prevent the packet from entering the port queue. At the same time, all the packets cached in the port queue are cleared, but the related configuration parameters, such as the priority and weight of the packet queue, and the rate limit policy for sending packets, are reserved.

After clearing the packets buffered in the blocked port queue and retaining the relevant configuration parameters required for resetting the port queue, the port queue is reset according to the reserved configuration parameters to restore the unsent packets to the port queue. The initial state corresponding to the time, and the port queue is activated by the enable mode, so that the port queue is opened again to send the packet.

After the traffic queues of the carrier-class routers are blocked, the common processing methods are as follows: First, insert or remove the interface subcard corresponding to the blocked port queue. The negative effect is that the number of the normal ports corresponding to the inserted interface subcards is ten. The second-level traffic is interrupted; the second is to plug or unplug the line card. The negative effect is that the traffic of the line card is interrupted for several minutes on all normal ports. The above operations on the insertion and removal of the interface daughter card and the line card will cause long-term traffic interruption of other normal port queues.

In this embodiment, by periodically monitoring the buffering and sending status of the port queue, when the port queue is blocked, only the blocked port queue is restored, thereby preventing the insertion and removal of the interface daughter card and the line card. The operation reduces the interruption time of the user traffic from tens of seconds to several seconds, and also reduces the impact range of the port queue congestion fault, and improves the stability of the carrier-class router.

Referring to FIG. 6, FIG. 6 is a schematic flowchart diagram of a third embodiment of a method for monitoring port queue congestion according to the present invention. In this embodiment, before step S110,

Step S001: Create a monitoring thread of the port queue and perform polling monitoring of the port queue congestion on all ports on the router in the preset monitoring period.

In this embodiment, after the router starts and the port queue is created, when the packet sending of the port queue is started, the monitoring thread of the port queue is created correspondingly, and the monitoring thread is also configured to perform polling monitoring of the port queue congestion of all the ports. The monitoring period, as well as setting the depth threshold used to determine the port queue buffer status and setting the delay duration for further determining the port queue transmission status.

In this embodiment, by monitoring the thread of the port queue and using the polling monitoring mode, the monitoring of the congestion of all the port queues is realized in real time, so that the blocked port queue can be determined more timely and accurately, and the solution is solved in a targeted manner. Blocked port queues, which reduce the impact of port queue congestion and reduce the interruption time of user traffic provide an effective solution.

Referring to FIG. 7, FIG. 7 is a schematic diagram of functional modules of a first embodiment of a monitoring system for port queue congestion according to the present invention. In this embodiment, the monitoring system for blocking the port queue includes:

The buffer status determining module 10 is configured to determine, when the current monitoring period arrives, whether the buffer status of the currently polled port queue meets a preset abnormal cache status condition;

When the current monitoring period arrives, it starts polling monitoring of one of all ports on the router. The cache state determination module 10 determines whether the cache state of the currently polled port queue satisfies a preset abnormal cache state condition. In this embodiment, the currently polled port queue specifically refers to the port queue that starts polling when the current monitoring period arrives. The cache status of a port queue is specifically the state of the amount of packet buffered in the port queue. For example, the cache status is empty, the cache status is not full (that is, it is not blocked), and the cache status is full (that is, the risk of congestion). )Wait. In this embodiment, the preset cache state is used to determine an abnormal cache state. If the preset condition is met, the state is abnormal. If not, the buffer state is normal. Continue to wait for the arrival of the next monitoring cycle.

Preferably, in order to facilitate timely or even early monitoring of port queue congestion, the abnormal cache state is determined to be in a state of congestion risk, and is also about to enter a cache state at the time of congestion. Therefore, for the abnormal cache state determined above, the preset cache state condition is determined correspondingly, and the condition is specifically set according to actual needs, for example, the condition may be the buffer size of the port queue or the size of the received message volume. Wait. Therefore, when the buffer status of the currently polled port queue satisfies the foregoing preset condition, at least the current polled port queue may be determined to have a congestion risk.

The first sending state determining module 20 is configured to determine, when the buffering state of the port queue meets the abnormal cache state condition, whether the first sending state of the port queue meets a preset abnormal sending state condition;

When the buffered state of the polled port queue satisfies the above-mentioned preset abnormal cache state condition, that is, when at least the current polled port queue may be determined to have a jam risk, the first sending state determining module 20 needs to further determine whether it is true. There is congestion, that is, whether the first transmission state of the port queue satisfies a preset abnormal transmission state condition.

In this embodiment, the sending status of the port queue (the first sending status is only a naming difference) is determined by a preset abnormal sending status condition. If the preset condition is met, the status is abnormal. If it is not satisfied, it is the normal transmission state, and it will continue to wait for the arrival of the next monitoring period.

Preferably, in order to facilitate accurate monitoring of port queue congestion, the abnormal transmission status is determined to be a state in which no message is sent. Therefore, for the abnormal transmission state determined above, the preset transmission state condition is determined correspondingly, and the condition is set according to actual needs, for example, the condition may be the number of packets in the port queue or the number of packets sent in a unit time. Wait. Therefore, when the transmission status of the currently polled port queue satisfies the corresponding preset condition, it is basically determined that there is congestion in the currently polled port queue.

The second sending state determining module 30 is configured to: when the first sending state of the port queue meets the abnormal sending state condition, extend the current monitoring period by a preset delay duration to determine the second queue queue. Whether the sending status satisfies the abnormal sending status condition;

The congestion determination module 40 is configured to determine that the port queue is blocked when the second transmission state of the port queue satisfies the abnormal transmission status condition.

When the first transmission status of the polled port queue meets the preset abnormal transmission status, there may be a misjudgment of the port queue congestion. For example, there may be a packet entering the port at the critical moment when the monitoring period arrives. Queue, but the message has not been sent out from the port queue at this time, so the message is not counted, resulting in a false positive.

Therefore, based on the above-mentioned misjudgment of port clogging that may exist, in this embodiment, the second transmission state determination mode Block 30 further extends the current monitoring period by a preset delay duration, thereby further accurately determining whether the currently polled port queue transmission state (the second transmission state is merely a naming difference) satisfies a preset abnormal transmission state condition, and if The congestion determination module 40 can accurately determine that the current polled port queue is sent without sending a message, that is, it determines that the current polled port queue is blocked; There is no congestion in the polled port queue, and it will continue to wait for the arrival of the next monitoring period.

In this embodiment, all port queue jams are monitored by means of periodic polling, so that the port queue jam problem can be locked to one or some port queues more finely, and only the targeted lock is needed. The port queue is restored to avoid the impact on other normal port queues. In addition, by monitoring the cache status of the port queue, it is possible to determine whether the port queue has a potential congestion risk in time or even in advance, and further monitor the transmission status of the port queue to determine whether there is currently a jam. In addition, when monitoring the transmission status of the port queue, in order to eliminate the misjudgment of the transmission status, the delay is used to further determine whether the current port queue is blocked, thereby further improving the accuracy of the port queue clogging judgment, thereby enabling In order to accurately and accurately determine the port queues that are blocked, the port queues that are blocked are processed in time to reduce the negative impact caused by port queue congestion.

Referring to FIG. 8, FIG. 8 is a schematic diagram of a refinement function module of the cache state determination module of FIG. Based on the foregoing embodiment, in the embodiment, the cache state determining module 10 includes:

The depth value obtaining unit 101 is configured to acquire a real-time depth value of the port queue when the current monitoring period arrives, where the depth value is used to measure a buffer size of the port queue;

The depth value determining unit 102 is configured to determine whether the real-time depth value of the port queue is greater than or equal to a preset depth threshold to determine whether the cache state of the port queue satisfies the abnormal cache state condition;

The cache state determining unit 103 is configured to determine, when the real-time depth value of the port queue is greater than or equal to the depth threshold, that the cache state of the port queue is the abnormal cache state.

In this embodiment, the depth value of the port queue is used to measure the amount of the message buffered in the port queue, that is, the depth value obtaining unit 101 obtains the real-time depth value of the port queue, and the depth value determining unit 102 performs the real-time port queue. The depth value is compared with the preset depth threshold to determine the buffer status of the port queue, that is, whether a certain number of packets to be sent are buffered in the port queue. And if the real-time depth value of the port queue is greater than or equal to the depth threshold, the cache state determining unit 103 determines that the cache state of the port queue satisfies the abnormal cache state condition.

It should be noted that the preset depth threshold must be smaller than the port queue cache. Otherwise, when the port queue is blocked, the real-time depth value of the port queue does not exceed the preset depth threshold. At the same time, the preset depth threshold cannot be set too large. Otherwise, when the port queue is blocked, it cannot be quickly monitored in time, which causes the traffic interruption time to be too long. Therefore, in this embodiment, the preset depth threshold is specifically set according to an actual situation, for example, set to a kilobyte level. In addition, it should be further explained that the port queue cache of all ports on the general router is the default allocation, and the default allocated cache size is the same, therefore, for multiple port queues on the router, only one preset is needed. The depth threshold is OK.

In this embodiment, the abnormal cache state condition is preferably that the real-time depth value of the port queue is greater than or equal to a preset depth. Threshold. If the real-time depth of the port queue is less than the preset threshold, the port queue is in the normal cache state, that is, the port queue is not blocked. Otherwise, if the real-time depth value is greater than or equal to the preset depth threshold, the port queue caches the packet. The storage capacity exceeds the preset depth threshold, which causes the port queue to be blocked, that is, the port queue is in an abnormal cache state. By setting the preset depth threshold, you can monitor the risk of port queue congestion in time, and then intervene in the port queue with the risk of congestion to minimize the impact of port congestion.

Referring to FIG. 9, FIG. 9 is a schematic diagram of a refinement function module of the first transmission state determining module in FIG. Based on the foregoing embodiment, in the embodiment, the first sending state determining module 20 includes:

The packet counting first obtaining unit 201 is configured to: when the buffering state of the port queue meets the abnormal cache state condition, acquire the port queue in the current monitoring period and in the previous monitoring period Corresponding to the first message count and the second message count respectively sent, wherein the message count is counted by using a high-order and a low-level double counter;

When it is determined that the buffer status of the currently polled port queue satisfies the abnormal cache state condition, that is, when the risk of congestion of the currently polled port queue is determined, it is necessary to further determine whether the port queue blockage does exist, that is, further Determine whether the sending status of the currently polled port queue meets the preset abnormal sending status condition.

In this embodiment, the abnormal transmission state is preferably that the number of packets sent in the two adjacent monitoring periods is equal, that is, no packets are sent in the two adjacent monitoring periods. The two adjacent monitoring periods specifically refer to the current monitoring period and the previous monitoring period of the current monitoring period. Specifically, the first acquiring unit 201 reads the first packet count (assumed as s1) and the upper packet sent in the current monitoring period from the preset dual counter for counting the statistics in the non-reading manner. The second message count sent during a monitoring period (assumed to be s0). The non-reading mode means that the counting data is not cleared when the counting data is read.

In this embodiment, for the TM chip that supports the port queue to send the packet count, the double counter used for counting the statistics of the packet directly obtains the count of the packet sent by the port queue provided by the hardware of the TM chip; The TM chip that counts the packets is sent to the corresponding port on the interface sub-card. Therefore, in this embodiment, the dual-counter for counting the statistics is obtained. The number of packets received by the corresponding port on the interface subcard corresponding to the port queue is also counted as the number of packets sent by the port queue.

Optionally, in view of the fact that the existing single counter has the problem that the count of two consecutively transmitted messages overflows and falls over a short period of time, it is preferable to determine whether the counts of the messages sent in the adjacent two monitoring periods are equal. The high and low double counters count the message.

The larger the counter width used by the port queue to send the message count, the smaller the probability that the sent message count overflows and flips during a monitoring period, and the more accurate the comparison is. In order to accurately compare the number of packets sent by the port queue, the present embodiment specifically uses the high-order and low-level dual counters to count the packets sent by the port queue. For example, when using an 8-bit double counter, the double counters are high 8 bits and low 8 bits respectively; when using a 16-bit double counter, the double counters are The upper 16 bits and the lower 16 bits are counted; when the 32-bit double counter is used, the double counters are high 32 bits and low 32 bits respectively. When the low-order counter is full, the high-order counter count is incremented by one, and the low-order counter count is cleared. When the high-order counter is full, the high-order counter is cleared.

The packet counting first comparing unit 202 is configured to compare whether the upper and lower bits of the first packet count are respectively equal to the high and low bits of the second packet count to determine the first sending of the port queue. Whether the state satisfies the abnormal transmission state condition;

The first sending state determining unit 203 is configured to: when the high and low bits of the first packet count are respectively equal to the high and low bits of the second packet count, determine that the first sending state of the port queue is The abnormal transmission state.

After acquiring the first packet count s1 sent in the current monitoring period and the second packet count s0 sent in the previous monitoring period, the packet counting first comparing unit 202 compares s1 and s0 to determine the port. The sending status of the queue determines whether the port queue is sending packets. If the port queue is sending packets, it is in the normal sending state. Otherwise, it is in the abnormal sending state.

In order to avoid the misjudgment caused by the overflow of the message count overflow, the size of s1 and s0 is not compared in this embodiment, but it is preferable to compare whether s1 and s0 are equal. Since the message count in the embodiment is counted by the high-order and low-order double counters, it is necessary to respectively compare whether the upper or lower bits of s1 and s0 are equal: if the upper bits corresponding to s1 and s0 are not equal and/or the low bits are not equal The port queue is sent in the adjacent two counting times; if the upper bits corresponding to s1 and s0 are equal and the lower bits are equal, the first sending state determining unit 203 determines that the port queue is in the adjacent two counting times. No message was sent.

In this embodiment, the abnormal transmission status condition is that the port queue does not send a message within two consecutive counting times, that is, no message is sent in two adjacent monitoring periods. By judging the buffer status of the port queue, the port queue with the risk of blocking is locked; and the status of the port queue is judged to further accurately determine the blocked port queue. The double-layer progressive monitoring mechanism of the buffer state and the transmission state used in this embodiment can lock the blocked port queues in a timely and accurate manner, thereby enabling timely intervention, thereby minimizing the impact caused by port queue congestion. .

In addition, in the embodiment, the high-order and low-level dual counters are used to count the packets sent by the port queue, which solves the problem that the use of the single counter has the overflow of the two consecutively transmitted message counts in a short time, thereby enabling more accurate Determine whether the counts of two consecutively sent messages are equal. At the same time, in this embodiment, the number of received packets of the corresponding port of the interface subcard is obtained, and the count is used as the port queue to send the packet count, thereby solving the problem that the TM chip hardware cannot support the port queue to send the packet count and cannot determine the port queue sending. State problem.

Referring to FIG. 10, FIG. 10 is a schematic diagram of a refinement function module of the second transmission state determining module in FIG. Based on the foregoing embodiment, in the embodiment, the second sending state determining module 30 includes:

The delay unit 301 is configured to extend the current monitoring period by the delay duration when the first sending status of the port queue satisfies the abnormal sending status condition;

When the transmission status of the polled port queue meets the abnormal transmission status condition, that is, when the port queue does not send packets in the two adjacent monitoring periods, the port queue cannot be determined to be blocked and cannot send packets. For example, when a packet enters the port queue at the critical time when the monitoring period arrives, the packet is not sent out from the port queue at this time (that is, the counter does not count at this time). Therefore, the packet may be sent during the current monitoring period. The count s1 is equal to the s0 sent in the last monitoring period, that is, the port queue is not sent, and if the port queue is not blocked, but the judgment result of the unsent packet is obtained, the adjacent two are generated. No misjudgement of the message was sent during the monitoring period.

In this embodiment, based on the above-mentioned misidentification problem of the port queue clogging, the delay unit 301 delays the time length T1 to perform counting and then compares the counts of the adjacent two monitoring period messages. The delay time T1 introduced must be smaller than the monitoring period T0 of the port queue. Otherwise, after the delay time T1 arrives and the port queue sends a message count, the time has reached the next monitoring period, instead of the current monitoring period. There is no guarantee that the count of sent messages will be compared during the same monitoring period.

In addition, it is preferable that the product of the port queue delay duration T1 and the number of ports polled in the monitoring period T0 must be smaller than the port queue monitoring period T0. Otherwise, if all the ports are delayed by T1, at least one port is acquired. After the port queue sends the packet count, the time has reached the next monitoring period, instead of the current monitoring period. Therefore, it is impossible to compare the number of sent packets in the same monitoring period.

The message counting second obtaining unit 302 is configured to: when the delay duration arrives, acquire a third packet count sent by the port queue in the last monitoring period and the delay duration;

The packet counting second comparing unit 303 is configured to compare whether the upper and lower bits of the first packet count are respectively equal to the high and low bits of the third packet count to determine the second sending of the port queue. The state satisfies the abnormal transmission state condition;

The second sending state determining unit 304 is configured to: when the high and low bits of the first packet count are respectively equal to the high and low bits of the third packet count, determine that the second sending state of the port queue is The abnormal transmission state.

When the preset delay duration arrives, the packet count second obtaining unit 302 reads the packet received by the corresponding port on the interface daughter card corresponding to the port queue in a non-reading manner, and at this time, the statistics are collected. The third packet count s2 is specifically the sum of the sent packet count s1 and the delayed packet count in the last monitoring period, and the packet count second comparing unit 303 compares whether the upper and lower bits of s1 and s2 respectively correspond to each other. That is, it is determined whether the count of the transmitted message within the delay time is zero, so that the previous judgment result is further determined to avoid misjudgment of the port queue blockage. If the transmission status of the port queue still meets the preset abnormal transmission status condition after the delay processing, that is, the message is still not sent, the second transmission status determining unit 304 can accurately determine the current polled port queue. Blocked.

In this embodiment, in order to avoid the misjudgment caused by the packet entering the port queue at the critical time when the monitoring period arrives, the delay count is introduced to further judge the sending state of the extended time, thereby achieving accurate locking and blocking. The port queue improves the accuracy of the judgment that the port queue is blocked.

Referring to FIG. 11, FIG. 11 is a schematic diagram of functional modules of a second embodiment of a monitoring system for port queue congestion according to the present invention. In this embodiment, the monitoring system for blocking the port queue further includes:

The port queue closing module 50 is configured to close the port queue when it is detected that the port queue is blocked.

The cache message clearing module 60 is configured to clear the buffered message in the port queue and retain relevant configuration parameters required for resetting the port queue.

The port queue resetting module 70 is configured to reset the port queue according to the retained related configuration parameter to restore an initial state corresponding to when the port queue does not send a message;

The port queue enable module 80 is configured to enable the port queue to enable the port queue to send packets after the port queue is restored to the initial state.

In this embodiment, when the port queue of the polled port is blocked, the port queue closing module 50 closes the port queue to prevent the packet from entering the port queue. At the same time, the buffered message clearing module 60 clears all the messages buffered in the port queue but retains the relevant configuration parameters required for resetting the port queue, such as the priority and weight of the packet scheduling in the port queue; Speed limit strategy, etc.

After clearing the buffered message in the blocked port queue and retaining the relevant configuration parameters required for resetting the port queue, the port queue reset module 70 will reset the port queue to restore the port according to the reserved configuration parameters. When the queue does not send a packet, the port queue enable module 80 activates the port queue by enabling the port queue to enable the port queue to send the packet again.

After the traffic queues of the carrier-class routers are blocked, the common processing methods are as follows: First, insert or remove the interface subcard corresponding to the blocked port queue. The negative effect is that the number of the normal ports corresponding to the inserted interface subcards is ten. The second-level traffic is interrupted; the second is to plug or unplug the line card. The negative effect is that the traffic of the line card is interrupted for several minutes on all normal ports. The above operations on the insertion and removal of the interface daughter card and the line card will cause long-term traffic interruption of other normal port queues.

In this embodiment, by periodically monitoring the buffering and sending status of the port queue, when the port queue is blocked, only the blocked port queue is restored, thereby preventing the insertion and removal of the interface daughter card and the line card. The operation reduces the interruption time of the user traffic from tens of seconds to several seconds, and also reduces the impact range of the port queue congestion fault, and improves the stability of the carrier-class router.

Referring to FIG. 12, FIG. 12 is a schematic diagram of functional modules of a third embodiment of a monitoring system for port queue congestion according to the present invention. In this embodiment, the monitoring system for blocking the port queue further includes:

The monitoring thread creation module 90 is configured to create a monitoring thread of the port queue and perform polling monitoring of the port queue congestion on all ports on the router in the preset monitoring period.

In this embodiment, when the packet is started after the router starts and the port queue is created, the monitoring thread creation module 90 creates a monitoring thread of the port queue, and the monitoring thread creation module 90 also sets the monitoring thread to all. The monitoring period of the polling monitoring of the port queue blocking, and the setting is used to determine the port team. The depth threshold for the column cache state and the length of the delay that is set to further determine the port queue transmission status.

In this embodiment, by monitoring the thread of the port queue and using the polling monitoring mode, the monitoring of the congestion of all the port queues is realized in real time, so that the blocked port queue can be determined more timely and accurately, and the solution is solved in a targeted manner. Blocked port queues, which reduce the impact of port queue congestion and reduce the interruption time of user traffic provide an effective solution.

An embodiment of the present invention further provides a monitoring device for port queue congestion, which is applied to a carrier-grade router including a plurality of ports, including:

processor;

a memory for storing processor executable instructions;

Wherein the processor is configured to:

When the current monitoring period arrives, it is determined whether the buffer status of the currently polled port queue satisfies a preset abnormal cache state condition;

Determining whether the first sending state of the port queue meets a preset abnormal sending state condition when the buffering state of the port queue meets the abnormal cache state condition;

And when the first sending state of the port queue meets the abnormal sending state, the current monitoring period is extended by a preset delay duration to determine whether the second sending state of the port queue meets the abnormal sending state. condition;

When the second transmission state of the port queue satisfies the abnormal transmission state condition, it is determined that the port queue is blocked.

Embodiments of the present invention also provide a non-transitory computer readable storage medium having stored therein instructions that cause the router to implement a port when executed by a processor of a carrier-grade router including a plurality of ports A method of monitoring queue congestion, the method comprising the steps of:

When the current monitoring period arrives, it is determined whether the buffer status of the currently polled port queue satisfies a preset abnormal cache state condition;

Determining whether the first sending state of the port queue meets a preset abnormal sending state condition when the buffering state of the port queue meets the abnormal cache state condition;

And when the first sending state of the port queue meets the abnormal sending state, the current monitoring period is extended by a preset delay duration to determine whether the second sending state of the port queue meets the abnormal sending state. condition;

When the second transmission state of the port queue satisfies the abnormal transmission state condition, it is determined that the port queue is blocked.

The above is only a preferred embodiment of the present invention, and thus does not limit the scope of the patent of the present invention. The equivalent structure or equivalent process transformation of the contents of the specification and the drawings, or directly or indirectly applied to other related technical fields, are included in the scope of patent protection of the present invention.

Industrial applicability

The monitoring method of the port queue jam in the present application can be applied to a carrier-class router, and all port queue jams are monitored by periodic polling, so that the port queue jam problem can be locked to some or some more finely. The port queue, in turn, only needs to recover the locked port queue in a targeted manner to avoid affecting the normal port queue.

Claims (13)

A monitoring method for port queue congestion is applied to a carrier-class router, where the router includes a plurality of ports, wherein the monitoring method for blocking the port queue includes: When the current monitoring period arrives, it is determined whether the buffer status of the currently polled port queue satisfies a preset abnormal cache state condition; Determining whether the first sending state of the port queue meets a preset abnormal sending state condition when the buffering state of the port queue meets the abnormal cache state condition; And when the first sending state of the port queue meets the abnormal sending state, the current monitoring period is extended by a preset delay duration to determine whether the second sending state of the port queue meets the abnormal sending state. condition; When the second transmission state of the port queue satisfies the abnormal transmission state condition, it is determined that the port queue is blocked. The method for monitoring the congestion of a port queue according to claim 1, wherein, when the current monitoring period arrives, determining whether the buffer status of the currently polled port queue satisfies a preset abnormal cache state condition includes: Obtaining a real-time depth value of the port queue when the current monitoring period arrives, where the depth value is used to measure the amount of packet buffered in the port queue; Determining whether the real-time depth value of the port queue is greater than or equal to a preset depth threshold to determine whether the cache state of the port queue satisfies the abnormal cache state condition; When the real-time depth value of the port queue is greater than or equal to the depth threshold, determining that the cache status of the port queue is the abnormal cache state. The method for monitoring the congestion of a port queue according to claim 1 or 2, wherein when the buffer state of the port queue satisfies the abnormal cache state condition, it is determined whether the first transmission state of the port queue is satisfied. The preset abnormal transmission status conditions include: When the buffer status of the port queue satisfies the abnormal cache state condition, the first packet count corresponding to the port queue sent in the current monitoring period and in the previous monitoring period is obtained. Two message counts, wherein the message count is counted by using a high and low double counter; Comparing whether the upper and lower bits of the first packet count are respectively equal to the upper and lower bits of the second packet count to determine whether the first transmission state of the port queue satisfies the abnormal transmission state condition; When the upper and lower bits of the first packet count are respectively equal to the upper and lower bits of the second packet count, determining that the first transmission state of the port queue is the abnormal transmission state. The method for monitoring port queue congestion according to claim 3, wherein said number of said port queues When a sending status meets the abnormal sending status condition, extending the current monitoring period by a preset delay duration to determine whether the second sending status of the port queue satisfies the abnormal sending status condition includes: And when the first sending state of the port queue meets the abnormal sending state condition, extending the current monitoring period by the delay duration; And acquiring, when the delay duration arrives, a third packet count sent by the port queue in the last monitoring period and the delay duration; Comparing whether the upper and lower bits of the first packet count are respectively equal to the upper and lower bits of the third packet count to determine that the second transmission state of the port queue satisfies the abnormal transmission state condition; When the upper and lower bits of the first packet count are respectively equal to the upper and lower bits of the third packet count, determining that the second transmission state of the port queue is the abnormal transmission state. The method for monitoring the congestion of a port queue according to claim 4, wherein when the second transmission state of the port queue satisfies the abnormal transmission state condition, after determining that the port queue is blocked, the method includes: When it is detected that there is congestion in the port queue, the port queue is closed; Clearing the buffered message in the port queue and retaining relevant configuration parameters required for resetting the port queue; Resetting the port queue to restore an initial state corresponding to when the port queue does not send a message according to the related configuration parameter that is reserved; After the port queue is restored to the initial state, the port queue is enabled to enable the port queue to send packets. The method for monitoring the congestion of a port queue according to claim 5, wherein, when the current monitoring period arrives, determining whether the buffer status of the currently polled port queue satisfies a preset abnormal cache state condition includes: A monitoring thread of the port queue is created and polling of the port queue jam is performed on all ports on the router in the preset monitoring period. A monitoring system for port queue congestion is applied to a carrier-grade router, and the router includes a plurality of ports, wherein the monitoring system for blocking the port queue includes: The buffer status determining module is configured to determine, when the current monitoring period arrives, whether the buffer status of the currently polled port queue meets a preset abnormal cache status condition; a first sending state determining module, configured to determine, when the buffering state of the port queue meets the abnormal cache state condition, whether the first sending state of the port queue meets a preset abnormal sending state condition; a second sending state determining module, configured to: when the first sending state of the port queue meets the abnormal sending state condition, extend the current monitoring period by a preset delay duration to determine a second sending of the port queue Whether the state satisfies the abnormal transmission state condition; The congestion determination module is configured to determine that the port queue is blocked when the second transmission state of the port queue satisfies the abnormal transmission status condition. The monitoring system for port queue clogging according to claim 7, wherein the cache status determining module comprises: a depth value obtaining unit, configured to acquire a real-time depth value of the port queue when the current monitoring period arrives, where the depth value is used to measure a buffered amount of packet storage in the port queue; The depth value determining unit is configured to determine whether the real-time depth value of the port queue is greater than or equal to a preset depth threshold to determine whether the cache state of the port queue satisfies the abnormal cache state condition; The buffer status determining unit is configured to determine that the cache status of the port queue is the abnormal cache status when the real-time depth value of the port queue is greater than or equal to the depth threshold. The monitoring system of the port queue clogging according to claim 7 or 8, wherein the first transmission state determining module comprises: The packet counting first obtaining unit is configured to: when the buffering state of the port queue satisfies the abnormal buffering state condition, obtain the port queue in the current monitoring period and in the previous monitoring period respectively Corresponding to the first packet count and the second packet count, wherein the packet count is counted by using a high-order and a low-order double counter; The packet counting first comparing unit is configured to compare whether the upper and lower bits of the first packet count are respectively equal to the high and low bits of the second packet count to determine the first sending state of the port queue. Whether the abnormal transmission status condition is satisfied; The first sending state determining unit is configured to: when the high and low bits of the first packet count are respectively equal to the high and low bits of the second packet count, determine that the first sending state of the port queue is The abnormal transmission status is described. The monitoring system of the port queue clogging according to claim 9, wherein the second transmission state determining module comprises: a delay unit, configured to extend the current monitoring period by the delay duration when the first transmission state of the port queue satisfies the abnormal transmission state condition; The second counting unit is configured to: when the delay duration arrives, obtain a third packet count sent by the port queue in the last monitoring period and the delay duration; The packet counting second comparing unit is configured to compare whether the upper and lower bits of the first packet count are respectively equal to the high and low bits of the third packet count to determine the second sending state of the port queue. Satisfying the abnormal transmission state condition; a second transmission state determining unit, configured to: when the high and low bits of the first message count correspond to the first When the upper and lower bits of the three message counts are equal, it is determined that the second transmission state of the port queue is the abnormal transmission state. The monitoring system of the port queue jam according to claim 10, wherein the monitoring system of the port queue jam further comprises: The port queue closing module is configured to close the port queue when it is detected that the port queue is blocked. Cache message clearing module, configured to clear the buffered message in the port queue and retain relevant configuration parameters required for resetting the port queue; The port queue reset module is configured to reset the port queue according to the retained related configuration parameter to restore an initial state corresponding to when the port queue does not send a message; The port queue enable module is configured to enable the port queue to enable the port queue to send packets after the port queue is restored to the initial state. The monitoring system of the port queue jam according to claim 11, wherein the monitoring system of the port queue jam further comprises: The monitoring thread creation module is configured to create a monitoring thread of the port queue and perform polling monitoring of the port queue congestion on all ports on the router in the preset monitoring period. A non-transitory computer readable storage medium having stored therein instructions that, when executed by a processor of a carrier-grade router comprising a plurality of ports, cause the router to implement a method of monitoring port clogging, The method includes the following steps: When the current monitoring period arrives, it is determined whether the buffer status of the currently polled port queue satisfies a preset abnormal cache state condition; Determining whether the first sending state of the port queue meets a preset abnormal sending state condition when the buffering state of the port queue meets the abnormal cache state condition; And when the first sending state of the port queue meets the abnormal sending state, the current monitoring period is extended by a preset delay duration to determine whether the second sending state of the port queue meets the abnormal sending state. condition; When the second transmission state of the port queue satisfies the abnormal transmission state condition, it is determined that the port queue is blocked.

Images