CN1878131A - Method and system for carrying out flow-control - Google Patents

Abstract

The present invention relates to a method and system for flow control, the core of which is: first obtain the comprehensive parameters of the length of the internal queue and the average residence time of message packets in the internal queue, or the detection result of the CPU occupancy rate, and when the obtained comprehensive parameters are confirmed Or when the CPU occupancy detection result data exceeds a set threshold, flow control is performed by setting global shared parameters. Through the dual detection mechanism, the present invention can grasp the overload situation of the system more accurately, ensuring the effective use of the system processing capacity; moreover, by setting global shared parameters for flow control, it ensures that important services are processed preferentially when the system is overloaded . In addition, by setting global shared parameters to control the messages that pre-enter the input buffer and/or the messages in the internal queues with different priorities, it can ensure that important services are prioritized when the system is overloaded, and the effective processing rate of the system is improved. and the processing rate of valid messages.

Description

Method and system for flow control

technical field

The invention relates to the fields of electronics and communication, in particular to a flow control method and system.

Background technique

In the field of electronics and communications, system resources, such as CPU processing power, timers, and memory buffers, are very limited. With the increase of system communication traffic, various communication processes often have connection requests terminated abnormally due to lack of system resources. Connection requests are under heavy load. When the connection request load is too heavy, it may cause the CPU usage to be too high, resulting in a sharp decline in system performance, which in turn leads to a decrease in the density of successful calls.

A lower density of successful calls means lower connection rates, lower throughput, and higher communication delays. The greater the proportion of rejected calls to the total number of calls, the faster the rate of throughput decline will eventually lead to the loss of traffic and the decline of user satisfaction. Therefore, deciding to reject or accept the connection request as early as possible, that is, performing flow control on the connection request message, can effectively utilize limited system resources.

In the prior art related to the present invention, the technical solution for flow control is described in conjunction with FIG. 1:

First put the pending message from the previous link into the input buffer of the system;

At the same time, the system detects the CPU usage in real time. If the average CPU usage of the system does not exceed the set threshold within a period of time, the admission module normally reads the messages in the input buffer and processes them, and then sends the processed messages to Send to the output buffer; if the average CPU usage of the system exceeds the set threshold for a period of time, the admission module will be notified to start flow control, and the admission module will discard the excessive message packets entering the system for a period of time until the CPU Occupancy drops back to another set threshold and the system resumes normal operation.

As can be seen from the technical solutions of the above-mentioned prior art, it has the following deficiencies:

1. The prior art only uses the CPU occupancy rate as the basis for flow control, resulting in that the system processing capacity cannot be effectively utilized.

Because the essential factor affecting the system processing capability is that the waiting time of message packets in the input buffer is too long, and the increase of CPU usage rate is only an appearance. If only the CPU usage rate is used as the basis for flow control, it may lead to excessive If the system overload state is ended early, the CPU cannot work at full capacity at this time, so the system processing capacity cannot be effectively utilized.

2. The prior art directly discards the messages without distinction, resulting in a decrease in the effective processing rate of the messages by the system.

For a process, because the priority of the message in the middle of the process is higher than the initial message of the process, if the message is discarded without distinction, the effective processing rate of the message will be reduced by the system.

3. Some messages have been waiting in the input buffer for too long, and there is no need to process them. If the message packets that are too long and invalid are not removed, some resources of the system will be spent on invalid processing, and the processing rate of valid messages will decrease.

Contents of the invention

The purpose of the present invention is to provide a method and system for communication flow control. Through the present invention, the problem in the prior art that the system processing capacity cannot be effectively utilized due to the fact that only the CPU occupancy rate is used as the basis for flow control is solved. . It solves the problem in the prior art that the effective processing rate of the message is decreased by the system because the message is directly discarded without distinction.

In addition, the purpose of the present invention is to provide another method and system for communication flow control, which solves the problem in the prior art that the processing rate of effective messages decreases due to the fact that the invalid message packets are not eliminated due to too long delay. question.

The purpose of the present invention is achieved through the following technical solutions:

The invention provides a method for flow control, including:

A. Obtain the comprehensive parameters of the internal queue length and the average residence time of message packets in the internal queue, or the CPU occupancy detection results;

B. When it is confirmed that the acquired comprehensive parameter or the CPU occupancy rate detection result data exceeds the set threshold, control the communication flow through the set global shared parameters.

Wherein, the step B specifically includes:

B1. Judging whether the obtained comprehensive parameter or the CPU occupancy rate detection result data exceeds a set threshold, if not, then execute step B2; if yes, execute step B3;

B2. Clear the flow control flag, enter the normal operation state, and execute step B1;

B3. Set the flow control flag, control the communication flow through the set global shared parameters, and then proceed to step B1.

Wherein, the global shared parameters include:

Message priority configuration parameters and/or scheduled scan weight parameters.

Wherein, the step B3 includes:

B31. Read the global shared parameters;

B32. Perform corresponding level flow control on the message flow in internal queues of different priority levels according to the global shared parameters, and then continue to execute step B1.

Wherein, the step B32 specifically includes:

B321. Determine whether the internal queue with the highest priority level is full according to the global shared parameters, if so, perform emergency flow control, and then continue to execute step B1; otherwise, execute step B322;

B322. According to the global shared parameters, it is judged whether internal queues other than the internal queue with the highest priority level are full, if yes, perform general flow control, and then continue to execute step B1; otherwise, execute step B1.

Wherein, the process of emergency flow control described in step B321 specifically includes:

B3211. When it is confirmed that the message in the internal queue is an initial message or a retry message, according to the set message priority configuration parameter, the priority level of the initial message or the retry message is reduced by one level;

B3212. Scheduling the initial message or retry message after the priority level has dropped according to the order of the priority level after the drop from high to low;

or,

B3213. When it is confirmed that the message in the internal queue is an initial message or a retry message, increase the scheduling scanning weight of the internal queue with the highest priority level by one level according to the set scheduling scanning weight parameter, and increase the scheduling scanning weight of the internal queue with the lowest priority level. The scheduling scan weight of the queue is reduced by one level;

B3214. Schedule the initial message or the retry message according to the changed scheduling scan weight parameter and the order of priority levels from high to low;

or,

B3215. When confirming that the message in the internal queue is an initial message or a retry message, according to the set message priority configuration parameter, the priority level of the initial message is reduced by one level; and, according to the set scheduling scan weight parameter, the The scheduled scan weight of the internal queue with the highest priority level is increased by one level, and the scheduled scan weight of the internal queue with the lowest priority level is decreased by one level;

B3216. Scheduling the initial message or retry message after the priority level has dropped according to the changed scheduling scan weight parameter, and according to the descending priority level from high to low;

or,

B3217. When it is confirmed that the message in the internal queue is the initial message, according to the set message priority configuration parameters, the priority level of the initial message is lowered by one level.

B3218. Judging whether the message in the internal queue is a retry initial message, if so, then according to the set message priority configuration parameters, the priority level of the initial message is lowered by one level, and then step B3219 is performed; otherwise, directly Execute step B3220;

B3219. Schedule the retry initial message after the priority level drops according to the descending priority level from high to low;

B3220. Scheduling the initial messages with reduced priority levels in descending order of the reduced priority levels.

or,

B3221. When it is confirmed that the message in the internal queue is the initial message, then according to the set message priority configuration parameter, the priority level of the initial message is lowered by one level;

B3222. Determine whether the message in the internal queue is a retry initial message, and if so, lower the priority level of the initial message by one level according to the set message priority configuration parameter, and scan according to the set schedule The weight parameter increases the scheduling scanning weight of the internal queue with the highest priority level by one level, and decreases the scheduling scanning weight of the internal queue with the lowest priority level by one level, and then executes step B3223; otherwise, directly executes step B3224;

B3223. According to the changed scheduling scanning weight parameter, and according to the order of the descending priority rank from high to low, schedule the retry initial message after the descending priority rank;

B3224. According to the changed scheduling scan weight parameter, and according to the order of the decreased priority level from high to low, schedule the initial message with the decreased priority level.

Wherein, the process of performing general flow control described in step B322 specifically includes:

B3221. According to the set message priority configuration parameters, the priority level of the retry message in the internal queue is reduced by one level;

B3222. Scheduling the retry messages after the priority level has dropped according to the order of the priority level after the drop from high to low.

Wherein, the step B specifically includes:

B4. Judging whether the acquired comprehensive parameter or the CPU occupancy rate detection result data exceeds the set threshold, if not, clear the flow control flag and enter the normal operation state; if so, execute step B5;

B5. Set the flow control flag, and control the communication flow through the set global shared parameters.

B6. Control the communication flow according to the message volume in the input buffer and/or the message volume in the internal queue.

Wherein, the step B6 specifically includes:

B61. Determine whether there is a message in the input buffer, if so, perform admission control on the message that is about to enter the input buffer, and then continue to execute step B4; otherwise, clear the flow control flag and enter the normal operation state;

or,

B62, judging whether there is a message in the internal queue, if so, then carry out scheduling processing by the scheduling scan weight parameters set for different priority queues, and then continue to perform step B4; otherwise, clear the flow control flag and enter the normal operation state;

or,

B63. Determine whether there is a message in the input buffer, and if so, perform admission control on the message that is about to enter the input buffer, and then continue to execute step B4; otherwise, execute step B64;

B64. Determine whether there is a message in the internal queue, if so, control the communication flow through the scheduling scanning weight parameters set for different priority queues, and then continue to perform step B4; otherwise, clear the flow control flag and enter the normal operation state.

Wherein, said step B6 also includes:

B65. Determine whether the number of scheduled scans has been reached, and if so, end the process; otherwise, go to step B6.

Wherein, the process of performing admission control on the messages ready to enter the input buffer described in step B6 specifically includes:

C1. Read a message from the input buffer, and set the message lifetime of the message;

C2. Judging whether the message is an initial message or a retry message according to the user identity, if so, lowering the priority of the initial message or the retry message by one level, and then performing step C3; otherwise, directly performing step C3;

C3. According to the priority level of the message, judge whether the corresponding internal queue is full, if so, discard the message; otherwise, put the message carrying the message lifetime into the internal queue, and then end the process;

or,

C4. Read a message from the input buffer, and set the message lifetime of the message;

C5. Judging whether the message is an initial message according to the user ID, if so, lowering the priority of the initial message by one level, and then performing step C6; otherwise, directly performing step C6;

C6. Judging whether the message is a retry message according to the user identity, if so, then lowering the priority of the retry message by one level, and then performing step C7; otherwise, directly performing step C7;

C7. According to the priority level of the message, judge whether the corresponding internal queue is full, and if so, discard the message; otherwise, put the message into the internal queue, and then end this process.

Wherein, the process of scheduling the scheduling process through the scheduling scanning weight parameters set for different priority queues described in step B6 specifically includes:

D1. According to the weighted scheduling algorithm, read a message from the internal queue in order of priority from high to low;

D2. Determine whether the messages in the internal queue of this priority have been scheduled, and if so, schedule the messages in the internal queue of the next priority level; otherwise, go to step D1;

D3. Determine whether all priority internal queues have been scheduled and scanned, and if so, end the process; otherwise, go to step D1.

Wherein, the weighted scheduling algorithm described in step D1 specifically includes:

scan all internal queues each round; and,

According to the set scheduling scan weight parameter, the number of messages scheduled to be read from internal queues with high priority levels in each round is greater than the number of messages scheduled to be read from internal queues with low priority levels; and,

The sum of the number of messages scheduled to be read from all priority internal queues per round is balanced against the sum of the number of messages scheduled to be read from the input buffer per round; and,

If you increase the scheduled scan weight parameter of the high-priority internal queue, decrease the scheduled scan weight parameter of the low-priority internal queue.

Wherein, before said step D1 also includes:

Judging whether the message in the internal queue has reached the message lifetime set for it, if so, discarding the message from the internal queue directly; otherwise, performing step D1.

A system for flow control provided by the present invention includes:

Main control module, flow control module, input buffer and internal queue;

The main control module obtains the comprehensive parameters of the internal queue length and the average residence time of the message packet in the internal queue, or the CPU occupancy detection result; and when it is confirmed that the obtained comprehensive parameter or the CPU occupancy detection result data exceeds the set When the threshold is set, the flow control module performs flow control on the messages to be entered into the input buffer and the messages in the internal queue according to the set global shared parameters.

Wherein, the system also includes:

The admission module is configured to accept the control of the main control module or the control of the main control module and the flow control module, and control the flow of messages ready to enter the input buffer.

Wherein, the system also includes:

The scheduling module is used to accept the control of the main control module or the control of the main control module and the flow control module, and schedule and read the message flow in the internal queue according to the set global shared parameters.

As can be seen from the technical solution provided by the present invention above, the present invention first obtains the comprehensive parameters of the internal queue length and the average residence time of the message packet in the internal queue, or the CPU occupancy rate detection result, when it is confirmed that the obtained comprehensive parameters or the When the CPU occupancy rate detection result data exceeds the set threshold, the communication flow control is performed through the set global shared parameters. Through the dual detection mechanism of the CPU occupancy rate and the buffer delay/length adopted by the present invention, the overload situation of the system can be grasped more accurately, thereby ensuring that the system processing capacity can be effectively utilized; moreover, the present invention can pass the set Globally shared parameters are used to control the communication flow, thereby ensuring that important services are prioritized when the system is overloaded, and solving the problem in the prior art that the effective processing rate of messages by the system decreases due to indiscriminate direct discarding of messages.

In addition, the present invention also controls the message flow that is ready to enter the input buffer and/or the message flow in the internal queue with different priorities through the set global shared parameters, and through the message lifetime set for the message, it further guarantees Important businesses are prioritized when the system is overloaded, which improves the system's effective message processing rate and effective message processing rate.

Description of drawings

Fig. 1 is the schematic diagram of prior art;

Fig. 2 is the schematic diagram of the first embodiment of the present invention;

Fig. 3 is the schematic diagram of the second embodiment of the present invention;

Fig. 4 is the schematic diagram of the third embodiment of the present invention;

Fig. 5 is the schematic diagram of the fourth embodiment of the present invention;

Fig. 6 is the flowchart of the fifth embodiment of the present invention;

FIG. 7 is a flow chart of controlling communication traffic through set global shared parameters described in the fifth embodiment of the present invention;

Fig. 8 is a flow chart of the sixth embodiment of the present invention;

FIG. 9 is a flow chart 1 of calling the admission module in the sixth embodiment of the present invention to perform admission control on messages that are about to enter the input buffer;

Fig. 10 is a flow chart 2 of calling the admission module in the sixth embodiment of the present invention to perform admission control on messages that are about to enter the input buffer;

Fig. 11 is a flowchart of the seventy-ninth embodiment of the present invention;

FIG. 12 is a flow chart of calling the scheduling module described in the seventh embodiment of the present invention to perform scheduling processing by scheduling and scanning weight parameters set for different priority queues;

Fig. 13 is a flow chart of the eighth embodiment of the present invention.

Detailed ways

The present invention provides a method and system for communication flow control, the core of which is: the main control module obtains the comprehensive parameters of the length of the internal queue and the average residence time of message packets in the internal queue, or the detection result of the CPU occupancy; When the acquired comprehensive parameter or the CPU occupancy detection result data exceeds the set threshold, the communication flow is controlled through the set global shared parameters.

The schematic diagram of the first embodiment of a communication flow control system provided by the present invention is shown in Figure 2, including: input buffer 00, output buffer 01, admission module 02, scheduling module 03, internal queue 04, main Control module 05, flow control module 06 and configurable global shared parameter database 07.

Input buffer 00, which is used to cache the message packets from the previous link;

Output buffer 01, which is used to cache processed messages and wait for the next link to process;

The admission module 02 is used to accept the control of the main control module or the control of the main control module and the flow control module, and control the flow of messages to be entered into the input buffer: to determine whether the message to be input into the input buffer 00 has access Conditions, if the access conditions are met, set the message priority according to the global flow control parameters, start the timer, and set the message lifetime (Time to Live, TTL), that is, add a TTL field in front of each message and fill in the initial value, such as 5s, and then put the message into the corresponding buffer according to the priority level of the message. Messages that do not meet the entry conditions are discarded directly.

The scheduling module 03 is used to accept the control of the main control module or the control of the main control module and the flow control module, and schedule the message flow in the internal queue according to the set global shared parameters. Read: schedule the scan message from the internal queue 04 and process it. Scheduling scanning is performed according to the scheduling scanning weight parameter set in the global shared parameters of the flow control module 06, that is, scanning all queues in each round, and the number of message packets processed by the internal queue scanning is from high to low according to the scheduling scanning weight of the internal queue The order decreases successively. The number of message packets processed each time is an algorithm parameter. Algorithm parameters can be dynamically adjusted by the flow control module through global shared parameters. In the process of processing, the timer is read every time a message packet is processed to obtain the waiting time of the message packet in the queue, and then the TTL of all message packets in the queue is subtracted from the waiting time. If the TTL is found to be 0, the message will be discarded; if the internal queue is found to be full, the message pre-entered in the internal queue will be discarded.

Internal queue 04, the present invention sets three priority internal queues, the internal queue with a priority level of 1 has the highest priority, and the internal queue with a priority level of 3 has the lowest priority. Of course, other numbers of internal queues can be set according to system needs .

The main control module 05 is used to complete functions such as flow detection, starting the flow control operation mode and restoring the normal operation mode. The main control module obtains the comprehensive parameters of the internal queue length and the average residence time of the message packet in the internal queue, or the CPU occupancy detection result; and when it is confirmed that the obtained comprehensive parameter or the CPU occupancy detection result data exceeds the set When the threshold is set, the flow control operation mode is started through the flow control module according to the set global shared parameters to control the flow of messages to be entered into the input buffer and messages in the internal queue.

The flow control module 06 receives the command of the main control module 05, controls the message flow in the internal queue by setting the global shared parameters, and affects the behavior of the admission module and the scheduling module through the set global shared parameters; The set global shared parameters are placed in the configurable global parameter database 07.

The configurable global shared parameter database 07 is writable for the flow control module and the main control module, and read-only for the admission module and the scheduling module. Global shared parameters include message priority configuration parameters and scheduling scan weight parameters, etc.

During implementation, the present invention is embedded in a host computer that will perform flow control, and the present invention starts and operates together with the host computer communication program.

When normal, the internal queue 04 reports the length of the internal queue and the average residence time of the message packet in the internal queue to the main control module 05;

The main control module 05 monitors the communication flow, that is, the input buffer length and the delay comprehensive parameter, and the additional CPU occupancy rate detection result. Only when the communication host is congested, the main control module 05 of the present invention detects that the internal queue or the CPU occupancy rate are abnormal , the flow control module 06 is notified to enable the entire system to enter the flow control mode by setting global shared parameters, that is, to control the communication flow of the system.

The process of controlling the communication flow of the system may include the following three processes:

1. The main control module 05 controls the message flow according to the set global shared parameters through the flow control module 06 .

2. The main control module 05 schedules or schedules the admission module 02 through the flow control module 06, and controls the message flow to be entered into the input buffer according to the set global shared parameters.

3. The main control module 05 schedules or schedules the schedule module 03 through the flow control module 06, and controls the message flow in the internal queue according to the set global shared parameters.

If a single-task operating system is adopted, the above three processes (main control, admission and scheduling) are processed serially. If a multitasking operating system is used, the above three processes are processed in parallel.

The second embodiment provided for the system of the present invention, as shown in FIG. 3 , differs from the first embodiment in that it does not include the scheduling module 03 and the admission module 02 . In the process of realizing the flow control, it only includes the process that the main control module 05 controls the message flow according to the set global shared parameters through the flow control module 06 .

The third embodiment provided for the system of the present invention, as shown in FIG. 4 , differs from the first embodiment in that it does not include the scheduling module 03 . In the process of realizing flow control, it includes the process that the main control module 05 controls the message flow according to the set global shared parameters through the flow control module 06, and the main control module 05 schedules or schedules admission through the flow control module 06 Module 02, the process of controlling the flow of messages ready to enter the input buffer according to the set global shared parameters.

The fourth embodiment provided for the system of the present invention, as shown in FIG. 5 , differs from the first embodiment in that it does not include an admission module 02 . In the process of realizing flow control, it includes the process that the main control module 05 controls the message flow according to the set global shared parameters through the flow control module 06, and the main control module 05 schedules or schedules the scheduling module through the flow control module 06 03. The process of controlling the message flow in the internal queue according to the set global shared parameters.

A fifth embodiment of a method for controlling communication flow provided by the present invention, as shown in FIG. 6 , includes:

Step S101 , acquiring comprehensive parameters of the length of the internal queue and the average stay time of message packets in the internal queue, or the detection result of the CPU occupancy rate.

Step S102, judging whether the acquired comprehensive parameter or the CPU occupancy rate detection result data exceeds a set threshold, if not, execute step S103; if yes, execute step S104.

Step S103, clear the flow control flag, and enter the normal operation state.

Step S104, set the flow control flag, and control the communication flow through the set global shared parameters, and then continue to execute step S102.

In step S104, the global shared parameters include: message priority configuration parameters and/or scheduled scanning weight parameters. In this step, the global shared parameter is first read; and then the flow of messages in the internal queues of different priority levels is controlled at a corresponding level according to the global shared parameter. The specific implementation process, as shown in Figure 7, includes the following steps:

Step S51, reading the global shared parameters;

Step S52, judge whether the internal queue with the highest priority level is full according to the global shared parameters, if so, execute step S53, that is, perform emergency flow control, and then execute step S56; otherwise, execute step S54;

Step S54, judge whether the internal queues other than the internal queue with the highest priority level are full according to the global shared parameters, if so, execute step S55, that is, perform general flow control; otherwise, execute step S56, that is, return to step S103, that is Clear the flow control flag and enter the normal operation state.

The process of performing emergency flow control in step S53 specifically includes the following situations.

In the first case, when it is confirmed that the message in the internal queue is an initial message or a retry message, the priority level of the internal queue is only adjusted according to the set message priority configuration parameters, and according to the adjusted priority The class controls the flow of messages on this internal queue. The specific implementation process is as follows:

Step 531, when it is confirmed that the message in the internal queue is the initial message or the retry message, according to the set message priority configuration parameters, the priority level of the initial message or the retry message is lowered by one level.

Step 532, schedule the initial message or the retry message whose priority level has decreased according to the order of the decreased priority level from high to low.

In the second case, when it is confirmed that the message in the internal queue is an initial message or a retry message, adjust the scheduling scan weight parameter of the priority queue according to the priority of the internal queue, and then adjust the internal queue according to the adjusted scheduling scan weight parameter. The flow of messages in the queue is controlled. The specific implementation process is as follows:

Step 533, when it is confirmed that the message in the internal queue is an initial message or a retry message, increase the scheduled scan weight of the internal queue with the highest priority level by one level according to the set scheduled scan weight parameter, and increase the scheduled scan weight of the internal queue with the lowest priority level The scheduled scan weight of the internal queue is lowered by one level.

Step 534, schedule the initial message or the retry message according to the changed scheduling scan weight parameter and the order of priority levels from high to low.

In the third case, when it is confirmed that the message in the internal queue is an initial message or a retry message, not only the priority of the internal queue is adjusted, but also the scheduling scan weight parameter of the internal queue is adjusted according to the priority of the current internal queue , and control the message flow in the internal queue according to the adjusted priority and the scheduling scan weight parameters. The specific implementation process is as follows:

Step 535, when confirming that the message in the internal queue is an initial message or a retry message, according to the set message priority configuration parameter, lower the priority level of the initial message by one level; and, according to the set scheduling scan weight parameter Increase the scheduled scan weight of the internal queue with the highest priority level by one level, and decrease the scheduled scan weight of the internal queue with the lowest priority level by one level;

Step 536, according to the changed scheduling scanning weight parameter, and according to the order of the descending priority level from high to low, schedule the initial message or the retry message after the priority level is decreased;

In the fourth case, when confirming that the message in the internal queue is the initial message, not only adjust the priority of the internal queue, but also further judge whether the message is a retry initial message, and further adjust the internal queue according to the judgment result priority, and then control the flow of messages in the internal queue according to the adjusted priority. The specific implementation process includes:

Step 537: When it is confirmed that the message in the internal queue is the initial message, the priority level of the initial message is lowered by one level according to the set message priority configuration parameters.

Step 538, judging whether the message in the internal queue is a retry initial message, if so, then according to the set message priority configuration parameters, the priority level of the initial message is lowered by one level, and then step 539 is executed; otherwise, Execute step 540 directly.

Step 539, schedule the retried initial message with the reduced priority level in descending order of the reduced priority level.

Step 540, schedule the initial messages with reduced priority levels in descending order of the reduced priority levels.

In the fifth case, the message flow in the internal queue is controlled by adjusting the scheduling scanning weight parameter of the internal queue on the basis of the fourth case. The specific implementation process includes:

Step 541, when it is confirmed that the message in the internal queue is the initial message, then according to the set message priority configuration parameters, the priority level of the initial message is lowered by one level.

Step 542, judging whether the message in the internal queue is a retry initial message, if so, according to the set message priority configuration parameter, the priority level of the initial message is lowered by one level, and, according to the set scheduling The scan weight parameter increases the scheduled scan weight of the internal queue with the highest priority level by one level, and decreases the scheduled scan weight of the internal queue with the lowest priority level by one level, and then executes step 543; otherwise, directly executes step 544.

Step 543 , according to the changed scheduling scan weight parameter, and according to the descending priority order of the reduced priority level, schedule the retry initial message with the reduced priority level.

Step 544: Schedule the initial messages with lowered priority levels according to the changed scheduling scan weight parameters and in descending order of the lowered priority levels.

In the above step S55, the process of performing general flow control specifically includes:

First, according to the set message priority configuration parameters, the priority level of the retry message in the internal queue is lowered by one level; then, the retry after the priority level is lowered is retried according to the descending priority level from high to low. Messages are scheduled.

The sixth embodiment provided for the method of the present invention, as shown in FIG. 8 , includes:

Step S201 , acquiring comprehensive parameters of the length of the internal queue and the average stay time of message packets in the internal queue, or the detection result of the CPU occupancy.

Step S202, judging whether the acquired comprehensive parameter or the CPU occupancy detection result data exceeds a set threshold, if not, execute step S203; if yes, execute step S204.

Step S203, clear the flow control flag, and enter the normal operation state.

Step S204, setting a flow control flag, and controlling the communication flow through the set global shared parameters. This step is the same as the description in the fifth embodiment above, and will not be described in detail.

Step S205, determine whether there is a message in the input buffer, if so, execute step S206; otherwise, execute step S203, clear the flow control flag, and enter the normal operation state;

Step S206, calling the admission module to perform admission control on the messages to be entered into the input buffer, and then continue to execute step S202.

The specific implementation process of step S206 includes two cases, the first case is shown in Figure 9, including the following steps:

21. Read a message from the input buffer, and set the message lifetime of the message;

22. Judging whether the message is an initial message or a retry message according to the user ID, if so, perform step 23, lower the priority of the initial message or retry message by one level, and then perform step 24; otherwise, directly Execute step 24;

24. According to the priority level of the message, judge whether the corresponding internal queue is full, if so, execute step 25, and discard the message; otherwise, execute step 26, and put the message carrying the message lifetime into the internal queue , and then perform step 27;

27. Determine whether the number of scheduled scans has been reached, if yes, execute step 28, end this process, and return to step S202; otherwise, go to step 21.

The second case is shown in Figure 10 and includes the following steps:

31. Read a message from the input buffer, and set the message lifetime of the message;

32. Determine whether the message is an initial message according to the user ID, and if so, perform step 33, that is, lower the priority of the initial message by one level, and then perform step 34; otherwise, directly perform step 34;

34. Determine whether the message is an initial retry message according to the user ID, if so, perform step 35, that is, lower the priority of the retry initial message by one level, and then perform step 36; otherwise, directly perform step 36;

36. According to the priority level of the message, judge whether the corresponding internal queue is full, if so, execute step 37, that is, discard the message, and then execute step 39; otherwise, execute step 38, that is, put the message inside queue, and then go to step 39.

39. Determine whether the number of scheduled scans has been reached, if yes, execute step 40, end this process, and return to step S202; otherwise, go to step 31.

The seventh embodiment provided for the method of the present invention, as shown in FIG. 11 , includes:

Step S301 , acquiring comprehensive parameters of the length of the internal queue and the average stay time of message packets in the internal queue, or the detection result of the CPU occupancy rate.

Step S302, judging whether the acquired comprehensive parameter or the CPU occupancy detection result data exceeds a set threshold, if not, execute step S303; if yes, execute step S304.

Step S303, clear the flow control flag, and enter the normal operation state.

Step S304, setting a flow control flag, and controlling the communication flow through the set global shared parameters. This step is the same as the description in the fifth embodiment above, and will not be described in detail.

Step S305, determine whether there is a message in the internal queue, if so, execute step S306; otherwise execute step S307, clear the flow control flag, and enter the normal operation state;

In step S306, the scheduling module is invoked, and the scheduling processing is performed through the scheduling scan weight parameters set for different priority queues, and then step S302 is continued.

In step S306, the dispatching module is called, and the dispatching process is performed through the dispatching and scanning weight parameters set for different priority queues, as shown in FIG. 12 , specifically including:

71. Determine whether the message in the internal queue has reached the set message lifetime, if so, perform step 72, and directly discard the message from the internal queue; otherwise, perform step 73.

73. According to the weighted scheduling algorithm, read a message from the internal queue in order of priority from high to low.

The weighted scheduling algorithm described in step 73 specifically includes:

scan all internal queues each round; and,

According to the set scheduling scan weight parameter, the number of messages scheduled to be read from internal queues with high priority levels in each round is greater than the number of messages scheduled to be read from internal queues with low priority levels; and,

The sum of the number of messages scheduled to be read from all priority internal queues per round is balanced against the sum of the number of messages scheduled to be read from the input buffer per round; and,

If you increase the scheduled scan weight parameter of the high-priority internal queue, decrease the scheduled scan weight parameter of the low-priority internal queue.

74. Determine whether the scheduling of the messages in the internal queue of this priority is completed, and if so, execute step 75 to schedule the messages in the internal queue of the next priority level, and then execute step 76; otherwise, go to step 71 ;

76. Determine whether all priority internal queues have been scheduled and scanned, if yes, execute step 77, end this process, and return to step S302; otherwise, go to step 71.

The eighth embodiment provided for the method of the present invention, as shown in FIG. 13 , includes:

Step S401, obtaining the comprehensive parameters of the length of the internal queue and the average stay time of message packets in the internal queue, or the detection result of CPU occupancy.

Step S402, judging whether the acquired comprehensive parameter or the CPU occupancy detection result data exceeds a set threshold, if not, execute step S403; if yes, execute step S404.

Step S403, clear the flow control flag, and enter the normal operation state.

Step S404, setting a flow control flag, and controlling the communication flow through the set global shared parameters. This step is the same as the description in the fifth embodiment above, and will not be described in detail.

Step S405, judging whether there is a message in the input buffer, if so, execute step S406; otherwise, execute step S407;

Step S406, that is, calling the admission module to perform admission control on the messages to be entered into the input buffer. This step is the same as the description in the sixth embodiment above, and will not be described in detail.

Step S407, determine whether there is a message in the internal queue, if so, execute step S408; otherwise, execute step S409, clear the flow control flag, and enter the normal operation state.

In step S408, call the scheduling module to control the communication flow through the scheduling scan weight parameters set for different priority queues, and then return to step S402. This step is the same as the description in the seventh embodiment above, and will not be described in detail.

As can be seen from the technical solution provided by the present invention above, the present invention first obtains the comprehensive parameters of the internal queue length and the average residence time of the message packet in the internal queue, or the CPU occupancy rate detection result, when it is confirmed that the obtained comprehensive parameters or the When the CPU occupancy rate detection result data exceeds the set threshold, the communication flow control is performed through the set global shared parameters. Through the dual detection mechanism of the CPU occupancy rate and the buffer delay/length adopted by the present invention, the overload situation of the system can be grasped more accurately, thereby ensuring that the system processing capacity can be effectively utilized; moreover, the present invention can pass the set Globally shared parameters are used to control communication flow, thereby ensuring that important services are prioritized when the system is overloaded, and solving the problem in the prior art that the effective processing rate of messages by the system decreases due to indiscriminate direct discarding of messages.

In addition, the present invention also controls the message flow that is ready to enter the input buffer and/or the message flow in the internal queue with different priorities through the set global shared parameters, and through the message lifetime set for the message, it further guarantees Important businesses are prioritized when the system is overloaded, which improves the system's effective message processing rate and effective message processing rate.

The above is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed in the present invention can easily think of changes or Replacement should be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention should be determined by the protection scope of the claims.

Claims (17)

1. A method for flow control, comprising: A. Obtain the comprehensive parameters of the internal queue length and the average residence time of message packets in the internal queue, or the CPU occupancy detection results; B. When it is confirmed that the acquired comprehensive parameter or the CPU occupancy rate detection result data exceeds the set threshold, control the communication flow through the set global shared parameters. 2. The method according to claim 1, characterized in that the step B specifically comprises: B1. Judging whether the obtained comprehensive parameter or the CPU occupancy rate detection result data exceeds a set threshold, if not, then execute step B2; if yes, execute step B3; B2. Clear the flow control flag, enter the normal operation state, and execute step B1; B3. Set the flow control flag, control the communication flow through the set global shared parameters, and then proceed to step B1. 3. The method according to claim 2, wherein the global shared parameters include: Message priority configuration parameters and/or scheduled scan weight parameters. 4. The method according to claim 3, wherein said step B3 comprises: B31. Read the global shared parameters; B32. Perform corresponding level flow control on the message flow in internal queues of different priority levels according to the global shared parameters, and then continue to execute step B1. 5. The method according to claim 4, characterized in that the step B32 specifically comprises: B321. Determine whether the internal queue with the highest priority level is full according to the global shared parameters, if so, perform emergency flow control, and then continue to execute step B1; otherwise, execute step B322; B322. According to the global shared parameters, it is judged whether internal queues other than the internal queue with the highest priority level are full, if yes, perform general flow control, and then continue to execute step B1; otherwise, execute step B1. 6. The method according to claim 5, wherein the process of emergency flow control in step B321 specifically includes: B3211. When confirming that the message in the internal queue is an initial message or a retry message, according to the set message priority configuration parameter, the priority level of the initial message or the retry message is lowered by one level; B3212. Scheduling the initial message or retry message after the priority level has dropped according to the order of the priority level after the drop from high to low; or, B3213. When it is confirmed that the message in the internal queue is an initial message or a retry message, increase the scheduling scanning weight of the internal queue with the highest priority level by one level according to the set scheduling scanning weight parameter, and increase the scheduling scanning weight of the internal queue with the lowest priority level. The scheduling scan weight of the queue is reduced by one level; B3214. Schedule the initial message or the retry message according to the changed scheduling scan weight parameter and the order of priority levels from high to low; or, B3215. When confirming that the message in the internal queue is an initial message or a retry message, according to the set message priority configuration parameter, the priority level of the initial message is reduced by one level; and, according to the set scheduling scan weight parameter, the The scheduled scan weight of the internal queue with the highest priority level is increased by one level, and the scheduled scan weight of the internal queue with the lowest priority level is decreased by one level; B3216. Scheduling the initial message or retry message after the priority level has dropped according to the changed scheduling scan weight parameter, and according to the descending priority level from high to low; or, B3217. When it is confirmed that the message in the internal queue is the initial message, according to the set message priority configuration parameters, the priority level of the initial message is lowered by one level. B3218. Judging whether the message in the internal queue is a retry initial message, if so, then according to the set message priority configuration parameters, the priority level of the initial message is lowered by one level, and then step B3219 is performed; otherwise, directly Execute step B3220; B3219. Schedule the retry initial message after the priority level drops according to the descending priority level from high to low; B3220. Scheduling the initial messages with reduced priority levels in descending order of the reduced priority levels. or, B3221. When it is confirmed that the message in the internal queue is the initial message, then according to the set message priority configuration parameter, the priority level of the initial message is lowered by one level; B3222. Determine whether the message in the internal queue is a retry initial message, and if so, lower the priority level of the initial message by one level according to the set message priority configuration parameter, and scan according to the set schedule The weight parameter increases the scheduling scanning weight of the internal queue with the highest priority level by one level, and decreases the scheduling scanning weight of the internal queue with the lowest priority level by one level, and then executes step B3223; otherwise, directly executes step B3224; B3223. According to the changed scheduling scanning weight parameter, and according to the order of the descending priority rank from high to low, schedule the retry initial message after the descending priority rank; B3224. According to the changed scheduling scan weight parameter, and according to the order of the decreased priority level from high to low, schedule the initial message with the decreased priority level. 7. The method according to claim 5, characterized in that the process of performing general flow control in step B322 specifically includes: B3221. According to the set message priority configuration parameters, the priority level of the retry message in the internal queue is reduced by one level; B3222. Scheduling the retry messages after the priority level has dropped according to the order of the priority level after the drop from high to low. 8. The method according to claim 1, wherein the step B specifically comprises: B4. Judging whether the obtained comprehensive parameter or the CPU occupancy rate detection result data exceeds the set threshold, if not, then clear the flow control flag and enter the normal operation state; if so, then perform step B5; B5. Set the flow control flag, and control the communication flow through the set global shared parameters. B6. Control the communication flow according to the message volume in the input buffer and/or the message volume in the internal queue. 9. The method according to claim 8, characterized in that the step B6 specifically comprises: B61. Determine whether there is a message in the input buffer, if so, perform admission control on the message to be entered into the input buffer, and then continue to execute step B4; otherwise, clear the flow control flag and enter the normal operation state; or, B62, judging whether there is a message in the internal queue, if so, then carry out scheduling processing by the scheduling scan weight parameters set for different priority queues, and then continue to perform step B4; otherwise, clear the flow control flag and enter the normal operation state; or, B63. Determine whether there is a message in the input buffer, and if so, perform admission control on the message that is about to enter the input buffer, and then continue to execute step B4; otherwise, execute step B64; B64. Determine whether there is a message in the internal queue, if so, control the communication flow through the scheduling scanning weight parameters set for different priority queues, and then continue to perform step B4; otherwise, clear the flow control flag and enter the normal operation state. 10. The method according to claim 9, characterized in that the step B6 further comprises: B65. Determine whether the number of scheduled scans has been reached, and if so, end the process; otherwise, go to step B6. 11. The method according to claim 9, characterized in that the process of performing admission control on the messages ready to enter the input buffer in step B6 specifically includes: C1. Read a message from the input buffer, and set the message lifetime of the message; C2. Judging whether the message is an initial message or a retry message according to the user identity, if so, lowering the priority of the initial message or the retry message by one level, and then performing step C3; otherwise, directly performing step C3; C3. According to the priority level of the message, judge whether the corresponding internal queue is full, if so, discard the message; otherwise, put the message carrying the message lifetime into the internal queue, and then end the process; or, C4. Read a message from the input buffer, and set the message lifetime of the message; C5. Judging whether the message is an initial message according to the user ID, if so, lowering the priority of the initial message by one level, and then performing step C6; otherwise, directly performing step C6; C6. Judging whether the message is a retry message according to the user identity, if so, then lowering the priority of the retry message by one level, and then performing step C7; otherwise, directly performing step C7; C7. According to the priority level of the message, judge whether the corresponding internal queue is full, and if so, discard the message; otherwise, put the message into the internal queue, and then end this process. 12. The method according to claim 9, wherein the process of scheduling in step B6 through scheduling scan weight parameters set for different priority queues specifically includes: D1. According to the weighted scheduling algorithm, read a message from the internal queue in order of priority from high to low; D2. Determine whether the messages in the internal queue of this priority have been scheduled, and if so, schedule the messages in the internal queue of the next priority level; otherwise, go to step D1; D3. Determine whether all priority internal queues have been scheduled and scanned, and if so, end the process; otherwise, go to step D1. 13. The method according to claim 12, wherein the weighted scheduling algorithm in step D1 specifically includes: scan all internal queues each round; and, According to the set scheduling scan weight parameter, the number of messages scheduled to be read from internal queues with high priority levels in each round is greater than the number of messages scheduled to be read from internal queues with low priority levels; and, The sum of the number of messages scheduled to be read from all priority internal queues per round is balanced against the sum of the number of messages scheduled to be read from the input buffer per round; and, If you increase the scheduled scan weight parameter of the high-priority internal queue, decrease the scheduled scan weight parameter of the low-priority internal queue. 14. The method according to claim 10, further comprising: before the step D1: Judging whether the message in the internal queue has reached the message lifetime set for it, if so, discarding the message from the internal queue directly; otherwise, performing step D1. 15. A system for flow control, comprising: Main control module, flow control module, input buffer and internal queue; The main control module obtains the comprehensive parameters of the internal queue length and the average residence time of the message packet in the internal queue, or the CPU occupancy detection result; and when it is confirmed that the obtained comprehensive parameter or the CPU occupancy detection result data exceeds the set When the threshold is set, the flow control module performs flow control on the messages to be entered into the input buffer and the messages in the internal queue according to the set global shared parameters. 16. The system of claim 15, further comprising: The admission module is configured to accept the control of the main control module or the control of the main control module and the flow control module, and control the flow of messages ready to enter the input buffer. 17. The system according to claim 15 or 16, further comprising: The scheduling module is used to accept the control of the main control module or the control of the main control module and the flow control module, and schedule and read the message flow in the internal queue according to the set global shared parameters.

Images