Self-adaptive scheduling optimization method suitable for intensive high-concurrency computing power shortage

Abstract

The invention relates to the technical field of 5G communication, and provides a self-adaptive scheduling optimization method suitable for intensive high-concurrency computing power shortage, which comprises the following five steps from two aspects of computing power vacation and service adjustment: acquiring current time at the start and end of protocol stack operation, the time precision being at least us level, and the time precision being at least us level; the time difference is the actual total consumed time of protocol stack processing. The time consumption proportion is calculated by dividing the actual total time consumption by the designed time consumption and is divided into three grades of less than 50%, less than 100% and timeout. Taking the design time consumption of one slot, namely 500us, as an example, the first gear is processed after 250us, the second gear is processed within 500us, and the third gear is processed after more than 500us. And the scheduling number overhead of a single slot is reduced, and the overall throughput rate is improved. According to the method, under the condition that the computing power resources are limited, the relation between other advantages and performance of the sorting algorithm, the matching relation between the uplink and the downlink and the relation between the computing power resources and the air interface utilization rate can be rebalanced, and timely and effective services are provided for users.

Description

Technical Field

[0001] The present application belongs to the field of 5G communication transmission technology, and in particular, relates to an adaptive scheduling optimization method suitable for intensive high-concurrency computing power shortage. Background Art

[0002] In large-scale public events in high-density user environments, especially in TDD systems, the base station's uplink computing resources are under great pressure due to the uneven data transmission requirements of user equipment (UE). In this scenario, traditional fixed parameters and scheduling strategies cannot meet the dynamically changing resource requirements.

[0003] Dynamic allocation of computing resources: When the number of users and business demands fluctuate, how to dynamically adjust computing resources to meet the needs of different users and businesses is a major challenge facing current technology; in addition, the problem of high concurrent user access: the surge in the number of users at large events has led to a shortage of computing resources at the base station MAC layer, especially in the TDD system, where uplink computing resources are particularly scarce. For large events, the network's emergency support capabilities are crucial. There needs to be a plan to deal with equipment failures, high traffic, high interference, and other problems, and to be able to quickly adjust network parameters to deal with emergencies. Summary of the invention

[0004] The embodiment of the present application provides an adaptive scheduling optimization method suitable for intensive high-concurrency computing power shortage, aiming to solve the technical problems that it is difficult to conduct network environment measurement and analysis manually, the labor cost is high and the accuracy of environmental analysis is low, and the business effect cannot be guaranteed efficiently.

[0005] The technical solution adopted by the present invention is: 1. An adaptive scheduling optimization method suitable for intensive high-concurrency computing power shortage, characterized in that the current time is obtained at the start and end of the protocol stack operation, the time accuracy is at least micron level, the time difference is the actual total time consumed by the protocol stack processing, and the time consumption ratio is calculated by dividing the actual total time consumption by the design time consumption, which is divided into three levels: less than 50%, less than 100%, and overtime; taking the time consumption of 1 slot, i.e. 500us, as an example, the processing is completed in 250us for the first level, the processing is completed within 500us for the second level, and the processing exceeds 500us for the third level; the specific method is as follows: under the monitoring of the 100ms computing power state, S1. When the probability of the time consumption ratio being in the second gear exceeds 80%, the first gear computing power optimization is started: Switch the high-complexity sorting algorithm to a high-performance sorting algorithm; S2: When the probability of the time consumption ratio being in the third gear exceeds 10%, the second gear computing power optimization is started: Dynamically adjust the maximum number of users scheduled for uplink and downlink. In the case of more downlink than uplink in TDD, the original 1:1 maximum number of users scheduled for uplink and downlink is changed to 2:1, which shortens the downlink processing time and leaves time for uplink to respond to more user uplink requests. S3: When the probability of the time consumption ratio being in the third gear exceeds 30%, the third gear computing power optimization is started: Dynamically adjust the timing structure so that the slot that is performing uplink scheduling will no longer perform downlink scheduling. If the deployment mode of sharing the same core for uplink and downlink is adopted, the time freed up by downlink is used for uplink scheduling. If the uplink and downlink cores are deployed separately, the downlink core resources are allocated to the uplink. S4. Continue to monitor the 100ms computing power status. When the probability that the time consumption ratio is in the third gear exceeds 60%, start the fourth gear computing power optimization: calculate the current average PRB utilization; If the average PRB utilization is less than 60%, the number of scheduled low-priority SRs is reduced. Scheduling SRs occupies less air interface RB resources, but consumes computing resources that are basically equivalent to BSR data services. Scheduling a large number of SRs will lead to a waste of air interface resources. Reducing the response of low-priority SRs can allocate the corresponding computing resources to data services and improve the utilization of air interface resources. 5. Continue to monitor the 100ms computing power status. When the probability that the time consumption ratio is in the third gear exceeds 80%, start the fifth gear computing power optimization: calculate the current average PRB utilization. If the average PRB utilization is less than 80%, increase the scheduling priority of the large BSR.

[0006] Preferably, in step S4, low-priority SR requests may also cause the terminal to accumulate more data requests due to delayed responses.

[0007] Preferably, in step 5, UEs with a value greater than or less than a threshold are divided into two queues according to the original scheduling priority; during scheduling, UEs in the large BSR queue are allocated 80% of the RB resources, and after 80% of the RB resources are allocated, they are allocated to UEs in the small BSR queue.

[0008] Compared with the prior art, the embodiments of the present invention have the following beneficial effects: 1) Through this invention, when computing resources are limited, the relationship between other advantages and performance of the sorting algorithm, the ratio between upstream and downstream, and the relationship between computing resources and air interface utilization can be rebalanced to provide users with timely and effective services.

[0009] 2) After the delayed response to the terminal data request with a small amount of data, the terminal will accumulate more data requests and send them at one time, reducing the scheduling overhead of a single slot and improving the overall throughput. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0011] Figure 1 It is a schematic diagram of the interaction between MAC and physical layer provided in an embodiment of the present application; Figure 2 It is a probability ratio diagram of different gears provided in an embodiment of the present application; Figure 3 is a timing structure diagram of a TDD 5ms single cycle provided in an embodiment of the present application, Figure 4 It is a schematic diagram of the principle provided by the embodiment of the present application; Figure 5 It is a diagram of algorithm steps provided in the embodiment of the present application; Figure 6 It is a schematic diagram of the overall solution provided by the embodiment of the present application. DETAILED DESCRIPTION

[0012] In the following description, specific details such as specific system structures, technologies, etc. are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present application. However, it should be clear to those skilled in the art that the present application may also be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present application.

[0013] The present invention provides an adaptive computing power optimization method in five steps from two aspects: computing power shifting and business adjustment.

[0014] like Figure 1 and 5 As shown in the figure, it is a schematic diagram of the interaction between MAC and the physical layer. Assuming that the designed running time of MAC is 500us, the time consumption ratio and gear position are calculated according to the actual end time.

[0015] like Figure 2 As shown in the figure, the computing power status detection method is: obtain the current time at the beginning and end of the protocol stack operation, with a time accuracy of at least us level, and the time difference is the actual total time consumed by the protocol stack processing. The time consumption ratio is calculated by dividing the actual total time consumption by the designed time consumption, which is divided into three levels: less than 50%, less than 100%, and timeout.

[0016] Taking the design time of one slot, i.e. 500us, as an example, the processing time completed in 250us is the first level, the processing time completed within 500us is the second level, and the processing time exceeding 500us is the third level.

[0017] Continue to monitor the 100ms computing power status; When the probability that the time consumption ratio is in the third level exceeds 10%, the second level computing power optimization is started.

[0018] The time consumption of each slot is determined by the data request of the terminal. The time consumption ratio of different slots varies to a certain extent. It is necessary to count the proportion of different gears within 100ms to obtain the exact computing power status; Specifically, Figure 6 As shown, S1. When the probability that the time consumption ratio is in the second level exceeds 80%, start the first level computing power optimization: switch the high-complexity sorting algorithm to a high-performance sorting algorithm.

[0019] For example, the commonly used proportional fairness algorithm is switched to the round-robin algorithm.

[0020] S2: When the probability that the time consumption ratio is in the third level exceeds 10%, the second level computing power optimization is started: the maximum number of upstream and downstream scheduled users is dynamically adjusted.

[0021] For the situation where TDD has more downlink than uplink, the original 1:1 maximum number of scheduled uplink and downlink users can be changed to 2:1, which can shorten the downlink processing time and leave time for the uplink to respond to more user uplink requests.

[0022] S3: When the probability that the time consumption ratio is in the third gear exceeds 30%, the third gear computing power optimization is started: the timing structure is dynamically adjusted, and the slots that are performing uplink scheduling are no longer scheduled downlink.

[0023] If the deployment method of sharing the same core for uplink and downlink is adopted, the time freed up by the downlink can be directly used for uplink scheduling; if the deployment method of separate cores for uplink and downlink is adopted, the downlink core resources need to be allocated to the uplink.

[0024] S4: When the probability of the time consumption ratio being in the third gear exceeds 60%, the fourth gear computing power optimization is started: Calculate the current average PRB utilization. If the average PRB utilization is less than 60%, reduce the number of scheduled low-priority SRs.

[0025] Scheduling SR occupies less air interface RB resources, but consumes computing resources that are basically equivalent to BSR data services. Scheduling a large number of SRs will lead to a waste of air interface resources.

[0026] Reducing the response to low-priority SRs can allocate the corresponding computing resources to data services and improve the utilization of air interface resources. At the same time, low-priority SR requests will also cause the terminal to accumulate more data requests due to delayed responses.

[0027] Reduce the scheduling overhead of a single slot and improve the overall throughput.

[0028] S5. When the probability of the time consumption ratio being in the third gear exceeds 80%, Start the fifth level of computing power optimization: calculate the current average PRB utilization. If the average PRB utilization is less than 80%, increase the scheduling priority of large BSRs.

[0029] The principle of the fifth gear here is similar to that of the fourth gear. After the delayed response to the terminal data request with a small amount of data, the terminal will accumulate more data requests and send them up at one time, reducing the scheduling overhead of a single slot and improving the overall throughput. It is worth mentioning here that the method to improve the large BSR is to set the large BSR threshold to 1000 bytes, and divide the UEs larger and smaller than the threshold into two queues according to the original scheduling priority. During scheduling, the UEs in the large BSR queue are allocated 80% of the RB resources. After 80% of the RB resources are allocated, they are allocated to the UEs in the small BSR queue.

[0030] like Figure 3 As shown, there are 8 downlinks and 2 uplinks. K1 refers to the time from when the terminal receives the downlink data to when it performs HARQ feedback. K2 refers to the time from when the terminal receives the uplink scheduling indication to when it sends PUSCH. As shown in the figure, the uplink data of slot 8 is scheduled in slot 3. Slot 3 does not perform downlink scheduling to reserve more time for uplink operations.

[0031] like Figure 4 As shown in the figure, the SR scheduling authorization is smaller, and the BSR scheduling authorization is larger. If the SR authorization is 4RB, and the BSR user requests 40RB, assuming that the maximum scheduling capacity of the base station is 6 UEs, there are 4 SR schedulings before optimization, which is reduced to 2 after optimization. After optimization, there are fewer idle resources, which can meet more data services of users. The fragmented SR requests are converted into a whole BSR request, which can reduce the uplink requests of the terminal as a whole. An embodiment of the present application provides a computer program product. When the computer program product runs on a mobile terminal, the mobile terminal can implement the steps in the above-mentioned method embodiments when executing the computer program product.

[0032] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0033] The embodiments described above are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. An adaptive scheduling optimization method suitable for intensive high-concurrency computing power shortage, characterized in that: Get the current time at the beginning and end of the protocol stack operation. The time accuracy is at least micron level. The time difference is the actual total time consumed by the protocol stack. The time consumption ratio is calculated by dividing the actual total time consumption by the designed time consumption. It is divided into three levels: less than 50%, less than 100%, and timeout. Taking the time consumption of 1 slot, i.e. 500us, as an example, the processing time of 250us is the first level, the processing time of 500us is the second level, and the processing time of more than 500us is the third level. The specific method is as follows: monitor the computing power state of 100ms, S1. When the probability of the time consumption ratio being in the second gear exceeds 80%, the first gear computing power optimization is started: Switch the high-complexity sorting algorithm to a high-performance sorting algorithm; S2: When the probability of the time consumption ratio being in the third gear exceeds 10%, the second gear computing power optimization is started: Dynamically adjust the maximum number of users scheduled for uplink and downlink. In the case of more downlink than uplink in TDD, the original 1:1 maximum number of users scheduled for uplink and downlink is changed to 2:1, which shortens the downlink processing time and leaves time for uplink to respond to more user uplink requests. S3: When the probability of the time consumption ratio being in the third gear exceeds 30%, the third gear computing power optimization is started: Dynamically adjust the timing structure so that the slot that is performing uplink scheduling will no longer perform downlink scheduling. If the deployment mode of sharing the same core for uplink and downlink is adopted, the time freed up by downlink is used for uplink scheduling. If the uplink and downlink cores are deployed separately, the downlink core resources are allocated to the uplink. S4. Continue to monitor the 100ms computing power status. When the probability that the time consumption ratio is in the third gear exceeds 60%, start the fourth gear computing power optimization: calculate the current average PRB utilization; If the average PRB utilization is less than 60%, the number of scheduled low-priority SRs is reduced. Scheduling SRs occupies less air interface RB resources, but consumes computing resources that are basically equivalent to BSR data services. Scheduling a large number of SRs will lead to a waste of air interface resources. Reducing the response of low-priority SRs can allocate the corresponding computing resources to data services and improve the utilization of air interface resources. S5. Continue to monitor the 100ms computing power status. When the probability that the time consumption ratio is in the third gear exceeds 80%, start the fifth gear computing power optimization: calculate the current average PRB utilization. If the average PRB utilization is less than 80%, increase the scheduling priority of the large BSR.

2. The adaptive scheduling optimization method for intensive high-concurrency computing power shortage as described in claim 1, characterized in that: In step S4, low-priority SR requests may also cause the terminal to accumulate more data requests due to delayed responses.

3. The adaptive scheduling optimization method for intensive high-concurrency computing power shortage as described in claim 1 is characterized in that: In step 5, UEs with a BSR greater than or less than the threshold are divided into two queues according to the original scheduling priority. During scheduling, UEs in the large BSR queue are allocated 80% of the RB resources, and after 80% of the RB resources are allocated, they are allocated to UEs in the small BSR queue.

Images