A computational method for unloading blocked tasks based on user retransmission mechanism in cloud-edge fusion

Abstract

The present invention belongs to the field of electronic information, and discloses a calculation method for unloading blocked tasks based on a user retransmission mechanism under cloud-edge fusion. First, the initial parameters related to transmission, processing, and task arrival time in the cloud-edge fusion system are counted, and the parameters related to task processing and transmission are calculated. The maximum initial task blocking probability in the edge computing system is calculated, and the maximum initial task arrival rate is further calculated; then, compared with the initial task arrival rate, the maximum retransmission time cost of the system under the corresponding situation is calculated; finally, compared with the actual retransmission time cost, the optimal value of the retransmission parameter under the corresponding situation is calculated, so as to unload the tasks blocked by the edge computing system between the cloud computing system and the user. The present invention uses the user retransmission mechanism to unload each blocked task between the edge and the cloud and the user through the method, thereby making full use of the low latency advantage of the edge computing system.

Description

Technical Field

[0001] The present invention belongs to the field of electronic information, and specifically relates to a calculation method for unloading blocked tasks based on a user retransmission mechanism under cloud-edge fusion. Background Art

[0002] Both edge computing and cloud computing systems can provide computing services to users of IoT devices. Edge computing systems are usually deployed in areas close to users and can provide users with low-latency services; however, their scale is relatively small, resulting in limited computing and storage resources. Cloud computing systems are usually deployed in areas far from users. Although the communication latency with users is high, their computing and storage resources are sufficient. As user computing needs increase, edge computing and cloud computing systems are integrated to provide computing services to more users.

[0003] The cloud-edge fusion computing system provides users with reliable computing services through sufficient computing resources. In order to further improve the service quality, the system needs a corresponding task offloading algorithm to optimize the resource allocation and task scheduling management between the edge and the cloud. The current task offloading method usually relies on the acquisition of information about cloud and edge real-time computing resources, and does not consider the randomness of task arrival and task processing time at the same time. Therefore, it is impossible to effectively and fully utilize the low latency advantage of edge computing.

[0004] The user retransmission mechanism can effectively utilize the low latency advantage of edge computing. The typical user retransmission mechanism is: when the task is not completed, the user resends the task at a fixed interval; or when the task is blocked by the edge computing system, the task is resent immediately. However, the former has a large interval, which reduces the low latency advantage of edge computing, and is not much different from sending tasks to the cloud computing system; the latter has a short interval, and when there are many tasks, a large number of retransmitted tasks will arrive at the edge computing system, exceeding the computing capacity of the system, thereby reducing the task computing efficiency of the edge computing system. Therefore, when to use the user retransmission mechanism, how to control the retransmission cost, and how to effectively adjust the proportion of task retransmission are one of the key technologies to improve the performance of cloud-edge fusion computing systems. Summary of the invention

[0005] To solve the above technical problems, the present invention provides a computing method for unloading blocked tasks based on a user retransmission mechanism under cloud-edge fusion, which is applied to a cloud-edge fusion computing system. By adjusting the task retransmission probability and utilizing the user retransmission mechanism, each blocked task is scheduled between the edge and the cloud, and between users, thereby making full use of the computing resources and low latency advantages of edge computing.

[0006] The present invention provides a method for calculating blocked task unloading based on a user retransmission mechanism under cloud-edge fusion, comprising the following steps:

[0007] Step 1: Based on the carrying range of the edge computing system's task processing capacity, calculate the upper limit of the initial task arrival rate in the cloud-edge fusion system;

[0008] Step 2: Compare the initial task arrival rate with the upper limit of the initial task arrival rate obtained in step 1. According to the comparison results, calculate the task blocking probability corresponding to different task retransmission parameter values ​​in the corresponding cases, so as to obtain the maximum retransmission time cost of the cloud-edge fusion system.

[0009] Step 3: compare the average retransmission time cost of the blocked task with the maximum retransmission time cost, and calculate the optimal value of the retransmission parameter under the corresponding situation according to the comparison result, so as to minimize the average task completion time;

[0010] Step 4: Use the optimal value of the retransmission parameter to offload the tasks blocked by the edge computing system between the cloud computing system and the user.

[0011] Furthermore, in step 1, the tasks in the cloud-edge fusion system include initial tasks, blocked tasks, and retransmitted tasks, wherein the initial task refers to the first time the task is sent by the user to the edge computing system, the blocked task refers to the task that is denied service by the edge computing, and the retransmitted task refers to the task that is sent by the user to the edge computing system again or multiple times; the retransmission time cost refers to the time from the task being blocked by the edge computing system to the time it is sent to the edge computing system again;

[0012] The task processing capacity of the edge computing system refers to the total task arrival rate. Less than or equal to the total computing speed of all computing nodes in the edge computing system ;The total task includes the initial task and the retransmission task;

[0013] Upper limit of initial task arrival rate The calculation formula is:

[0014] ,

[0015] in, is the maximum task blocking probability, defined as when the total task arrival rate When any task arrives at the edge computing system, there is a probability that the system cannot provide computing services; Processing task time for edge computing; is the number of computing nodes in the edge computing system.

[0016] Furthermore, step 2 is specifically as follows:

[0017] Statistics fixed time The initial number of tasks arriving at the edge computing system , then the initial task arrival rate ;

[0018] if , then calculate the task retransmission parameters respectively and When the task blocking probability of the edge computing system is and :

[0019] ,

[0020] ,

[0021] Among them, Σ is the summation symbol, which calculates the variables in the following formula Take 0 to The sum of all integer values ​​between; starting from 0 and gradually increasing the average retransmission time cost of the blocked task The value of and calculate the average task processing time under the two retransmission parameter conditions. When the average task processing time under the two retransmission parameter conditions is equal, The value is the maximum retransmission time cost of the system ;

[0022] if , then calculate the task retransmission parameters respectively and When , the corresponding task blocking probability of the edge computing system is and , The calculation formula is:

[0023] ,

[0024] in, is satisfied Under the condition of The maximum value; starting from 0 and gradually increasing the average retransmission time cost of the blocked task The value of and calculate the average task processing time under the two retransmission parameter conditions. When the average task processing time under the two retransmission parameter conditions is equal, The value is the maximum retransmission time cost of the system .

[0025] Furthermore, step 3 is specifically as follows:

[0026] if and , then the task retransmits the parameters The optimal value is ;

[0027] if and , then the task retransmits the parameters The optimal value is ;

[0028] if and , then the task retransmits the parameters The optimal value is ;

[0029] if and , then the task retransmits the parameters The optimal value is .

[0030] Furthermore, in step 4, when any task is blocked by the edge computing system, the system will The probability of notifying the user to resend the task to the edge computing system; otherwise, the edge computing system will The probability of offloading the task to the cloud computing system is , and the cloud computing system handles the task.

[0031] The beneficial effects described in the present invention are: the method described in the present invention utilizes the user retransmission mechanism to offload blocked tasks in the cloud-edge fusion computing system; determines the applicable scope of the user retransmission mechanism by calculating the upper limit of the initial task arrival rate, thereby avoiding negative impact on the computing stability of the edge computing system; effectively controls the usage cost of the user retransmission mechanism by calculating the maximum retransmission time cost of the cloud-edge fusion system; utilizes the retransmission parameters to adjust the retransmission ratio of blocked tasks, fully utilizing the low latency advantage of edge computing and improving the computing speed of tasks. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 It is a schematic diagram of the application of the method of the present invention;

[0033] Figure 2 is a flow chart of the method of the present invention;

[0034] Figure 3 It is a flowchart of the user retransmission mechanism;

[0035] Figure 4 is the simulation result of the task completion time changing with the initial task arrival rate;

[0036] Figure 5 It is the simulation result when the cloud-edge fusion system handles a large computing task;

[0037] Figure 6 It is the simulation result of the high retransmission time cost scenario. DETAILED DESCRIPTION

[0038] In order to make the contents of the present invention more clearly understood, the present invention is further described in detail below based on specific embodiments in conjunction with the accompanying drawings.

[0039] like Figure 1 As shown, the computing method for unloading blocked tasks based on user retransmission mechanism under cloud-edge fusion described in the present invention is applied to the edge computing system. When a task is blocked by the edge computing system, the blocked task is sent to the user and retransmitted to the edge computing system through the method of the present invention, or the task is offloaded to the cloud computing system for processing.

[0040] like Figure 2 and Figure 3 As shown, a calculation method for unloading blocked tasks based on a user retransmission mechanism under cloud-edge fusion according to the present invention includes the following steps:

[0041] Step 1: Initialize system parameters: Obtain relevant initial parameters of the cloud-edge fusion system. For example, the acquisition method uses statistical methods to count fixed time. The initial number of tasks arriving at the edge computing system , then the initial task arrival rate ; Similarly, statistics of fixed time The average time for a cloud computing system to process a task is the cloud computing task processing time. The average time it takes for a user to retransmit a blocked task is the average retransmission time cost of the blocked task. , the average time for the edge computing system to process tasks is the edge computing processing task time .

[0042] Calculate the upper limit of the initial task arrival rate: For example, in the following scenario, the number of computing nodes in the edge computing system is , and the system does not have enough storage space to queue the arriving tasks; at this time, when the total task arrival rate Equal to the total computing speed of all computing nodes in the edge computing system The probability that any task is blocked by the edge computing system , the formula is:

[0043] (1),

[0044] Σ is the summation symbol, which calculates the variables in the formula that follows it. Take 0 to The sum of all integer values ​​between ; therefore, the maximum initial task arrival rate of the edge computing system is , the formula is:

[0045] (2).

[0046] Step 2: Calculate the maximum retransmission time cost: If , task retransmission parameters The task blocking probability , the formula is:

[0047] (3);

[0048] Task retransmission parameters The task blocking probability , the formula is:

[0049] (4);

[0050] Using formula (4) The calculation of can be done using the following approximate algorithm: The value range of Inside, from Start by increasing the value by 0.01 each time. Substitute the right side of the formula for calculation, then calculate the difference between the left and right sides of the formula, and select the one corresponding to the smallest difference. The value of

[0051] The maximum retransmission time cost of the system , the formula is:

[0052] (5).

[0053] if , task retransmission parameters The task blocking probability The calculation of is the same as formula (3); the task retransmission parameter The task blocking probability , the formula is:

[0054] (6);

[0055] Using formula (6) The calculation of can be done using the same approximate algorithm as formula (4). Task retransmission parameter The maximum value of , the formula is:

[0056] (7);

[0057] At this time, the maximum retransmission time cost of the system , the formula is:

[0058] (8).

[0059] Step 3: Optimal value of task retransmission parameter: If and , then the task retransmits the parameters The optimal value is ;

[0060] if and , then the task retransmits the parameters The optimal value is ;

[0061] if and , then the task retransmits the parameters The optimal value is ;

[0062] if and , then the task retransmits the parameters The optimal value is .

[0063] Step 4: Unload blocked tasks: When any task is blocked by the edge computing system, the system will The probability of notifying the user to resend the task to the edge computing system; otherwise, the edge computing system will The probability of offloading the task to the cloud computing system is , and the cloud computing system handles the task.

[0064] The calculation method for unloading blocked tasks based on the user retransmission mechanism under cloud-edge fusion described in the present invention is applied to the following edge computing system. First, based on the simulation parameter settings shown in Table 1, the task completion time using the method of the present invention and not using the method of the present invention is compared; secondly, by changing the simulation parameters, the effectiveness of the method of the present invention in different environments is verified.

[0065] Table 1 Simulation example parameter settings

[0066]

[0067] like Figure 4 As shown in the figure, the task completion time increases with the increase of the initial task arrival rate. When the initial task arrival rate is low, only a small number of tasks are blocked by the system, so there is no obvious performance improvement using the method of the present invention; when the initial task arrival rate is high, more tasks in the edge computing system will be blocked. At this time, the use of the method of the present invention significantly reduces the task completion time.

[0068] (1) Changes in the time distribution of edge computing processing tasks:

[0069] When the edge computing system processes different types of tasks, the distribution function of the task processing time will change. For example, when the computing task is large, the time of the edge computing processing task in Table 1 is changed. The time of edge computing processing tasks is sampled using the displacement exponential distribution, thereby simulating the time variation of edge computing systems processing larger tasks. The scale parameter of the displacement exponential distribution is 1, the displacement parameter is 1, and the time mean is 2s.

[0070] like Figure 5 As shown, when the initial task arrival rate is high, the method of the present invention significantly reduces the task completion time.

[0071] (2) Changes in the average retransmission time cost of blocked tasks:

[0072] The average retransmission time cost of the blocked task is also one of the important factors affecting the performance of the method of the present invention. For example, increasing the average retransmission time cost of the blocked task in Table 1 , so that At this time, using the user retransmission mechanism to unload the blocked task has a large cost, which can verify the effectiveness of the method of the present invention in high-cost scenarios.

[0073] like Figure 6 As shown, the cost of using the method of the present invention in this environment is relatively high. When the initial task arrival rate is high, the method of the present invention is no longer effective; however, when the initial task arrival rate is low, the method of the present invention can still effectively reduce the task completion time.

[0074] The experimental results illustrate the feasibility and practicality of the present invention. Through simulation results, it is verified that when the method of the present invention is used under different environmental parameters, the task completion time is significantly shorter than when the method of the present invention is not used. Especially when the task arrival rate is large, the method of the present invention can further improve the performance of the cloud-edge fusion computing system.

[0075] The above description is only a preferred embodiment of the present invention and is not intended to be a further limitation of the present invention. All equivalent changes made using the contents of the present specification and drawings are within the protection scope of the present invention.

Claims

1. A computational method for unloading blocked tasks based on a user retransmission mechanism under cloud-edge fusion, characterized in that: The following steps are involved: Step 1: Based on the carrying range of the edge computing system's task processing capacity, calculate the upper limit of the initial task arrival rate in the cloud-edge fusion system; Among them, the tasks in the cloud-edge fusion system include initial tasks, blocked tasks and retransmission tasks. The initial task refers to the first time the task is sent by the user to the edge computing system, the blocked task refers to the task that is denied service by the edge computing, and the retransmission task refers to the task that is sent to the edge computing system by the user again or multiple times; the retransmission time cost refers to the time from the task being blocked by the edge computing system to the time it is sent to the edge computing system again; The task processing capacity of the edge computing system refers to the total task arrival rate. Less than or equal to the total computing speed of all computing nodes in the edge computing system ;The total task includes the initial task and the retransmission task; Upper limit of initial task arrival rate The calculation formula is: ， in, is the maximum task blocking probability, defined as when the total task arrival rate When any task arrives at the edge computing system, there is a probability that the system cannot provide computing services; Processing task time for edge computing; is the number of computing nodes in the edge computing system; Step 2: Compare the initial task arrival rate with the upper limit of the initial task arrival rate obtained in step 1. According to the comparison results, calculate the task blocking probability corresponding to different task retransmission parameter values ​​in the corresponding cases, so as to obtain the maximum retransmission time cost of the cloud-edge fusion system. Step 3: compare the average retransmission time cost of the blocked task with the maximum retransmission time cost, and calculate the optimal value of the retransmission parameter under the corresponding situation according to the comparison result, so as to minimize the average task completion time; Step 4: Use the optimal value of the retransmission parameter to offload the tasks blocked by the edge computing system between the cloud computing system and the user.

2. According to the calculation method of unloading blocked tasks based on user retransmission mechanism under cloud-edge fusion in claim 1, it is characterized in that: Step 2 is as follows: Statistics fixed time The initial number of tasks arriving at the edge computing system , then the initial task arrival rate ; if , then calculate the task retransmission parameters respectively and When the task blocking probability of the edge computing system and : ， ， Among them, Σ is the summation symbol, which calculates the variables in the following formula Take 0 to The sum of all integer values ​​between; starting from 0 and gradually increasing the average retransmission time cost of the blocked task The value of and calculate the average task processing time under the two retransmission parameter conditions. When the average task processing time under the two retransmission parameter conditions is equal, The value is the maximum retransmission time cost of the system ; if , then calculate the task retransmission parameters respectively and When , the corresponding task blocking probability of the edge computing system is and , The calculation formula is: ， in, is satisfied Under the condition of The maximum value; starting from 0 and gradually increasing the average retransmission time cost of the blocked task The value of and calculate the average task processing time under the two retransmission parameter conditions. When the average task processing time under the two retransmission parameter conditions is equal, The value is the maximum retransmission time cost of the system .

3. According to claim 2, a calculation method for unloading blocked tasks based on a user retransmission mechanism under cloud-edge fusion, characterized in that: Step 3 is as follows: if and , then the task retransmits the parameters The optimal value is ; if and , then the task retransmits the parameters The optimal value is ; if and , then the task retransmits the parameters The optimal value is ; if and , then the task retransmits the parameters The optimal value is .

4. According to claim 3, a calculation method for unloading blocked tasks based on a user retransmission mechanism under cloud-edge fusion, characterized in that: In step 4, when any task is blocked by the edge computing system, the system will The probability of notifying the user to resend the task to the edge computing system; otherwise, the edge computing system will The probability of offloading the task to the cloud computing system is , and the cloud computing system handles the task.

Images