Service scheduling method for limited edge computing system

Abstract

The invention discloses a service scheduling method for a limited edge computing system, and relates to the technical field of shipborne and airborne edge computing systems. The method comprises the steps of S1, measurement of communication frequency, S2, measurement of communication duration, S3, measurement of communication quality and S4, optimal scheduling algorithm. According to the method, the communication frequency, the communication duration and the communication quality are measured through the communication pedometer and the scheduling method of optimal instruction execution is calculated through the multi-dimensional weighted average algorithm, so that uniform scheduling of service consistency under different network conditions is realized, the scheduling service is stable and reliable, the scheduling service cost is low, and the scheduling efficiency is high. And according to the business growth demand, the edge service is flexibly expanded.

Description

Technical Field

[0001] The present invention relates to the technical field of shipborne and airborne edge computing systems, and specifically to a service scheduling method for a restricted edge computing system. Background Art

[0002] In shipborne and airborne systems, network connectivity and server hardware performance are greatly limited. In particular, network connectivity is limited. Due to the particularity of the environment, such systems cannot maintain continuous network access. They usually need to use satellite networks, 4G\5G networks, site WIFI networks and other means to complement the network. This has a great impact on the operation of modern application system services, mainly in terms of operation and maintenance.

[0003] Take a cruise ship as an example. The service provider hopes to provide nearly a thousand tourists on board with a rich itinerary service during the cruise, including entertainment content, business content, online itinerary services, etc. However, in the ocean far away from the shore, communication can only be carried out through satellite. Satellite communication not only has low bandwidth and high latency, but also very expensive traffic fees. For operators, it is impossible to operate and maintain shipboard services in the same way as in a conventional Internet environment. Similar problems are more prominent in airborne systems.

[0004] At present, the solutions for this scenario on the market are not perfect. They usually only solve a specific business point, but do not fully consider the commonality of such edge systems as a whole and provide systematic solutions. For example, the existing solutions usually transmit data back at a fixed period and report the edge system status regularly. However, if you want to temporarily maintain and upgrade the edge service, it is impossible to do so.

[0005] However, existing technologies still have some defects:

[0006] 1. Existing methods only provide customized solutions to specific problems in specific business scenarios, and fail to achieve the ability to remotely dispatch and manage services in such restricted environments in a universal manner. As a result, they are unable to take into account the characteristics of multi-network communication of ship-borne and airborne servers and achieve scheduling of consistent services under different network conditions.

[0007] 2. It is impossible to quickly troubleshoot and recover edge service failures under different network conditions;

[0008] 3. It is impossible to achieve reliable service scheduling in an environment with extremely high airborne safety requirements;

[0009] 4. Unable to flexibly expand the scheduling method of edge services according to the needs of business growth;

[0010] In response to the above problems, the inventors propose a service scheduling method for a restricted edge computing system to solve the above problems. Summary of the invention

[0011] In order to solve the above problems, the present invention aims to provide a service scheduling method for a restricted edge computing system.

[0012] To solve the above technical problems, the present invention adopts the following technical solution: a service scheduling method for a restricted edge computing system, comprising the following steps:

[0013] S1. Measurement of communication frequency

[0014] Through the communication pedometer, the communication between the edge system and the scheduling system is maintained by continuously sending packets at evenly spaced intervals;

[0015] S2. Measurement of communication duration

[0016] Through the communication pedometer, the continuous step counting data is superimposed to measure the communication time length under each network condition;

[0017] S3. Measurement of communication quality

[0018] By recording each communication result through the communication pedometer, the real-time quality of communication under each network condition is identified to complete the measurement of communication quality;

[0019] S4. Optimal Scheduling Algorithm

[0020] The instructions to be executed are scored by a multi-dimensional weighted average algorithm, and the specific steps of the multi-dimensional weighted average algorithm are as follows:

[0021] A: Dimension division, set corresponding weight values ​​for each dimension to calculate weighted scores. There can be many dimensions, and more can be added as needed;

[0022] B: Weighted scoring. The instruction score is calculated based on the estimated time required for the instruction to be executed. The current network real-time bandwidth must be obtained through the communication pedometer, and the time evaluation is calculated based on the estimated data volume of the instruction to be executed.

[0023] C: Finally, the weighted average scores of each instruction are sorted, and the ones with higher scores are executed first. This is the optimal instruction execution scheduling method that takes into account the influencing factors of various dimensions.

[0024] Preferably, in S1, the higher the communication frequency between the edge system and the scheduling system, the higher the ability to measure network connectivity.

[0025] Preferably, in S3, the communication quality measurement includes two dimensions:

[0026] Dimension 1: evaluates the implementation availability of the network through the loss of step counting;

[0027] Dimension 2: Evaluate the effectiveness of the network through the delay of step counting.

[0028] Preferably, in S1-S3, the specific method of the communication pedometer is as follows:

[0029] The communication pedometer is implemented by a custom private protocol. Each time the edge system is started, the communication pedometer begins. The pedometer uses 2 bytes, i.e., 16 bits of binary, for counting.

[0030] Preferably, in S4, the service scheduling instructions are centrally managed, the center side is responsible for issuing instructions uniformly, and the edge side is provided with an instruction proxy module for arranging and executing instructions.

[0031] Compared with the prior art, the present invention has the following beneficial effects:

[0032] In the present invention, through the methods S1-S4, the communication frequency, duration and quality are measured by a communication pedometer, and the scheduling method for the optimal instruction execution is calculated by a multi-dimensional weighted average algorithm, thereby realizing unified scheduling of service consistency under different network conditions, stable and reliable scheduling services, low scheduling service costs, and flexible expansion of edge services according to the needs of business growth. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, 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 invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0034] Figure 1 The present invention is a flow chart of a service scheduling method for a restricted edge computing system.

[0035] Figure 2 This is a data diagram of the step counting method of the communication pedometer of the present invention.

[0036] Figure 3 Schematic diagram of the weight of dimension one of an embodiment of the present invention.

[0037] Figure 4 Schematic diagram of the weight of dimension 2 of an embodiment of the present invention.

[0038] Figure 5 Schematic diagram of the weight of dimension three of an embodiment of the present invention. DETAILED DESCRIPTION

[0039] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0040] Example: Figure 1-5 As shown, the present invention provides a service scheduling method for a restricted edge computing system, comprising the following steps:

[0041] S1. Measurement of communication frequency

[0042] The communication between the edge system and the scheduling system is maintained by continuously sending packets at even intervals through the communication pedometer. The communication pedometer is an algorithm tool specially designed for the restricted edge environment to measure the network communication situation. The higher the communication frequency between the edge system and the scheduling system, the higher the ability to measure network connectivity.

[0043] S2. Measurement of communication duration

[0044] Through the communication pedometer, the continuous step counting data is superimposed to measure the communication time length under each network condition;

[0045] S3. Measurement of communication quality

[0046] By recording each communication result through the communication pedometer, the real-time quality of communication under each network condition is identified to complete the measurement of communication quality;

[0047] The measurement of communication quality includes two dimensions:

[0048] Dimension 1: evaluates the implementation availability of the network through the loss of step counting;

[0049] Dimension 2: evaluate the effectiveness of the network through the delay of step counting;

[0050] The specific method of communicating with the pedometer is as follows:

[0051] The communication pedometer is implemented by a custom private protocol. Each time the edge system is started, the communication pedometer starts. The pedometer uses 2 bytes, i.e. 16 bits of binary, to count.

[0052] The first two bits of the 16-bit binary code identify the current step counting mode, that is, the type of network channel. A total of four modes can be identified.

[0053] The last 14 bits of the 16-bit binary are used for counting. Each time a step is counted, the binary number is increased by 1. When the number is full of 14 bits, it is a step counting cycle. The step count is reset to zero and accumulated again.

[0054] The benefits of the communication pedometer are as follows:

[0055] 1. Minimum cost to achieve network measurement. For example, if the step frequency is once per second, each step count consumes 2 Bytes of traffic, and a total of 168K traffic is consumed in 24 hours a day. This traffic consumption is almost negligible.

[0056] 2. The measurement of pedometer network quality can be used to guide the arrangement of service scheduling;

[0057] S4. Optimal Scheduling Algorithm

[0058] The instructions to be executed are scored through a multi-dimensional weighted average algorithm. The higher the score, the higher the execution order. The service scheduling instructions are centrally managed. The center is responsible for issuing instructions uniformly, and the edge side sets an instruction proxy module for arranging and executing instructions. The specific steps of the multi-dimensional weighted average algorithm are as follows:

[0059] A: Dimension division, set corresponding weight values ​​for each dimension to calculate weighted scores. There can be many dimensions, and more can be added as needed;

[0060] B: Weighted scoring. The instruction score is calculated based on the estimated time required for the instruction to be executed. The current network real-time bandwidth must be obtained through the communication pedometer, and the time evaluation is calculated based on the estimated data volume of the instruction to be executed.

[0061] C: Finally, the weighted average scores of each instruction are sorted, and the ones with higher scores are executed first, that is, the optimal instruction execution scheduling method after taking into account the influencing factors of various dimensions;

[0062] Take 30 seconds as a segment and score it based on the time consumed. The score ranges from 10 to 1, with a minimum of 1. After obtaining the score of each instruction to be executed, a weighted score is performed based on the current dimension information.

[0063] The specific scoring method is shown in Table 1:

[0064]

[0065] Table 1 Rating data table

[0066] The scoring formula is as follows:

[0067] T = x ÷ (y ÷ 8);

[0068] M = a*b;

[0069] N = (m1 score + m2 score + mn score) ÷ n;

[0070] Among them, T represents time consumption, x represents the amount of instruction data, and y represents the network bandwidth;

[0071] M represents the dimension score, a represents the score, and b represents the dimension weight;

[0072] N represents the weighted average score, and m is the dimension.

[0073] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A service scheduling method for a constrained edge computing system, characterized in that: The following steps are involved: S1. Measurement of communication frequency Through the communication pedometer, the communication between the edge system and the scheduling system is maintained by continuously sending packets at evenly spaced intervals; S2. Measurement of communication duration Through the communication pedometer, the continuous step counting data is superimposed to measure the communication time length under each network condition; S3. Measurement of communication quality By recording each communication result through the communication pedometer, the real-time quality of communication under each network condition is identified to complete the measurement of communication quality; S4. Optimal Scheduling Algorithm The instructions to be executed are scored by a multi-dimensional weighted average algorithm, and the specific steps of the multi-dimensional weighted average algorithm are as follows: A: Dimension division, set corresponding weight values ​​for each dimension to calculate weighted scores. There can be many dimensions, and more can be added as needed; B: Weighted scoring. The instruction score is calculated based on the estimated time required for the instruction to be executed. The current network real-time bandwidth must be obtained through the communication pedometer, and the time evaluation is calculated based on the estimated data volume of the instruction to be executed. C: Finally, the weighted average scores of each instruction are sorted, and the ones with higher scores are executed first. This is the optimal instruction execution scheduling method that takes into account the influencing factors of various dimensions.

2. A service scheduling method for a restricted edge computing system according to claim 1, characterized in that: In S1, the higher the communication frequency between the edge system and the scheduling system, the higher the ability to measure network connectivity.

3. A service scheduling method for a restricted edge computing system according to claim 1, characterized in that: In S3, the communication quality measurement includes two dimensions: Dimension 1: evaluates the implementation availability of the network through the loss of step counting; Dimension 2: Evaluate the effectiveness of the network through the delay of step counting.

4. A service scheduling method for a restricted edge computing system according to claim 1, characterized in that: In S1-S3, the specific method of the communication pedometer is as follows: The communication pedometer is implemented by a custom private protocol. Each time the edge system is started, the communication pedometer begins. The pedometer uses 2 bytes, that is, 16 bits of binary, for counting.

5. A service scheduling method for a restricted edge computing system according to claim 1, characterized in that: In S4, the service scheduling instructions are centrally managed, the center side is responsible for issuing instructions uniformly, and the edge side is equipped with an instruction proxy module for arranging and executing instructions.

Images