Carbon emission evaluation method and device for edge-cloud vision collaborative architecture

Abstract

The present application provides a kind of carbon emission evaluation method and device for end edge cloud vision coordination architecture, comprising: according to the carbon emission parameter of each traditional vision terminal, intelligent vision terminal, edge server and cloud server, respectively construct end device carbon emission model, edge device carbon emission model, cloud device carbon emission model according to function, according to the carbon emission parameter that network transmission is generated due to network transmission when end edge cloud vision coordination Internet of Things runs, construct network transmission carbon emission model, adopt hungarian algorithm, solve when network transmission carbon emission model carbon emission minimum, and the computing resource that vision task occupies is less than or equal to the upper limit of each device computing resource unload selection strategy matrix, as the task deployment strategy of multiple vision tasks, decide each vision task executes in end side, edge side or cloud side, the carbon emission amount that is lower than that produced by directly executing in cloud, than directly in edge processing, due to the resource limit of edge, part of task cannot be executed task processing rate is high.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of carbon emission evaluation, and particularly relates to a carbon emission evaluation method of an end-edge cloud collaborative architecture, and gives a specific Hungarian algorithm to realize the reduction of carbon emission. BACKGROUND

[0002] In recent years, with the continuous development of deep learning and big data technologies in the visual field, the application fields such as intelligent transportation and intelligent security have higher visual perception, analysis, prediction and control capabilities. The operation efficiency and management level of the system under the construction of smart city are effectively improved. However, this change is at the cost of increasing computing and communication resources. The higher the complexity of the deep learning model, the greater the amount of calculation required, which will result in huge energy consumption and carbon emission. Edge computing, as a new computing paradigm, is used to replace cloud computing with high latency and high carbon emission, which is conducive to offloading computing tasks in real time to resource-rich infrastructure under low-carbon emission. In addition, edge devices do not need to consume a large number of cooling equipment, and usually choose GPU parallel to process resource-intensive edge workloads. Compared with CPU, it improves the running speed, saves energy consumption, and thus reduces the carbon footprint of the cloud service center. However, the current carbon emission evaluation of the end-edge cloud system mainly evaluates the carbon emission of the cloud data center, and the carbon emission of the end and edge sides has not given a specific evaluation method.

[0003] For the carbon emission evaluation of the end-edge cloud system, the existing method mainly designs the carbon use efficiency of the data center by the consumption of electricity; designs the carbon-free energy score measurement method for the carbon use efficiency of the power plant and the power purchase contract of the data operator; at the same time, considering the type of energy, renewable energy is given priority, and a renewable energy index is introduced to obtain the grid carbon intensity and emission source index. With the rise of edge computing, in order to meet the latency requirements of terminals, the large-scale computing demand of the end-edge cloud system sinks to edge devices, and the carbon footprint of the cloud server center also shifts to the end-edge device. The carbon emission of edge devices and terminal devices is different from that of cloud devices. Only using energy consumption to evaluate carbon emission and the carbon emission evaluation method of data centers cannot measure the carbon emission and the granularity division of carbon emission nodes. Specifically, the carbon emission of edge devices and terminal devices is different from that of cloud devices in that:

[0004] Cloud devices: carbon emission generated by IT computing devices, refrigeration systems and basic system device operation energy consumption, carbon emission caused by refrigerant leakage of refrigeration systems.

[0005] Edge devices: carbon emission generated by IT computing devices and basic system device operation energy consumption;

[0006] Traditional terminal devices: carbon emission generated by basic system device (video acquisition) operation energy consumption;

[0007] Smart visual terminal device: carbon emissions generated by IT computing device and basic operation device operation energy consumption. SUMMARY

[0008] In order to solve the deficiency of the prior art, more accurate measurement of the carbon emissions of the end-edge-cloud collaborative system is realized, the carbon emissions of the end-edge-cloud visual Internet of Things are unified under the smart city, and an end-edge-cloud system model is constructed according to the visual task offloading strategy. A multi-to-multi Hungarian matching strategy is used to obtain an offloading task decision scheme to determine the execution of each visual task at the end side, the edge side or the cloud side. The carbon emissions generated by direct execution on the cloud side are lower than those generated by direct execution on the edge side. Due to the resource limitation of the edge, the task processing rate of the part of the tasks that cannot be executed is high.

[0009] In view of the deficiency of the prior art, as shown in Figure 4 The present application proposes a carbon emission evaluation method for an end-edge-cloud visual collaborative architecture, which comprises:

[0010] In the initial step, an end-edge-cloud visual collaborative Internet of Things including a traditional visual terminal, a smart visual terminal, an edge server and a cloud server is obtained.

[0011] In the model construction step, according to the carbon emission parameters of the traditional visual terminal, the smart visual terminal, the edge server and the cloud server, an end device carbon emission model, an edge device carbon emission model and a cloud device carbon emission model are respectively constructed according to the functions. According to the carbon emission parameters generated by network transmission when the end-edge-cloud visual collaborative Internet of Things is running, a network transmission carbon emission model is constructed.

[0012] In the task preparation step, a plurality of visual tasks are obtained, and an offloading selection strategy matrix of the plurality of visual tasks is constructed according to the deployment strategy of each visual task. According to the network transmission carbon emission model, the end device carbon emission model, the edge device carbon emission model and the cloud device carbon emission model, a carbon emission matrix corresponding to the offloading selection strategy matrix is constructed. The carbon emission matrix represents the carbon emissions generated by each visual task in each server of the end-edge-cloud visual collaborative Internet of Things. According to the upper limit of the computing resources of each device in the end-edge-cloud visual collaborative Internet of Things, a computing resource load matrix is constructed.

[0013] In the task deployment step, the Hungarian algorithm is used to solve the offloading selection strategy matrix when the carbon emission of the network transmission carbon emission model is the smallest and the computing resource occupied by the visual task is less than or equal to the upper limit of the computing resources of each device. The task deployment strategy of the plurality of visual tasks is obtained, and the plurality of visual tasks are executed by using the task deployment.

[0014] The carbon emission evaluation method for the end-edge-cloud visual collaborative architecture, wherein the task deployment step comprises:

[0015] Step 1, based on the visual task and the unloading selection strategy matrix, set the carbon emission matrix and the computing resource load matrix, each row of the carbon emission matrix increases in turn according to the order of the end, the edge and the cloud; when outputting the task deployment strategy, it is necessary to ensure that each row of the task deployment strategy has only one element of 1, and meet the demand of the computing resource load matrix;

[0016] Step 2, select the edge server with the minimum carbon emission for each visual task, subtract the corresponding carbon emission from the corresponding row and column of the carbon emission matrix, and subtract the corresponding load from the corresponding server of the computing resource load matrix;

[0017] Step 3, according to the 0 value of the carbon emission matrix and the computing resource load matrix, assign the execution device of the visual task, if the assigned 0 value is equal to the total number of tasks of the plurality of visual tasks, end; otherwise, execute the step 2 again.

[0018] The carbon emission evaluation method of the end-edge-cloud visual collaborative architecture, wherein the model construction step comprises:

[0019] The cloud server comprises computing devices for training, learning and data calculation, a refrigeration system and basic devices for cloud operation; the energy consumption of all devices running in the whole cloud center Divide the energy consumption of IT devices running , the energy utilization efficiency of the cloud server ;

[0020] The refrigeration coefficient of the cloud server is , wherein , and are real numbers, is the temperature set by the refrigeration device, and the running power consumption of the cloud server is ; the carbon emission and the power carbon intensity of energy consumption ; wherein and represent the renewable power generation energy set and the regional set respectively; represents the renewable energy type power consumption in the region ; is the carbon emission rate of the renewable energy type ; the computing resource load rate of the cloud server is , wherein and are the current load and the maximum load of the cloud server respectively; the carbon emission generated by the operation of the cloud server is ;

[0021] Based on the carbon emissions from the refrigerant leak of the cloud server and the quality of the refrigerant used. Global warming potential (GWP), refrigerant leakage rate and the overall uptime of the equipment The carbon emissions generated by the use of refrigerants ;

[0022] The carbon emission model for this cloud device is as follows:

[0023]

[0024] The edge server includes IT equipment for edge computing and basic system operating equipment; this edge-cloud vision collaborative IoT includes... Each edge server; obtain the power energy utilization efficiency of each edge server. ;

[0025] Obtain the computing power consumption of the edge server's own computing device. and running time The computing power consumption of edge servers With full load power consumption during operation Power consumption during idle time Current required workload and maximum workload Relevant; the load rate of this edge server; The power consumption is: The edge server's working and idle times are divided into two segments. and Because different sites provide different hybrid energy sources to the edge servers, the carbon intensity of electricity at each edge is... Difference, i.e., edge end The overall carbon emission calculation model is as follows: The carbon emission model for this equipment is as follows:

[0026]

[0027] Traditional visual terminals use visual perception to compress captured videos, upload and store them, and then analyze, identify, and process them at the edge and cloud. Intelligent visual terminals, on the other hand, possess visual feature extraction capabilities and store the compressed video stream locally, retrieving it on demand based on edge and cloud requirements. They also support the migration and compression of video encoding and feature analysis models between the edge and cloud. This enables the construction of visual terminals. Equipment categories If it is a traditional visual terminal, then Otherwise ;

[0028] This traditional visual terminal is based on its power consumption and running time and the area where it is located Types of energy that use electricity and electrical carbon strength Based on the current required workload The energy consumption model of this traditional visual terminal ;

[0029] This intelligent vision terminal also has data analysis capabilities, but its power consumption efficiency is lower than that of edge devices. Based on the power consumption of this intelligent vision terminal Running time Full load power consumption during operation Power consumption during idle time Current required workload and maximum workload The load rate of this intelligent vision terminal is: The power consumption of this intelligent vision terminal is... The intelligent vision terminal Work and free time are divided into two segments. and The carbon intensity of this intelligent vision terminal With the region and types of energy that generate electricity The carbon emission model for this intelligent vision terminal is as follows: ;

[0030] The carbon emission model for this terminal device is as follows:

[0031]

[0032] Obtain the network transmission carbon emissions and communication transmission efficiency under edge-cloud collaboration in this edge-cloud visual collaborative IoT. The amount of data transmitted Energy consumption during transmission and electrical carbon strength ; Obtain the energy consumption per unit bit of data in this edge-cloud visual collaborative IoT. ;

[0033] The amount of data transmitted by edge-cloud vision-coordinated IoT With the resolution of video frames Related, that is ,in It is a fixed value; and the energy consumption of network transmission is the same as the energy consumption per unit bit of data. Related to the resolution of video frames. At the node and nodes The carbon emission model for transmission between nodes is: ; network transmission carbon emission and energy consumption per unit bit data volume , electric carbon intensity , video frame resolution , frame sampling rate , and communication transmission efficiency The carbon emission model for transmission between nodes is:

[0034]

[0035] The carbon emission evaluation method of the end-edge-cloud vision collaborative architecture, wherein

[0036] The offloading selection strategy matrix , x belongs to 0 or 1 decision task is executed in the intelligent vision terminal, edge server or cloud server; m is the id of the terminal device; V is the set of all vision tasks; wherein, indicates that the intelligent vision terminal is locally offloaded, indicates that the edge server is offloaded, indicates that the cloud server is offloaded; for the edge server , the load rate of the edge computing resource is; ; for the cloud server, the load rate of the computing resource is: The overall carbon emission model is:

[0037]

[0038] wherein the first part is the carbon emission model of local offloading; the second part is the carbon emission model of edge offloading; the third part is the carbon emission model of cloud offloading; and the fourth part is the network transmission carbon emission model of terminal transmission to the edge and the cloud.

[0039] As shown in Figure 5 , the present application further provides a carbon emission evaluation device for an end-edge-cloud vision collaborative architecture, which comprises:

[0040] An initial module acquires an end-edge-cloud vision collaborative Internet of Things including a traditional vision terminal, an intelligent vision terminal, an edge server and a cloud server;

[0041] A model construction module respectively constructs an end device carbon emission model, an edge device carbon emission model and a cloud device carbon emission model according to the carbon emission parameters of the traditional vision terminal, the intelligent vision terminal, the edge server and the cloud server according to functions, and constructs a network transmission carbon emission model according to the carbon emission parameters generated by network transmission during operation of the end-edge-cloud vision collaborative Internet of Things;

[0042] The task preparation module obtains a plurality of visual tasks, constructs an offloading selection strategy matrix of the plurality of visual tasks according to a deployment strategy of each visual task, constructs a carbon emission matrix corresponding to the offloading selection strategy matrix according to the network transmission carbon emission model, the end device carbon emission model, the edge device carbon emission model and the cloud device carbon emission model, and the carbon emission matrix represents carbon emissions generated by each visual task in each server in the end-edge-cloud visual collaborative Internet of Things; and constructs a computing resource load matrix according to the upper limit of the computing resources of each device in the end-edge-cloud visual collaborative Internet of Things.

[0043] The task deployment module uses the Hungarian algorithm to solve the offloading selection strategy matrix when the carbon emission of the network transmission carbon emission model is the smallest and the computing resources occupied by the visual tasks are less than or equal to the upper limit of the computing resources of each device, as the task deployment strategy of the plurality of visual tasks, and executes the plurality of visual tasks using the task deployment.

[0044] The carbon emission evaluation device for the end-edge-cloud visual collaborative architecture, wherein the task deployment module comprises:

[0045] Module 1: based on the visual tasks and the offloading selection strategy matrix, set the carbon emission matrix and the computing resource load matrix, each row in the carbon emission matrix is sequentially increased according to the order of the end, the edge and the cloud; when the task deployment strategy is output, it is necessary to ensure that there is only one element of 1 in each row of the task deployment strategy, and the demand of the computing resource load matrix is met;

[0046] Module 2: each visual task selects the edge server with the smallest carbon emission, and the corresponding row and column of the carbon emission matrix are subtracted by the corresponding carbon emission, and the corresponding server of the computing resource load matrix is subtracted by the corresponding load;

[0047] Module 3: according to the 0 value of the carbon emission matrix and the computing resource load matrix, assign the execution device of the visual task, if the assigned 0 value is equal to the total number of tasks of the plurality of visual tasks, end; otherwise, call the module 2 again.

[0048] The carbon emission evaluation device for the end-edge-cloud visual collaborative architecture, wherein the model construction module comprises:

[0049] The cloud server includes a computing device for training, learning and data calculation, a refrigeration system and a basic device for cloud operation; the energy consumption of all devices running through the whole cloud center Divide the energy consumption of the IT device running , to obtain the energy utilization efficiency of the cloud server ;

[0050] The refrigeration coefficient of the cloud server is , wherein , and is a real number, is the temperature set for the refrigeration equipment, the operating power consumption of the cloud server is ; the carbon emission and the power carbon intensity of energy consumption ; wherein and respectively represent the renewable power generation energy set and the regional set; represents the renewable energy type power consumption in the region ; is the carbon emission rate of the renewable energy type ; the computing resource load rate of the cloud server is , wherein and are the current load and the maximum load of the cloud server respectively; the carbon emission generated by the operation of the cloud server is ;

[0051] According to the carbon emission generated by the refrigerant leakage of the cloud server and the use quality of the refrigerant, the global warming potential GWP, the leakage rate of the refrigerant , and the overall operation time of the equipment , the carbon emission generated by the use of the refrigerant is constructed;

[0052] The cloud device carbon emission model is:

[0053]

[0054] The edge server includes IT equipment and system basic operation equipment for edge computing; the end-edge-cloud visual collaborative Internet of Things includes edge servers; the power energy utilization efficiency of each edge server is obtained ;

[0055] The computing power consumption and the operation time of the computing device of the edge server itself are obtained, and the computing power consumption of the edge server is related to the full load power consumption , the power consumption when idle , the current required workload and the maximum workload ; the load rate of the edge server is , and the power consumption is ; the time of the edge server working and idling is divided into two segments and Because different sites provide different hybrid energy sources to the edge servers, the carbon intensity of electricity at each edge is... Difference, i.e., edge end The overall carbon emission calculation model is as follows: The carbon emission model for this equipment is as follows:

[0056]

[0057] Traditional visual terminals use visual perception to compress captured videos, upload and store them, and then analyze, identify, and process them at the edge and cloud. Intelligent visual terminals, on the other hand, possess visual feature extraction capabilities and store the compressed video stream locally, retrieving it on demand based on edge and cloud requirements. They also support the migration and compression of video encoding and feature analysis models between the edge and cloud. This enables the construction of visual terminals. Equipment categories If it is a traditional visual terminal, then Otherwise ;

[0058] This traditional visual terminal is based on its power consumption and running time and the area where it is located Types of energy that use electricity and electrical carbon strength Based on the current required workload The energy consumption model of this traditional visual terminal ;

[0059] This intelligent vision terminal also has data analysis capabilities, but its power consumption efficiency is lower than that of edge devices. Based on the power consumption of this intelligent vision terminal Running time Full load power consumption during operation Power consumption during idle time Current required workload and maximum workload The load rate of this intelligent vision terminal is: The power consumption of this intelligent vision terminal is... The intelligent vision terminal Work and free time are divided into two segments. and The carbon intensity of this intelligent vision terminal With the region and types of energy that generate electricity The carbon emission model for this intelligent vision terminal is as follows: ;

[0060] The carbon emission model for this terminal device is as follows:

[0061]

[0062] The network transmission carbon emission and communication transmission efficiency in the edge-cloud vision collaborative Internet of Things under the edge-cloud collaboration are obtained , the amount of data transmitted , the energy consumption consumed by transmission , and the electric carbon intensity ; the energy consumption consumed per bit of data amount in the edge-cloud vision collaborative Internet of Things is obtained ;

[0063] The amount of data transmitted in the edge-cloud vision collaborative Internet of Things is related to the resolution of the video frame , that is , wherein is a fixed value; and the network transmission energy consumption is related to the energy consumption consumed per bit of data amount , the resolution of the video frame transmitted between the node and the node , the network transmission carbon emission model is ; the network transmission carbon emission is related to the energy consumption consumed per bit of data amount , the electric carbon intensity , the resolution of the video frame , the frame sampling rate , and the communication transmission efficiency , and the network transmission carbon emission model is

[0064]

[0065] The offloading selection strategy matrix , x belongs to 0 or 1 decision task is executed in the intelligent vision terminal local, edge server or cloud server; m is the id of the terminal device; V is the set of all vision tasks; wherein represents the intelligent vision terminal local offloading, represents the edge server offloading, represents the cloud server offloading; for the edge server , the load rate of the edge computing resource is ; for the cloud server, the load rate of the computing resource is , the overall carbon emission model is

[0066]

[0067] The first part is a local offloading carbon emission model; the second part is an edge offloading carbon emission model; the third part is a cloud offloading carbon emission model; and the fourth part is a network transmission carbon emission model of terminal transmission to the edge and the cloud.

[0068] The application further provides an electronic device comprising the carbon emission evaluation device for the end-edge-cloud visual collaborative architecture.

[0069] The application further provides a computer readable storage medium having a computer program stored thereon, and the computer program is executed by a processor to implement the steps of the carbon emission evaluation method for the end-edge-cloud visual collaborative architecture.

[0070] The application further provides a computer program product comprising a computer program, and the computer program is executed by a processor to implement the steps of any of the carbon emission evaluation methods for the end-edge-cloud visual collaborative architecture. BRIEF DESCRIPTION OF DRAWINGS

[0071] Figure 1 The end-edge-cloud collaborative carbon emission architecture of the application;

[0072] Figure 2 The end-edge-cloud collaborative carbon emission related parameter diagram of the application;

[0073] Figure 3 The carbon emission evaluation process diagram under the end-edge-cloud collaboration of the application;

[0074] Figure 4 The method flowchart of the application;

[0075] Figure 5 The device module diagram of the application;

[0076] Figure 6 The first electronic device structure schematic diagram of the application;

[0077] Figure 7 The first electronic device application environment structure schematic diagram of the application;

[0078] Figure 8 The second electronic device structure schematic diagram of the application.

[0079] REFERENCE NUMERALS:

[0080] A-First electronic device;

[0081] B-Carbon emission evaluation device for the end-edge-cloud visual collaborative architecture;

[0082] C-Data acquisition device;

[0083] D-information display device

[0084] 1000-second electronic device

[0085] I-computing unit

[0086] II-ROM

[0087] III-RAM

[0088] IV-bus

[0089] V-interface

[0090] VI-input unit

[0091] VII-output unit

[0092] VIII-storage medium

[0093] IX-communication unit DETAILED DESCRIPTION

[0094] It should be noted that the relationship terms such as first and second, and the like, are used only to differentiate one entity or operation from another, and do not necessarily require or imply such actual relationship or order between these entities or operations. In addition, the terms "comprising", "containing", or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, method, article, or apparatus including a list of elements does not only include those elements, but also includes other elements not explicitly listed, or further includes elements inherent in such process, method, article, or apparatus.

[0095] Without more limitations, the element defined by the phrase "comprising a" does not exclude the presence of additional identical elements in the process, method, article, or apparatus including the element.

[0096] The processor of the present application is the control center of the electronic device, which can be one processor or a collective term of multiple processing elements. For example, it is one or more central processing units (CPU), or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement embodiments of the present application, such as one or more microprocessors (digital signal processor, DSP), or one or more field programmable gate arrays (FPGA).

[0097] Optionally, the processor can execute various functions of the electronic device by running or executing software programs stored in the memory and calling data stored in the memory.

[0098] In a specific implementation, as an example, the processor can include one or more CPUs. Each of the processors can be a single-CPU or a multi-CPU. The processor herein can refer to one or more devices, circuits, and / or processing cores for processing data (e.g., computer program instructions). The electronic device can include a server, a desktop computer, a notebook computer, a smart phone, a tablet computer, an embedded computer, and the like, wherein the embedded computer includes a vehicle and a robot, and the like.

[0099] The memory is used to store software programs for implementing the solutions of the present application, and is controlled by the processor to execute. The specific implementation can refer to the above-mentioned method embodiments, which will not be described here.

[0100] It should be noted that the structure of the electronic device shown in the drawings of the present application does not constitute a limitation thereon, and the actual knowledge structure recognition device can include more or fewer components than shown, or combine certain components, or different component arrangements.

[0101] The above embodiments can be implemented, wholly or partially, by software, hardware (such as a circuit), firmware, or any other combination. When implemented by software, the above embodiments can be implemented, wholly or partially, in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the flow or function described in the embodiments of the present application is wholly or partially generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium or transferred from one computer-readable storage medium to another, for example, the computer instructions can be transferred from one website, computer, server, or data center to another website, computer, server, or data center by wired (such as infrared, wireless, microwave, etc.) mode. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server, data center, etc. containing one or more available medium collections. The available medium can be a magnetic medium (such as a floppy disk, a hard disk, a magnetic tape), an optical medium (such as a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state disk.

[0102] It should also be understood that, in the specification, terms "and / or" merely describes an associated relationship, which means that there can be three relationships, for example, A and / or B can mean that A exists alone, A and B exist together, and B exists alone, where A and B can be singular or plural. In addition, the character " / " generally represents an "or" relationship between the front and rear associated objects, but can also represent an "and / or" relationship, which can be understood in combination with the context before and after.

[0103] In the present application, "at least one" means one or more, and "multiple" means two or more. "At least one of the following" or the like means any combination of the items, including any combination of single or multiple items. For example, at least one of a, b, or c can mean a, b, c, a-b, a-c, b-c, or a-b-c, where a, b, and c can be singular or plural.

[0104] It should also be understood that the order of the above processes in various embodiments of the present application does not mean the order of execution, and the execution order of the processes should be determined by its function and inherent logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0105] In several embodiments provided by the present application, it should be understood that the disclosed devices, apparatuses and methods can be implemented in other ways. For example, the above-described device embodiments are only schematic, and the division of the units is only a logical function division, and actual implementation can have another division manner, for example, a plurality of units or components can be combined or integrated into another device, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the units shown or discussed can be indirect coupling or communication connection through some interface, device or unit, and can be electrical, mechanical or other forms.

[0106] The units described as separate components can or can not be physically separated, and the components shown as units can or can not be physical units, that is, they can be located in one place, or can be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the present embodiment.

[0107] In addition, each functional unit in each embodiment of the present application can be integrated into a processing unit, or each unit can exist physically, or two or more units can be integrated into one unit.

[0108] If the functions are realized in the form of software function units and sold or used as independent products, they can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application or the parts of the present application that essentially contribute to the prior art or the parts of the technical solutions of the present application can be embodied in the form of a software product, which is stored in a storage medium and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in the embodiments of the present application. The aforementioned storage medium includes a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and various media that can store program codes.

[0109] The present application adopts the following technical solutions:

[0110] In a first aspect, according to an end-edge-cloud visual system, an end-edge-cloud system carbon emission evaluation model is constructed. As shown in Figure 1 the end-edge-cloud visual architecture. At the cloud level, the carbon emission of the system is related to the energy consumption of cloud computing, the operating energy consumption of the refrigeration system, the carbon emission released by the leakage of the refrigerant used in the refrigeration system, and the carbon emission generated by the energy consumption of the infrastructure operation; at the edge level, the carbon emission of the system operation is related to the energy consumption of edge computing and the energy consumption of the infrastructure operation; and at the terminal, it is divided into a traditional visual terminal and an intelligent visual terminal. It is assumed that the system operating energy consumption of the traditional visual terminal is determined by its own power consumption and time; and the intelligent visual terminal has an additional energy consumption of end computing operation compared with the traditional visual terminal. Therefore, the carbon emissions generated at the cloud layer, the edge layer, the terminal layer and the network layer are evaluated respectively.

[0111] Cloud layer carbon emission evaluation: as Figure 2 shown, the cloud layer carbon emission calculation involves parameters including: carbon emission evaluation of cloud service and power energy utilization efficiency (Power Usage Effectiveness, PUE) of cloud device , power carbon intensity of cloud device , operating power consumption of evaluation and calculation stage device , operating time , load rate of cloud service center computing resource , and use quality of refrigerant and use time . The operating power consumption of the cloud device includes calculation parameters: computing power consumption of the cloud server and the temperature set by the refrigeration device And the power carbon intensity of the cloud device is renewable energy type And the region is related And the carbon emission rate of the renewable energy type And the carbon emission caused by refrigerant leakage is related to the refrigerant use quality (kg) , the global warming potential (GWP), the refrigerant leakage rate (% / year) , and the working time period (years) .

[0112] Based on the current energy type used for power generation: for example: coal, nuclear power, coal, natural gas, oil, water power, wind energy, etc. Cloud power energy: due to the huge demand for electricity, the energy used in this model includes coal, nuclear power, coal, natural gas, oil, water power, wind energy, etc. These all have carbon emission statistics. Edge side: both commercial electricity and small power generation equipment can be used, and other green energy can be used. End side: commercial electricity energy is used.

[0113] Edge layer carbon emission evaluation: as Figure 2 , the edge layer carbon emission calculation involves parameters including: the power energy utilization efficiency of each edge terminal , the power carbon intensity of the edge terminal , the computing power consumption of the edge terminal device , the full load power consumption during device operation , the idle power consumption , the operation time , the idle operation time , the load rate of edge computing resources . The overall operation energy consumption of the edge server is related to the visual task being worked on and the operation of the underlying device. Here, the computing power consumption of the edge terminal device includes: full load power consumption during device operation , idle power consumption , and load rate of edge computing resources . The energy consumption of the edge device is provided by a mixture of different sites, which is different from the cloud layer, and the power carbon intensity is obtained according to the type of energy used .

[0114] Terminal layer carbon emission evaluation: as Figure 2 , the terminal layer carbon emission calculation involves parameters including: the device category of the visual terminal , the terminal carbon intensity , the computing power consumption ​, full load power consumption of the device runtime of the intelligent terminal and idle power consumption , runtime , idle runtime , intelligent terminal computing resource load rate . Among them, the carbon intensity of the terminal , the computing power consumption and the runtime are the carbon emission calculation parameters of the traditional terminal. The computing power consumption parameters of the intelligent visual terminal are similar to the edge device, including full load power consumption , idle power consumption and runtime , idle runtime , intelligent terminal computing resource load rate . In addition, the carbon intensity of the terminal layer is similar to the cloud.

[0115] Network layer carbon emission evaluation: as Figure 2 , the network layer carbon emission calculation involves parameters related to energy consumption per unit bit data volume , communication transmission efficiency , video frame resolution , frame sampling rate , and carbon intensity . Among them, the resolution of the video frame determines the data volume on the transmission path.

[0116] Secondly, the carbon emission evaluation scheme under the end-edge-cloud collaboration. As Figure 3 shown in the process, according to the end-edge-cloud system carbon emission evaluation model, the carbon emission model scheme based on a single device\subsystem is constructed; and through the end-edge-cloud collaborative task offloading model, the overall system carbon emission evaluation model is constructed, and the appropriate end-edge-cloud collaborative offloading strategy is obtained by using the many-to-many Hungarian matching strategy.

[0117] The advantages and beneficial effects of the present application are: based on the carbon footprint under the visual Internet of Things architecture, the carbon emissions of each subsystem device of the end-edge-cloud under the architecture are modeled and unified. At the same time, the network level carbon emission evaluation model of each subsystem device data transmission under the end-edge-cloud collaboration is provided. Considering the diversity of large-scale visual tasks, the end-edge-cloud collaborative system carbon emission optimization model is constructed, and a many-to-many Hungarian matching algorithm is provided to determine the decision scheme of each visual task executed locally, at the edge or in the cloud.

[0118] In order to make the above features and effects of the present application more explicit and easy to understand, the following embodiments are specifically described below, and the detailed description is made in conjunction with the accompanying drawings. The present specification discloses one or more embodiments comprising the features of the present application. The disclosed embodiments are only for illustration. The scope of protection of the present application is not limited to the disclosed embodiments, and the present application is defined by the appended claims.

[0119] The present application mainly describes the carbon emissions of different devices through cloud layer carbon emission evaluation, edge layer carbon emission evaluation, terminal layer carbon emission evaluation and network layer carbon emission evaluation, and proposes an end-edge-cloud collaborative carbon emission evaluation technology according to a carbon emission evaluation model.

[0120] I. Carbon emission evaluation model of end-edge-cloud system

[0121] As Figure 1 , the present application is a specific end-edge-cloud collaborative visual Internet of Things architecture, and the system network is composed of one conventional visual terminal device, one intelligent visual terminal device, a cloud server and one edge server and a plurality of infrastructures, the terminal set is defined as , the server set is defined as , wherein is the cloud server. The terminal is divided into two categories: the conventional visual terminal is responsible for video acquisition, acquisition, encoding and transmission; the intelligent terminal, in addition to the above functions, also includes video data analysis, processing and control functions. The edge server is responsible for local data analysis and processing, and transmits part of the unprocessed data to the cloud, and returns the related results to the client. In addition to analyzing and processing the data of the entire system, the cloud server also needs to implement control functions such as distributing instructions to the entire system. Due to the different functions and the different basic running devices contained in each node, the energy consumption required is different, and the corresponding carbon emissions are also different. The carbon emission evaluation method of the network transmission between different terminals, edges and clouds is described in detail below, in order to realize the evaluation and prediction of the overall carbon emission under the operation of the entire system.

[0122] (1) Cloud layer carbon emission evaluation

[0123] The devices of the cloud service include IT devices for training, learning and data calculation, refrigeration systems and basic devices for cloud operation. Therefore, in the end-edge-cloud system, the carbon footprint of the cloud service center is primarily considered the power energy utilization efficiency (Power Usage Effectiveness, PUE), which is equal to the energy consumption of all devices running in the cloud center divided by the energy consumption of the IT device running Consider the power consumption of the device at different evaluation periods and running time to obtain the power utilization efficiency PUE of the power source.

[0124] In addition to considering the power consumption of the IT device itself and running time , the power consumption and running time of air conditioners and other devices are also considered. Here, the power consumption of the cloud server IT device is defined in relation to the refrigeration coefficient. The refrigeration coefficient is , where , and are heterogeneous real numbers, is the temperature set by the refrigeration device and the like. The actual running power consumption of the cloud server is . The carbon emission of the cloud device and the carbon intensity (g / KW) of the power consumption are . Among them and represent the renewable power generation energy set and the regional set, respectively; represents the type of renewable energy and the power consumption in the region ; and is the carbon emission rate of the renewable energy type . Let the load rate of the cloud service center computing resource be , where and are the current required load and the maximum load of the cloud service center, respectively. Therefore, the carbon emission generated by the energy consumption of the cloud device running is

[0125] In addition, since the cloud uses air conditioners and other refrigeration devices to maintain the operation of the server and other IT devices and ensure the safety of the cloud device, in addition to the carbon emission generated by power consumption, there is a carbon emission risk of refrigerant leakage. The carbon emission generated by refrigerant leakage is related to the use quality (kg) of the refrigerant, the global warming potential (GWP), the leakage rate of the refrigerant (% / year) , and the overall running time of the device (years) . The carbon emission calculation formula caused by the use of refrigerant is

[0126] Therefore, the evaluation of the carbon emission of the cloud includes the carbon emission caused by energy consumption and the carbon emission caused by the use of refrigerant and the like, using power consumption / time / carbon intensity index / usage of refrigerant / leakage rate of refrigerant parameters; wherein the power consumption, time, carbon intensity index, etc. determine the carbon emission model of energy consumption ; the usage of refrigeration equipment and the leakage rate of refrigerant and the refrigerant usage time period determine the carbon emission model caused by the use of refrigerant . The overall cloud carbon emission evaluation model is:

[0127] (1)

[0128] (2) Edge layer carbon emission evaluation

[0129] The edge device center mainly includes IT equipment and system basic operation equipment for edge computing. However, under the vision of IoT collaborative edge cloud , the edge layer here is the sum of the carbon emissions of all edge layers. Similar to the cloud server, the power energy utilization efficiency of each edge layer needs to be considered firstly, and since there is no related refrigeration equipment compared with the cloud server center, the PUE here is reduced, that is .

[0130] The overall operation carbon emission of the edge layer mainly considers the computing power consumption and operation time of the IT equipment itself , and the computing power consumption of the edge server is related to the full load power consumption , the power consumption when idle , the current required workload and the maximum workload of the device when running. Let the load rate of edge computing resources be , that is, the current computing power consumption is . Therefore, the time when the edge layer is working and idle is divided into two segments and . Since different sites provide mixed energy to the edge server, the power carbon intensity of each edge layer is different, that is, the overall carbon emission calculation model of the edge layer is . The overall carbon emission calculation model of the edge layer is

[0131] (2)

[0132] (3) Terminal layer carbon emission evaluation

[0133] The terminal defined in this invention is a visual device, which is divided into two categories: traditional terminals and smart terminals. Traditional cameras use visual perception to compress captured video, upload and store it, and then analyze, identify, and process it at the edge and cloud. Smart cameras possess high-quality encoding and visual feature extraction capabilities, and store the compressed video stream locally, retrieving it on demand according to edge and cloud requirements. They also support the migration and compression of video encoding and feature analysis models between the edge and cloud. This invention enables visual terminals... The equipment category is If it is a traditional terminal Otherwise .

[0134] For traditional terminals, electricity consumption is generally categorized into four main types: residential, agricultural, large industrial, and general commercial / industrial. Therefore, this analysis primarily focuses on the camera's computational power consumption. and running time and regions Types of energy that use electricity and electrical carbon strength Relevant, and the current required workload is .Right now .

[0135] In addition to visual perception capabilities, smart terminals also possess certain data analysis capabilities, and therefore also have a certain power usage efficiency (PUE), which is lower than that of edge devices. Therefore, smart terminals can be viewed as analyzable small computing nodes; that is, the carbon emissions from the operation of a smart terminal mainly depend on its own computing power consumption. and running time Similar to edge computing, smart terminals also have high computing power consumption. With the device's full-load power consumption during operation Power consumption during idle time Current required workload and maximum workload Related. This sets the load rate of the computing resources on the smart terminal at; That is, the current computing power consumption is: Therefore, smart terminals Work and leisure time can also be divided into two segments. and Similar to traditional terminals, smart terminals also use electricity sold through sales; therefore, their electrical carbon intensity... With the region and types of energy that generate electricity Related. That is, the carbon emission model is: .

[0136] As shown above, the carbon emission model for all terminals is as follows:

[0137] (3)

[0138] (4) Carbon emission assessment of network layer

[0139] The edge-cloud collaborative visual IoT architecture achieves full-process data transmission, cleaning, synchronization, and retrieval through collaboration between endpoints, edges, and cloud devices. Since 2010, global internet traffic has increased 25-fold, undoubtedly leading to increased energy consumption in data transmission networks and a significant increase in carbon emissions. This paper discusses the relationship between network transmission carbon emissions and communication transmission efficiency under edge-cloud collaboration. The amount of data transmitted Energy consumption during transmission and electrical carbon strength Related. And communication transmission efficiency and the energy consumed in transmission All of these are related to the transmission method (such as wireless network transmission and core network transmission). Wireless network transmission is less efficient than core network transmission and consumes more energy. The energy consumption per unit bit of data varies depending on the transmission method. Different (J / bit) results in different carbon emissions.

[0140] Data volume transmitted over the network in the visual Internet of Things With the resolution of video frames Related, that is ,in It is a fixed value. Furthermore, the network transmission energy consumption is the same as the energy consumption per unit bit of data. (J / bit) is related to the resolution of the video frame, depending on the node. At the node and nodes The carbon emission transfer model between them is as follows: Carbon emissions from network transmission and energy consumption per unit bit of data. Electrocarbon strength Video frame resolution Frame sampling rate and communication transmission efficiency The overall calculation model is as follows:

[0141] (4)

[0142] II. Carbon Emission Assessment Scheme under Edge-Cloud Collaboration

[0143] Based on the constructed edge-cloud carbon emission assessment method, this invention proposes a carbon emission assessment technology under edge-cloud collaboration, which is applied to the field of edge-cloud collaborative vision. The specific steps are as follows:

[0144] (1) Carbon emission model measurement of individual subsystems / equipment

[0145] First, at the terminal, carbon emission calculations are based on local calculations and quantified conversion parameters to achieve the terminal-level carbon emission model design. This clarifies the visual terminal. Equipment categories If it is a traditional terminal Otherwise, it is an intelligent vision terminal. The parameters and characteristics of the terminal in each time period are analyzed, i.e., the task requirements. , and Etc., clarify the carbon emission model for this task to be performed locally. Terminal tasks that cannot be processed locally are uploaded to the corresponding edge server or cloud server based on proximity and latency requirements. The latency requirement is either a requirement inherent to the terminal task itself, or it is the maximum latency available in the current market based on the task scenario and type.

[0146] Next, at the edge or cloud, based on data such as video / compressed video / video features obtained from the terminal device, specific visual tasks and requirements are implemented, such as face recognition, vehicle counting, deep learning training, etc. This involves carbon emission model evaluation of edge and cloud sub-devices and carbon emission evaluation of the edge-cloud collaborative network. The carbon emission models of the cloud service center and each edge service center at this stage are recorded according to formulas (1) and (2). and Meanwhile, the carbon emission model for each transmission path is recorded according to formula (4). The time will be adjusted according to actual needs.

[0147] (2) Carbon emission model of edge-cloud collaborative system

[0148] Based on the carbon emission models of individual subsystems and devices, and taking resource consumption under the edge-cloud collaborative visual IoT architecture as the criterion, a new paradigm for optimizing system carbon emissions is established, reforming and innovating the overall deployment strategy of the edge-cloud system. First, two-dimensional offloading decision variables are defined. x represents a 0 or 1 decision task, executed locally, on edge 1, on edge 2, or in the cloud; m is the ID of the terminal device; V is the set of all terminal tasks. This indicates local uninstallation. This indicates unloading at the edge. This indicates uninstallation in the cloud. So, what about edge servers? , the load rate of edge computing resources is; . For the cloud server, the load rate of computing resources is: The collaborative optimization in the visual task offloading process is completed under the condition of ensuring the network service quality and visual task delay are not affected. Under the adaptive adjustment of the offloading scheme of different terminal visual tasks, the carbon emission target is achieved. The overall carbon emission model is designed as follows:

[0149] (5)

[0150] Among them, the first part is the carbon emission model of local offloading; the second part is the carbon emission model of edge offloading; the third part is the carbon emission model of cloud offloading; the fourth part is the network transmission carbon emission model of terminal transmission to edge and cloud.

[0151] (3) Carbon emission deployment scheme of end-edge-cloud collaborative system

[0152] The above carbon emission model can be regarded as a many-to-many matching problem. As a classic combinatorial optimization algorithm, the Hungarian algorithm can solve the allocation and deployment problem in polynomial time. Here, the Hungarian algorithm is used as the solution scheme of the carbon emission deployment of the end-edge-cloud collaborative system. Specifically as follows:

[0153] Step 1: First, consider the cloud as a matching node based on the near edge, based on the terminal visual task and offloading selection strategy matrix Set different ID labels (carbon emission and computing resource load) to generate the carbon emission matrix And the computing resource load matrix . Wherein the carbon emission matrix of each row increases in turn according to the order of end, edge and cloud. When the decision output , it is necessary to ensure that each row of the carbon emission matrix There is only one element 0, and each row of the decision output matrix There is only one element 1, and it meets the demand of the computing resource load matrix . It should be noted that the 0 of the carbon emission matrix corresponds to the 1 of the decision matrix x.

[0154] Step 2: Then according to the row and column reduction, that is, the visual task selected by the service node corresponds to the minimum carbon emission ; the overall carbon emission matrix is reduced by , and the corresponding server of the computing resource load matrix is reduced by the corresponding load, that is, .

[0155] Step 3: According to the carbon emission matrix all 0 values of the carbon emission matrix and the computing resource load matrix satisfies, sequentially find a trial assignment, if the independent 0 value of the assignment is equal to the number of visual tasks, end, at this time the row and column where 0 in the carbon emission matrix is located, namely the decision matrix ; otherwise, adjust the carbon emission matrix and the computing resource load matrix, and return to Step 2.

[0156] The following is a system embodiment corresponding to the above method embodiment. The technical details mentioned in the above embodiment are still valid in this embodiment. In order to reduce repetition, they will not be repeated here. Correspondingly, the technical details mentioned in this embodiment can also be applied in the above embodiment.

[0157] As shown in Figure 5 , the application further provides a carbon emission evaluation device for an end-edge-cloud visual collaborative architecture, which comprises:

[0158] An initial module acquires an end-edge-cloud visual collaborative Internet of Things including a traditional visual terminal, an intelligent visual terminal, an edge server and a cloud server;

[0159] A model construction module respectively constructs an end device carbon emission model, an edge device carbon emission model and a cloud device carbon emission model according to the carbon emission parameters of the traditional visual terminal, the intelligent visual terminal, the edge server and the cloud server according to functions, and constructs a network transmission carbon emission model according to the carbon emission parameters generated by network transmission when the end-edge-cloud visual collaborative Internet of Things runs;

[0160] A task preparation module acquires a plurality of visual tasks, constructs an offloading selection strategy matrix of the plurality of visual tasks according to the deployment strategy of each visual task, constructs a carbon emission matrix corresponding to the offloading selection strategy matrix according to the network transmission carbon emission model, the end device carbon emission model, the edge device carbon emission model and the cloud device carbon emission model, the carbon emission matrix representing the carbon emission generated by each server in the end-edge-cloud visual collaborative Internet of Things for each visual task, and constructs a computing resource load matrix according to the upper limit of the computing resources of each device in the end-edge-cloud visual collaborative Internet of Things;

[0161] A task deployment module uses the Hungarian algorithm to solve the offloading selection strategy matrix when the carbon emission of the network transmission carbon emission model is the smallest and the computing resources occupied by the visual tasks are less than or equal to the upper limit of the computing resources of each device, as the task deployment strategy of the plurality of visual tasks, and executes the plurality of visual tasks using the task deployment.

[0162] The carbon emission evaluation device for the end-edge-cloud visual collaborative architecture, wherein the task deployment module comprises:

[0163] Module 1, based on the visual task and the unloading selection strategy matrix, sets the carbon emission matrix and the computing resource load matrix, each row of the carbon emission matrix increases in turn according to the order of the end, the edge and the cloud; when the task deployment strategy is output, it is necessary to ensure that each row of the task deployment strategy has only one element of 1, and meets the demand of the computing resource load matrix;

[0164] Module 2, each of the visual tasks selects the edge server with the minimum carbon emission, the corresponding row and column of the carbon emission matrix are subtracted by the corresponding carbon emission, and the corresponding server of the computing resource load matrix is subtracted by the corresponding load;

[0165] Module 3, according to the 0 value of the carbon emission matrix and the computing resource load matrix, the execution device of the visual task is assigned, if the assigned 0 value is equal to the total number of tasks of the plurality of visual tasks, the process is ended; otherwise, the module 2 is called again.

[0166] The carbon emission evaluation device of the end-edge-cloud visual collaborative architecture, wherein the model construction module comprises:

[0167] The cloud server comprises a computing device for training, learning and data calculation, a refrigeration system and a basic device for cloud operation; the energy consumption of all devices running in the whole cloud center Divided by the energy consumption of the IT device running , the energy utilization efficiency of the cloud server ;

[0168] The refrigeration coefficient of the cloud server is , wherein , and are real numbers, is the temperature set by the refrigeration device, and the running power consumption of the cloud server is ; the carbon emission and the power carbon intensity consumed by the energy consumption ; wherein and respectively represent a renewable power generation energy set and a region set; represents the type of renewable energy on the region ; is the carbon emission rate of the renewable energy type ; the computing resource load rate of the cloud server is , wherein and are the current load and the maximum load of the cloud server respectively; the carbon emission generated by the operation of the cloud server is ;

[0169] Based on the carbon emissions from the refrigerant leak of the cloud server and the quality of the refrigerant used. Global warming potential (GWP), refrigerant leakage rate and the overall uptime of the equipment The carbon emissions generated by the use of refrigerants ;

[0170] The carbon emission model for this cloud device is as follows:

[0171]

[0172] The edge server includes IT equipment for edge computing and basic system operating equipment; this edge-cloud vision collaborative IoT includes... Each edge server; obtain the power energy utilization efficiency of each edge server. ;

[0173] Obtain the computing power consumption of the edge server's own computing device. and running time The computing power consumption of edge servers With full load power consumption during operation Power consumption during idle time Current required workload and maximum workload Relevant; the load rate of this edge server; The power consumption is: The edge server's working and idle times are divided into two segments. and Because different sites provide different hybrid energy sources to the edge servers, the carbon intensity of electricity at each edge is... Difference, i.e., edge end The overall carbon emission calculation model is as follows: The carbon emission model for this equipment is as follows:

[0174]

[0175] Traditional visual terminals use visual perception to compress captured videos, upload and store them, and then analyze, identify, and process them at the edge and cloud. Intelligent visual terminals, on the other hand, possess visual feature extraction capabilities and store the compressed video stream locally, retrieving it on demand based on edge and cloud requirements. They also support the migration and compression of video encoding and feature analysis models between the edge and cloud. This enables the construction of visual terminals. Equipment categories If it is a traditional visual terminal, then Otherwise ;

[0176] The traditional visual terminal is based on its power consumption and running time , and the area , the type of energy used and the carbon intensity of electricity , according to the current required workload , the energy consumption model of the traditional visual terminal ;

[0177] The intelligent visual terminal also has data analysis capability, and its power energy use efficiency is lower than that of edge device, that is ; according to the power consumption , running time , full load power consumption at runtime , power consumption at idle time , current required workload and maximum workload of the intelligent visual terminal; the load rate of the intelligent visual terminal is ; the power consumption of the intelligent visual terminal is ; the intelligent visual terminal The time of work and idle time of the intelligent visual terminal is divided into two segments and ; the carbon intensity of electricity of the intelligent visual terminal and the area and the type of energy produced ; the carbon emission model of the intelligent visual terminal is ;

[0178] The carbon emission model of the terminal device is

[0179]

[0180] Get the network transmission carbon emission and communication transmission efficiency , the amount of data transmitted , the energy consumption consumed by transmission and the carbon intensity of electricity in the end-to-end cloud vision collaborative Internet of Things under the end-to-end cloud collaborative ;

[0181] The amount of data transmitted in the end-to-end cloud vision collaborative Internet of Things is related to the resolution of video frame , that is , where is a fixed value; and the network transmission energy consumption is related to the energy consumption per bit data , the resolution of video frame is related to the node and nodes The carbon emission model for the transmission between nodes is: ; the network transmission carbon emission and the energy consumption per unit bit data volume , the carbon intensity of electricity , the video frame resolution , the frame sampling rate , and the communication transmission efficiency , the network transmission carbon emission model is:

[0182]

[0183] The offloading selection strategy matrix , x belongs to 0 or 1 decision task is executed in the intelligent visual terminal, edge server or cloud server locally; m is the id of the terminal device; V is the set of all visual tasks; wherein, represents the intelligent visual terminal local offloading, represents the edge server offloading, represents the cloud server offloading; for the edge server , the load rate of the edge computing resource is; ; for the cloud server, the load rate of the computing resource is: The overall carbon emission model is:

[0184]

[0185] Wherein, the first part is the carbon emission model of local offloading; the second part is the carbon emission model of edge offloading; the third part is the carbon emission model of cloud offloading; the fourth part is the network transmission carbon emission model of terminal transmission to edge and cloud.

[0186] As shown in Figure 6 , the present application further proposes a first electronic device A, wherein the carbon emission evaluation device of the end-edge-cloud visual collaborative architecture is included.

[0187] As shown in Figure 7 , the first electronic device A can also be connected with the data acquisition device C and the information display device D through wired or wireless information transmission scheme, the data acquisition device C is used for acquiring a plurality of visual tasks, such as semantic segmentation task, image classification task, and the information display device D is used for displaying the task deployment strategy and task execution result obtained by the analysis of the present application.

[0188] The information display device D can process the data output by the first electronic device A based on an information display mechanism to improve the readability of the data output by the first electronic device A. The information display mechanism can be manually preset, for example, the data output by the first electronic device A is visually displayed according to the display parameters and / or attributes set by the user, for example, the display parameters can be the display data range, and the display attributes can be the display font, color, whether to scroll and play, etc. The user can understand the information more timely without accessing the secondary page or scrolling the page, which saves the operation of the user. Or the information display mechanism can be an artificial intelligence AI display model, which can learn the key information of the user according to the previous use habits of the user, for example, the viewing time, the number of clicks, the number of edits, etc., and then automatically present the user with rich and necessary key information.

[0189] The application further provides a computer program product, which comprises a computer program, the computer program can be stored on a readable storage medium, and the computer program can be executed by a processor to enable the computer to execute the carbon emission evaluation method of the end-edge cloud vision collaborative architecture provided by each method.

[0190] The present application also proposes a storage medium VIII for storing a computer program for implementing the carbon emission evaluation method of the edge-cloud vision collaborative architecture. It should be understood that the storage medium in the embodiments of the present application can be a volatile memory or a non-volatile memory, or can include both volatile and non-volatile memories. Among them, the non-volatile memory can be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM) or a flash memory. The volatile memory can be a random access memory (RAM) used as an external cache. By way of example, but not limitation, many forms of random access memory (RAM) are available, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchlink dynamic random access memory (SLDRAM) and direct memory bus random access memory (direct rambus RAM, DR RAM).

[0191] Figure 8 A schematic block diagram of a second electronic device 1000 that can be used to implement embodiments of the present application is shown. The second electronic device 1000 is intended to represent various forms of digital computers, such as laptops, desktops, tablets, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. The second electronic device 1000 can also represent various forms of mobile devices such as personal digital assistants, cellular telephones, smartphones, wearable devices, and other similar computing devices. The components shown here, their connections and relationships, and their functions, are meant to be examples only, and are not intended to limit the inventiveness in the present document as described and / or claimed. The second electronic device 1000 can be the same as or different from the first electronic device A.

[0192] The second electronic device 1000 includes a computing unit I that can perform various appropriate actions and processes in accordance with a computer program stored in a read-only memory II (ROM) or a computer program loaded into a random access memory (RAM) III from a storage medium VIII. In the RAM III, various programs and data required for the operation of the device 1000 can also be stored. The computing unit I, the ROM II, and the RAM III are connected to each other through a bus IV. An input / output (I / O) interface V is also connected to the bus IV.

[0193] A plurality of components in the second electronic device 1000 are connected to the I / O interface V, including an input unit VI such as a keyboard, a mouse, and the like, an output unit VII such as various types of displays, a speaker, and the like, a storage medium VIII such as a magnetic disk, an optical disk, and the like, and a communication unit IX such as a network card, a modem, a wireless communication transceiver, and the like. The communication unit IX allows the second electronic device 1000 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0194] The computing unit I can be various general and / or special-purpose processing components having processing and computing capabilities. Some examples of the computing unit I include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, and the like. The computing unit I performs various methods and processes described above, such as the method steps S1-S4. For example, in some embodiments, the method can be implemented as a computer software program tangibly embodied in a machine-readable medium, such as the storage medium VIII. In some embodiments, part or all of the computer program can be loaded and / or installed on the device 1000 via the ROM II and / or the communication unit IX. When the computer program is loaded into the RAM III and executed by the computing unit I, one or more steps of the method described above can be performed. Alternatively, in other embodiments, the computing unit I can be configured to perform the method by any other appropriate means, such as by means of firmware.

[0195] While the embodiments of the present application have been disclosed as above, they are not limited to only the applications listed in the specification and the embodiments, and can be fully applied to various fields suitable for the present application, and additional modifications can be easily made by those skilled in the art, and thus the present application is not limited to specific details and the figures shown and described herein, without departing from the general concept defined by the claims and the equivalent scope.

Claims

1. A carbon emission assessment method for an edge-cloud visual collaborative architecture, characterized in that, include: The initial step involves acquiring edge-cloud vision collaborative IoT, including traditional vision terminals, smart vision terminals, edge servers, and cloud servers. The model building steps are as follows: based on the carbon emission parameters of traditional visual terminals, intelligent visual terminals, edge servers, and cloud servers, respectively, end device carbon emission models, edge device carbon emission models, and cloud device carbon emission models are built according to their functions. Based on the carbon emission parameters generated by network transmission during the operation of the end-edge-cloud visual collaborative IoT, a network transmission carbon emission model is built. The task preparation steps include: acquiring multiple visual tasks; constructing an offloading selection strategy matrix for each visual task based on its deployment strategy; constructing a carbon emission matrix corresponding to the offloading selection strategy matrix based on the network transmission carbon emission model, the end device carbon emission model, the edge device carbon emission model, and the cloud device carbon emission model, whereby the carbon emission matrix represents the carbon emissions generated by each server in the edge-cloud visual collaborative IoT for each visual task; and constructing a computing resource load matrix based on the computing resource limits of each device in the edge-cloud visual collaborative IoT. The task deployment steps employ the Hungarian algorithm to solve for the offloading selection strategy matrix when the carbon emission model of the network transmission is minimized and the computing resources occupied by the visual tasks are less than or equal to the computing resource limits of each device. This matrix serves as the task deployment strategy for the multiple visual tasks, and the multiple visual tasks are deployed and executed using this strategy.

2. The carbon emission assessment method for edge-cloud visual collaborative architecture as described in claim 1, characterized in that, The deployment steps for this task include: Step 1: Based on the visual task and offloading selection strategy matrix, set up the carbon emission matrix and computing resource load matrix. Each row of the carbon emission matrix is ​​increased sequentially according to the order of end, edge and cloud. When the task deployment strategy is output, it is necessary to ensure that each row of the task deployment strategy has only one element with 1 and meets the requirements of the computing resource load matrix. Step 2: For each vision task, select the edge server with the lowest carbon emissions, subtract the corresponding carbon emissions from the corresponding rows and columns of the carbon emissions matrix, and calculate the resource load matrix by subtracting the corresponding load from the corresponding server. Step 3: Based on the 0 value of the carbon emission matrix and the computing resource load matrix, assign the execution device for the vision task. If the assigned 0 value is equal to the total number of vision tasks, then end; otherwise, repeat step 2.

3. The carbon emission assessment method for edge-cloud visual collaborative architecture as described in claim 1 or 2, characterized in that, The model construction steps include: The cloud server contains computing devices for training, learning, and data computation, a cooling system, and basic equipment for cloud operation; the energy consumption of all devices operating throughout the cloud center. Divide by the energy consumption of IT equipment operation The energy efficiency of the cloud server was obtained. ; The cooling coefficient of this cloud server is ,in , and For real numbers, The temperature set for the cooling equipment, and the power consumption of the cloud server are... , The computing power consumption of cloud servers; carbon emissions and the carbon intensity of electricity consumption. ;in and These represent the collection of renewable energy sources and the collection of regions, respectively. Indicates the type of renewable energy In the region Power consumption; For renewable energy types The carbon emission rate; the computing resource load rate of the cloud server: ,in and These are the current load and maximum load of the cloud server, respectively; the carbon emissions generated by the operation of this cloud server are: I c For electrical carbon intensity, Runtime; Based on the carbon emissions from the refrigerant leak of the cloud server and the quality of the refrigerant used. Global warming potential (GWP), refrigerant leakage rate and the overall uptime of the equipment The carbon emissions generated by the use of refrigerants ; The carbon emission model for this cloud device is as follows: The edge server includes IT equipment for edge computing and basic system operating equipment; this edge-cloud vision collaborative IoT includes... Each edge server; obtain the power energy utilization efficiency of each edge server. ; Obtain the computing power consumption of the edge server's own computing device. and running time The computing power consumption of edge servers With full load power consumption during operation Power consumption during idle time Current required workload and maximum workload Relevant; the load rate of this edge server is: The power consumption is: The edge server's working and idle times are divided into two segments. and Because different sites provide different hybrid energy sources to the edge servers, the carbon intensity of electricity at each edge is... Difference, i.e., edge end The overall carbon emission calculation model is as follows: The carbon emission model for this equipment is as follows: Traditional visual terminals use visual perception to compress captured videos, upload and store them, and then analyze, identify, and process them at the edge and cloud. Intelligent visual terminals, on the other hand, possess visual feature extraction capabilities and store the compressed video stream locally, retrieving it on demand based on edge and cloud requirements. They also support the migration and compression of video encoding and feature analysis models between the device, edge, and cloud. This enables the construction of visual terminals. Equipment categories If it is a traditional visual terminal, then Otherwise ; This traditional visual terminal is based on its power consumption and running time and the area where it is located Types of energy that use electricity and electrical carbon strength Based on the current required workload The carbon emission model of this traditional visual terminal ; This intelligent vision terminal also has data analysis capabilities, but its power consumption efficiency is lower than that of edge devices. Based on the power consumption of this intelligent vision terminal Running time Full load power consumption during operation Power consumption during idle time Current required workload and maximum workload The load rate of this intelligent vision terminal is: The power consumption of this intelligent vision terminal is... The intelligent vision terminal Work and free time are divided into two segments. and The carbon intensity of this intelligent vision terminal With the region and types of energy that generate electricity The carbon emission model for this intelligent vision terminal is as follows: ; The carbon emission model for this terminal device is as follows: Obtain the network transmission carbon emissions and communication transmission efficiency under edge-cloud collaboration in this edge-cloud visual collaborative IoT. The amount of data transmitted Energy consumption during transmission and electrical carbon strength ; Obtain the energy consumption per unit bit of data in this edge-cloud visual collaborative IoT. ; The amount of data transmitted by edge-cloud vision-coordinated IoT With the resolution of video frames Related, that is ,in It is a fixed value; and the energy consumption of network transmission is the same as the energy consumption per unit bit of data. Related to the resolution of video frames. At the node and nodes The carbon emission transfer model between them is as follows: Carbon emissions from network transmission and energy consumption per unit bit of data Electrocarbon strength Video frame resolution Frame sampling rate and communication transmission efficiency Relatedly, the carbon emission model for network transmission is as follows: .

4. The carbon emission assessment method for edge-cloud visual collaborative architecture as described in claim 3, characterized in that, The uninstallation selection strategy matrix In this context, x represents a 0 or 1 decision task executed locally on the intelligent vision terminal, on an edge server, or on a cloud server; m is the ID of the terminal device; and V is the set of all vision tasks. This indicates that the intelligent vision terminal will uninstall the software locally. This indicates that the data is being unloaded from the edge server. This indicates unloading on the cloud server; for edge servers... The load factor of edge computing resources is: For cloud servers, the load factor of computing resources is: Overall carbon emission model: The model consists of four parts: the first part is the carbon emission model for local offloading; the second part is the carbon emission model for edge offloading; the third part is the carbon emission model for cloud offloading; and the fourth part is the carbon emission model for network transmission from the terminal to the edge and the cloud.

5. A carbon emission assessment device for an edge-cloud visual collaborative architecture, characterized in that, include: The initial module acquires edge-cloud visual collaborative IoT, including traditional vision terminals, smart vision terminals, edge servers, and cloud servers; The model building module constructs carbon emission models for terminal devices, edge devices, and cloud devices based on the carbon emission parameters of traditional visual terminals, intelligent visual terminals, edge servers, and cloud servers, respectively, according to their functions. It also constructs a network transmission carbon emission model based on the carbon emission parameters generated by network transmission during the operation of the terminal-edge-cloud visual collaborative IoT. The task preparation module acquires multiple visual tasks and constructs an offloading selection strategy matrix for each visual task based on its deployment strategy. It then constructs a carbon emission matrix corresponding to the offloading selection strategy matrix, based on the network transmission carbon emission model, the end device carbon emission model, the edge device carbon emission model, and the cloud device carbon emission model. This carbon emission matrix represents the carbon emissions generated by each server in the edge-cloud visual collaborative IoT for each visual task. Finally, it constructs a computing resource load matrix based on the computing resource limits of each device in the edge-cloud visual collaborative IoT. The task deployment module uses the Hungarian algorithm to solve for the offloading selection strategy matrix when the carbon emission model of the network transmission is minimized and the computing resources occupied by the visual tasks are less than or equal to the computing resource limits of each device. This matrix serves as the task deployment strategy for the multiple visual tasks, and the multiple visual tasks are deployed and executed using this strategy.

6. The carbon emission assessment device for edge-cloud visual collaborative architecture as described in claim 5, characterized in that, The task deployment module includes: Module 1: Based on the visual task and offloading selection strategy matrix, set the carbon emission matrix and computing resource load matrix. Each row of the carbon emission matrix increases sequentially according to the order of end, edge, and cloud. When making a decision to output the task deployment strategy, it is necessary to ensure that each row of the task deployment strategy has only one element with 1, and meets the requirements of the computing resource load matrix. Module 2: For each visual task, select the edge server with the lowest carbon emissions, subtract the corresponding carbon emissions from the corresponding rows and columns of the carbon emissions matrix, and calculate the resource load matrix by subtracting the corresponding load from the corresponding server. Module 3: Based on the carbon emission matrix and the zero value of the computing resource load matrix, assign the execution device for the vision task. If the assigned zero value is equal to the total number of vision tasks, then the process ends; otherwise, Module 2 is called again.

7. The carbon emission assessment device for edge-cloud visual collaborative architecture as described in claim 5 or 6, characterized in that, The model building module includes: The cloud server contains computing devices for training, learning, and data computation, a cooling system, and basic equipment for cloud operation; the energy consumption of all devices operating throughout the cloud center. Divide by the energy consumption of IT equipment operation The energy efficiency of the cloud server was obtained. ; The cooling coefficient of this cloud server is ,in , and For real numbers, The temperature set for the cooling equipment, and the power consumption of the cloud server are... Carbon emissions and the carbon intensity of electricity consumption ;in and These represent the collection of renewable energy sources and the collection of regions, respectively. Indicates the type of renewable energy In the region Power consumption; For renewable energy types The carbon emission rate; the computing resource load rate of the cloud server: ,in and These are the current load and maximum load of the cloud server, respectively; the carbon emissions generated by the operation of this cloud server are: ; Based on the carbon emissions from the refrigerant leak of the cloud server and the quality of the refrigerant used. Global warming potential (GWP), refrigerant leakage rate and the overall uptime of the equipment The carbon emissions generated by the use of refrigerants ; The carbon emission model for this cloud device is as follows: The edge server includes IT equipment for edge computing and basic system operating equipment; this edge-cloud vision collaborative IoT includes... Each edge server; obtain the power energy utilization efficiency of each edge server. ; Obtain the computing power consumption of the edge server's own computing device. and running time The computing power consumption of edge servers With full load power consumption during operation Power consumption during idle time Current required workload and maximum workload Relevant; the load rate of this edge server is: The power consumption is: The edge server's working and idle times are divided into two segments. and Because different sites provide different hybrid energy sources to the edge servers, the carbon intensity of electricity at each edge is... Difference, i.e., edge end The overall carbon emission calculation model is as follows: The carbon emission model for this equipment is as follows: Traditional visual terminals use visual perception to compress captured videos, upload and store them, and then analyze, identify, and process them at the edge and cloud. Intelligent visual terminals, on the other hand, possess visual feature extraction capabilities and store the compressed video stream locally, retrieving it on demand based on edge and cloud requirements. They also support the migration and compression of video encoding and feature analysis models between the device, edge, and cloud. This enables the construction of visual terminals. Equipment categories If it is a traditional visual terminal, then Otherwise ; This traditional visual terminal is based on its power consumption and running time and the area where it is located Types of energy that use electricity and electrical carbon strength Based on the current required workload The carbon emission model of this traditional visual terminal ; This intelligent vision terminal also has data analysis capabilities, but its power consumption efficiency is lower than that of edge devices. Based on the power consumption of this intelligent vision terminal Running time Full load power consumption during operation Power consumption during idle time Current required workload and maximum workload The load rate of this intelligent vision terminal is: The power consumption of this intelligent vision terminal is... The intelligent vision terminal Work and free time are divided into two segments. and The carbon intensity of this intelligent vision terminal With the region and types of energy that generate electricity The carbon emission model for this intelligent vision terminal is as follows: ; The carbon emission model for this terminal device is as follows: Obtain the network transmission carbon emissions and communication transmission efficiency under edge-cloud collaboration in this edge-cloud visual collaborative IoT. The amount of data transmitted Energy consumption during transmission and electrical carbon strength ; Obtain the energy consumption per unit bit of data in this edge-cloud visual collaborative IoT. ; The amount of data transmitted by edge-cloud vision-coordinated IoT With the resolution of video frames Related, that is ,in It is a fixed value; and the energy consumption of network transmission is the same as the energy consumption per unit bit of data. Related to the resolution of video frames. At the node and nodes The carbon emission transfer model between them is as follows: Carbon emissions from network transmission and energy consumption per unit bit of data Electrocarbon strength Video frame resolution Frame sampling rate and communication transmission efficiency Relatedly, the carbon emission model for network transmission is as follows: The uninstallation selection strategy matrix In this context, x represents a 0 or 1 decision task executed locally on the intelligent vision terminal, on an edge server, or on a cloud server; m is the ID of the terminal device; and V is the set of all vision tasks. This indicates that the intelligent vision terminal will uninstall the software locally. This indicates that the data is being unloaded from the edge server. This indicates unloading on the cloud server; for edge servers... The load factor of edge computing resources is: For cloud servers, the load factor of computing resources is: Overall carbon emission model: The model consists of four parts: the first part is the carbon emission model for local offloading; the second part is the carbon emission model for edge offloading; the third part is the carbon emission model for cloud offloading; and the fourth part is the carbon emission model for network transmission from the terminal to the edge and the cloud.

8. An electronic device, characterized in that, The carbon emission assessment device includes the edge-cloud visual collaborative architecture as described in any one of claims 5-7, wherein the electronic device is connected to an information display device, which is used to display the assessment results using user-set display parameters, attributes, or through an artificial intelligence model.

9. A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the carbon emission assessment method for an edge-cloud visual collaborative architecture as described in any one of claims 1-4.

10. A computer program product, comprising a computer program, characterized in that, When executed by a processor, the computer program implements the steps of the carbon emission assessment method for the edge-cloud visual collaborative architecture as described in any one of claims 1-4.

Images