Multi-level cooperative regulation and control method and system based on dynamic load of data center

Abstract

The invention provides a multi-level cooperative regulation and control method and system based on a data center dynamic load, and relates to the technical field of IT infrastructure low-carbon. In the method provided by the invention, non-real-time tasks matched with power consumption are obtained according to the generated power of renewable energy sources accessed by a user layer data center; the matched non-real-time task is operated by using the electric energy generated by the renewable energy source, so that the low-carbon requirement in the operation process of the user layer data center is met; meanwhile, unmatched non-real-time tasks are suspended to a waiting queue, and the non-real-time tasks reaching a waiting time threshold in the waiting queue are migrated to a user layer data center of a regional level or an overall level by utilizing a task space scale; the electric energy of the renewable energy accessed by the user layer data center of the regional level or the overall level is used for power supply and operation, so that the electric energy generated by the renewable energy is timely consumed, and the electric energy generated by the renewable energy is prevented from being wasted.

Description

Technical Field

[0001] The present invention relates to the technical field of low-carbonization of IT (Internet Technology) infrastructure, and in particular to a multi-level collaborative control method and system based on dynamic load of a data center. Background Art

[0002] With the widespread application of information and digital technologies, the demand for high-capacity data storage, high-performance computing, and high-speed transmission in different industries is growing. As an intelligent infrastructure that undertakes tasks such as massive data storage and computing, the scale and number of data centers are increasing rapidly around the world. However, the rapid growth in the scale and number of data centers has brought about significant problems of high energy consumption and high carbon emissions. Increasing the proportion of renewable energy (such as wind / light) in data centers is an important means to reduce carbon emissions from data centers.

[0003] The existing method for reducing carbon emissions from data centers is to establish a tracking and matching algorithm and system for computing loads on renewable energy power for a single data center, thereby achieving the goal of improving the utilization rate of renewable energy power.

[0004] However, the adjustable computing power and power capacity of a single data center are very limited. At the same time, due to the strong volatility of local renewable energy power generation, data centers still need to use a large amount of traditional energy and electricity to ensure task processing performance, resulting in high carbon emissions levels. When renewable energy power generation provides a higher power generation capacity, it is difficult to fully absorb the generated electricity through the local data center's own computing load, which in turn leads to the waste of electricity generated by renewable energy power generation. Summary of the invention

[0005] In order to solve the problems in the prior art that data centers use traditional energy to generate power for power processing tasks, resulting in high carbon emissions, and that it is difficult to fully consume the electricity generated by renewable energy using their own computing load when renewable energy power generation provides a higher power, the present invention proposes a multi-level collaborative control method and system based on the dynamic load of the data center.

[0006] In a multi-level collaborative control method based on the dynamic load of a data center proposed in the present invention, non-real-time tasks with matching power consumption are obtained according to the power generation power of renewable energy accessed by the user-level data center, and the matching non-real-time tasks are operated using the electric energy generated by the renewable energy, thereby meeting the low-carbon requirements during the operation of the user-level data center; at the same time, the non-matching non-real-time tasks are suspended to a waiting queue, and the non-real-time tasks in the waiting queue that reach the waiting time threshold are migrated to the user-level data center at the regional level or the overall level using task space scale migration, and are powered and operated by the electric energy of renewable energy accessed by the user-level data center at the regional level or the overall level, thereby ensuring the timely consumption of the electric energy generated by the renewable energy and avoiding the waste of the electric energy generated by the renewable energy. At the same time, the electric power generated by the renewable energy is used to process the tasks, which can reduce the carbon emission level caused by the processing tasks.

[0007] On the one hand, the present invention proposes a multi-level collaborative control method based on dynamic load of a data center, comprising:

[0008] Based on each user-level data center, obtain the power generation of the connected renewable energy and the power consumption of each connected non-real-time task;

[0009] Based on the power generation power of the renewable energy source and the power consumption of each non-real-time task, determining a non-real-time task matching the power generation power in each non-real-time task;

[0010] Utilizing the renewable energy to power matching non-real-time tasks;

[0011] The non-real-time tasks that do not match the generated power are suspended in a waiting queue, and the waiting time is obtained; if there are non-real-time tasks whose waiting time exceeds the waiting time threshold, the task space scale migration strategy is used to migrate the non-real-time tasks whose waiting time exceeds the waiting time threshold as migration tasks to a user-layer data center at the regional level or the overall level.

[0012] Optionally, for non-real-time tasks whose waiting time exceeds the waiting time threshold, the task space scale migration strategy is used to migrate the non-real-time tasks whose waiting time exceeds the waiting time threshold as migration tasks to a user-layer data center at the regional level or the overall level, including:

[0013] Non-real-time tasks whose waiting time exceeds the waiting time threshold are considered as migration tasks;

[0014] Determining computing resources required to run the migration task according to the power consumption of the migration task;

[0015] According to the computing resources, using a traversal search algorithm, a plurality of candidate user-layer data centers that meet the requirements for running the migration task are obtained in the current area or other areas;

[0016] Determine, according to the migration path with each candidate user layer data center, the carbon emissions required to complete the task space scale migration with each candidate user layer data center;

[0017] In each carbon emission amount, determine the candidate user layer data center with the minimum carbon emission amount as the target user layer data center;

[0018] Migrate the migration task to the target user layer data center.

[0019] Optionally, determining the carbon emissions required to complete the task space scale migration with each candidate user layer data center according to the migration path with each candidate user layer data center respectively includes:

[0020] According to the total amount of data of the migration task, the available transmission rate of the optical network in the migration path of each alternative user layer data center, the characteristics of the optical communication equipment and the number of routes in the transmission path are comprehensively considered to determine the carbon emissions required to complete the spatial scale migration of the task with each alternative user layer data center.

[0021] Optionally, the acquiring, based on the computing resources and using a traversal search algorithm, a plurality of candidate user-layer data centers that meet the requirements for running the migration task in the current area or other areas includes:

[0022] Based on the computing resources, a traversal search algorithm is used to determine whether there is a user-level data center in the current area that meets the requirements for running the migration task:

[0023] If yes, taking the user layer data center in the current area that meets the requirements for running the migration task as a candidate user layer data center;

[0024] If not, a user layer data center that meets the migration task is obtained in other regions, and the user layer data center that meets the migration task in other regions is used as a candidate user layer data center.

[0025] Optionally, after suspending the non-real-time tasks that do not match the generated power to a waiting queue and obtaining the waiting time, the method further includes:

[0026] If there is at least one non-real-time task whose waiting time exceeds the waiting time threshold and has not been migrated to the user-layer data center at the regional level or the overall level, the task space scale migration of the at least one non-real-time task is stopped, and the at least one non-real-time task is converted into a real-time task, the converted real-time task is powered by the power grid in the current area, and the converted real-time task is run.

[0027] Optionally, after using the renewable energy to power the matching non-real-time tasks, the method further includes:

[0028] Determine whether there is any suspended non-real-time task in the waiting queue;

[0029] If it does not exist, then receive and run the non-real-time tasks migrated from other user-layer data centers through the task space scale, clear the accumulated waiting time of the non-real-time tasks migrated from other user-layer data centers, and shorten the waiting time threshold of the non-real-time tasks migrated from other user-layer data centers.

[0030] Optionally, after the renewable energy is used to power the matching non-real-time tasks, the method further includes:

[0031] For the connected real-time task, obtaining the power consumption of running the connected real-time task;

[0032] Determine whether there is any suspended non-real-time task in the waiting queue;

[0033] If it is determined that there is no such a power source, the remaining power generated by the connected renewable energy source is obtained;

[0034] If the surplus power generated by the renewable energy matches the power consumption of the connected real-time task, the renewable energy is used to power the independent server, and the independent server runs the connected real-time task.

[0035] On the other hand, the present invention also proposes a multi-level coordinated control system based on dynamic load of data center, the system includes multiple user layer data centers, multiple regional layer data centers and overall layer data center; each of the regional layer data centers is communicatively connected with multiple user layer data centers, and each of the regional layer data centers is connected with the overall layer data center;

[0036] Each user layer data center is used to execute a multi-level coordinated control method of a central dynamic load as described in the above technical solution;

[0037] The regional data center is used to transfer the remaining computing resources of other user data centers in the region and the surplus power of the connected renewable energy to the user data center to be used for task space scale migration;

[0038] The overall layer data center is used to transfer the remaining computing resources of other user layer data centers in other areas and the surplus power generation power of the connected renewable energy sources to the user layer data center to be migrated to the task space scale through the corresponding regional layer data center.

[0039] Optionally, the user layer data center includes:

[0040] The power data acquisition module is used to obtain the power side parameters of the data center connected to the current user layer;

[0041] A computing resource acquisition module is used to obtain computing resources of the current user layer data center;

[0042] A data center information collection module, used to communicate with a target user layer data center receiving the non-real-time task when performing task space scale migration on the non-real-time task;

[0043] The task migration and tracking module is connected to the regional layer data center or the overall layer data center for communication, and is used to migrate the task space scale of non-real-time tasks and track the running status of the non-real-time tasks.

[0044] Optionally, the power side parameters received by the power data acquisition module include:

[0045] Green electricity / non-green electricity supply power, cumulative power supply, consumed power, cumulative consumed electricity, green electricity supply ratio, wind / solar / hydro ratio in green electricity, non-green electricity coal / gas / nuclear ratio, green electricity utilization rate and green electricity in energy storage equipment.

[0046] Optionally, the computing resources received by the computing resource acquisition module include:

[0047] Task type, number of tasks, task status, task data volume, task computing resource usage, energy consumption per unit computing resource, cumulative running time, cumulative waiting time, waiting time threshold, task power consumption, cumulative power consumption, proportion of green power in task power consumption, data center path for task space migration, total computing resources and idle computing resources in the data center.

[0048] Optionally, the user layer data center further includes:

[0049] The parameter statistics module is used to count the carbon emission information of the current user layer data center, and the carbon emission information includes the carbon emission of the current user layer data center, the carbon emission generated by the task space scale migration and / or the green electricity surplus power.

[0050] On the other hand, the present invention also proposes a multi-level coordinated control device based on the dynamic load of a data center, comprising:

[0051] The task power consumption acquisition module is used to acquire the power generation power of the connected renewable energy and the power consumption of each connected non-real-time task;

[0052] A task matching module, configured to determine, in each of the non-real-time tasks, a non-real-time task that matches the generated power based on the generated power of the renewable energy and the power consumption of each non-real-time task;

[0053] A task processing module, used to utilize the renewable energy to perform power supply processing for matching non-real-time tasks;

[0054] A task waiting module is used to suspend non-real-time tasks that do not match the generated power to a waiting queue and obtain the waiting time;

[0055] The task migration module is used to migrate the non-real-time tasks whose waiting time exceeds the waiting time threshold to the user-level data center at the regional level or the overall level as migration tasks by using the task space scale migration strategy.

[0056] Optionally, the task migration module includes:

[0057] A migration task determination unit, used for determining a non-real-time task whose waiting time exceeds a waiting time threshold as a migration task;

[0058] A computing resource and power consumption determination unit, configured to determine the computing resources required to run the migration task according to the power consumption of the migration task;

[0059] A target data center search unit, configured to obtain, based on the computing resources and using a traversal search algorithm, a plurality of candidate user-layer data centers that meet the requirements for running the migration task in the current area or other areas;

[0060] A carbon emission determination unit, used to determine the carbon emission required to complete the task space scale migration with each candidate user layer data center according to the migration path with each candidate user layer data center;

[0061] A target data center determination unit is used to determine, in each carbon emission amount, an alternative user layer data center with a minimum carbon emission value as the target user layer data center;

[0062] The migration implementation unit is used to migrate the migration task to the target user layer data center.

[0063] Optionally, the carbon emission determination unit is specifically used to:

[0064] According to the total amount of data of the migration task, the available transmission rate of the optical network in the migration path of each alternative user layer data center, the characteristics of the optical communication equipment and the number of routes in the transmission path are comprehensively considered to determine the carbon emissions required to complete the spatial scale migration of the task with each alternative user layer data center.

[0065] Optionally, the target data center search unit is specifically used to:

[0066] Based on the computing resources, a traversal search algorithm is used to determine whether there is a user-level data center in the current area that meets the requirements for running the migration task:

[0067] If yes, taking the user layer data center in the current area that meets the requirements for running the migration task as a candidate user layer data center;

[0068] If not, a user layer data center that meets the migration task is obtained in other regions, and the user layer data center that meets the migration task in other regions is used as a candidate user layer data center.

[0069] Optionally, a task conversion module is also included, which is used to:

[0070] When there is at least one non-real-time task whose waiting time exceeds the waiting time threshold and has not been migrated to the user-layer data center at the regional level or the overall level, the task space scale migration of the at least one non-real-time task is stopped, and the at least one non-real-time task is converted into a real-time task, the converted real-time task is powered by the power grid in the current area, and the converted real-time task is run.

[0071] Optionally, a task receiving module is also included, which is used to:

[0072] Determine whether there is any suspended non-real-time task in the waiting queue;

[0073] If it does not exist, then receive and run the non-real-time tasks migrated from other user-layer data centers through the task space scale, clear the accumulated waiting time of the non-real-time tasks migrated from other user-layer data centers, and shorten the waiting time threshold of the non-real-time tasks migrated from other user-layer data centers.

[0074] Optionally, a power supply switching module is also included, which is used to:

[0075] For the connected real-time task, obtaining the power consumption of running the connected real-time task;

[0076] Determine whether there is any suspended non-real-time task in the waiting queue;

[0077] If it is determined that there is no such a power source, the remaining power generated by the connected renewable energy source is obtained;

[0078] If the surplus power generated by the renewable energy matches the power consumption of the connected real-time task, the renewable energy is used to power the independent server, and the independent server runs the connected real-time task.

[0079] In another aspect, the present invention further provides an electronic device, comprising: at least one processor and a memory; the memory and the processor are connected via a bus;

[0080] The memory is used to store one or more programs;

[0081] When the one or more programs are executed by the at least one processor, a multi-level collaborative control method based on the dynamic load of the data center as described in the above technical solution is implemented.

[0082] On the other hand, the present invention further proposes a readable storage medium having an execution program stored thereon, and when the execution program is executed, the multi-level collaborative control method based on the dynamic load of the data center as described in the above technical solution is implemented.

[0083] Compared with the prior art, the present invention has the following beneficial effects:

[0084] The present invention proposes a multi-level collaborative control method based on the dynamic load of a data center. The method obtains non-real-time tasks with power consumption matching according to the power generation power of renewable energy connected to the user-level data center, and uses the electric energy generated by renewable energy to run the matching non-real-time tasks, thereby meeting the low-carbon requirements during the operation of the user-level data center; at the same time, the non-matching non-real-time tasks are suspended to a waiting queue, and the non-real-time tasks that reach the waiting time threshold in the waiting queue are migrated to the user-level data center at the regional level or the overall level by using the task space scale migration, and are powered and operated by the electric energy of renewable energy connected to the user-level data center at the regional level or the overall level, thereby ensuring the full absorption of renewable energy power generation and avoiding the waste of electric energy generated by renewable energy. At the same time, the electric power generated by renewable energy is used to process tasks, which can reduce the carbon emission level caused by processing tasks.

[0085] The present invention proposes a multi-level coordinated control system based on the dynamic load of a data center. The regional layer data center is used to obtain whether other user layer data centers in the region meet the requirements of running non-real-time tasks with a waiting time exceeding the waiting time threshold, and other user layer data centers in the regional layer are used to run non-real-time tasks with a waiting time exceeding the waiting time threshold; at the same time, when other user layer data centers in the current region do not meet the requirements of running non-real-time tasks with a waiting time exceeding the waiting time threshold, the overall layer data center is used to migrate to other user layer data centers through regional layer data centers in other regions, thereby realizing the operation of non-real-time tasks with a waiting time exceeding the waiting time threshold using renewable energy power supply accessed by other user layer data centers, thereby reducing the probability of wasting power generated by renewable energy. BRIEF DESCRIPTION OF THE DRAWINGS

[0086] Figure 1 It is a flow chart of the multi-level collaborative control method based on data center dynamic load of the present invention;

[0087] Figure 2 It is a flow chart of the task space scale migration strategy of the multi-level collaborative control method based on the dynamic load of the data center of the present invention;

[0088] Figure 3 It is a control schematic diagram of the multi-level coordinated control method based on the dynamic load of the data center of the present invention;

[0089] Figure 4 It is a structural schematic diagram of a user layer data center in a multi-level coordinated control system based on dynamic load of a data center of the present invention;

[0090] Figure 5 It is a schematic structural diagram of the electronic device of the present invention. DETAILED DESCRIPTION

[0091] Embodiment 1:

[0092] Reference Figure 1 The present invention proposes a multi-level collaborative control method based on the dynamic load of a data center, comprising:

[0093] S1, based on each user layer data center, obtain the power generation power of the connected renewable energy and the power consumption of each connected non-real-time task;

[0094] S2, based on the power generation power of the renewable energy and the power consumption of each non-real-time task, determining a non-real-time task matching the power generation power in each non-real-time task;

[0095] S3, utilizing the renewable energy to power matching non-real-time tasks;

[0096] S4, suspending the non-real-time tasks that do not match the generated power to a waiting queue, and obtaining the waiting time;

[0097] S5, if there are non-real-time tasks whose waiting time exceeds the waiting time threshold, the task space scale migration strategy is used to migrate the non-real-time tasks whose waiting time exceeds the waiting time threshold as migration tasks to the user-layer data center at the regional level or the overall level; if there are no non-real-time tasks whose waiting time exceeds the waiting time threshold, they are kept in the waiting queue and wait for processing.

[0098] In order to support the country's strategic goal of green, low-carbon and sustainable development, the current development focus is on: optimizing the energy structure of data centers and improving the utilization rate of clean energy and renewable energy. The key challenge is the coordinated regulation method of computing power and electricity. Since renewable energy generation has strong temporal and spatial fluctuations, and the computing load of different data centers distributed across the country (i.e. user task volume, which is proportional to energy consumption) is also highly random in terms of temporal and spatial scales, it is necessary to match the computing load of data centers and the output of renewable energy generation in terms of both time and space, so as to take into account the high task processing performance and low carbon emissions of data centers.

[0099] For example, in terms of computing power-electricity matching on a spatial scale, the country has proposed the "Eastern Data Western Computing" strategy, which will migrate non-real-time data processing tasks that allow for higher latency to data center clusters in the west, which are rich in renewable energy. This puts forward higher requirements for the current computing power-electricity coordinated control method, that is, to form a large-scale computing power-electricity matching capability covering multiple levels of user-region-country, in order to effectively support the country's "Eastern Data Western Computing" strategy.

[0100] Currently available low-carbon technology solutions for data centers include: (1) establishing a tracking and matching algorithm and system for computing loads with respect to renewable energy power, thereby achieving coordinated regulation of computing power and power in a single data center and improving the utilization rate of renewable energy in the region; (2) configuring a single data center with advanced energy-saving equipment such as low-power chips, Panama power supplies, liquid cooling, and hydrogen backup power supplies, thereby reducing the energy consumption and carbon emissions of the user-level data center itself; (3) recovering the heat generated by the data center during the operation of tasks, and using the recovered heat for building heating or industrial and agricultural heating; (4) merging existing small-scale, dispersed user-level data centers, and using the merged user-level data centers to run tasks, thereby improving local energy efficiency.

[0101] The computing power and power capacity that can be regulated by a single data center are very limited, and renewable energy generation has strong volatility. In order to ensure the task processing efficiency of a single data center, a large amount of electricity generated by traditional energy in the current area is used to operate and receive computing loads, resulting in a high level of carbon emissions from the data center; at the same time, when the renewable energy generation in the current area provides a higher power generation, it is difficult for a single data center to absorb the power provided by renewable energy generation in a timely manner, and the excess electricity generated by renewable energy is wasted. As for the current three technologies of replacing energy-saving equipment, recycling waste heat or integrating data centers, it is necessary to purchase, install or replace a large number of new equipment such as low-power server chips, high-efficiency coolers, waste heat recycling heat exchangers, and even involve the dismantling, migration and reorganization of existing data centers, which will take a high time and economic cost, and is not conducive to the large-scale promotion of data centers in existing technologies.

[0102] Combining current technology, it is known that renewable energy generation has strong temporal and spatial fluctuations, and the user task volume of different data centers (which is proportional to energy consumption) also has strong randomness in terms of temporal and spatial scales. Therefore, in order to coordinate the computing power and electricity of the data center, the present invention matches the computing load of the data center and the power generation power of renewable energy in terms of both time and space, thereby taking into account the high task processing performance and low carbon emission requirements of the data center.

[0103] In the multi-level collaborative control method based on the dynamic load of the data center provided by the present invention, the multiple levels include multiple user-level data centers, multiple regional-level data centers and an overall-level data center, wherein the user-level data is implemented as a server cluster distributed in various regions, which is used to use the power provided by the region to run the received data processing tasks, and the regional-level data center is used to manage multiple user-level data centers to achieve data transmission of spatial scale migration of tasks at the regional level. For example, the regional-level data center can be a provincial-level data center; the overall-level data center is used to manage multiple regional-level data centers to achieve data transmission of spatial scale migration of tasks at the overall level. For example, the overall-level data center can be a national-level data center.

[0104] Since real-time tasks are non-interruptible, that is, real-time tasks cannot be added to the waiting queue and cannot be migrated through the task space scale migration strategy, the ability to utilize the electricity generated by renewable energy is poor. Therefore, it is considered that the non-real-time tasks of the user-level data center are powered and operated by the electricity generated by renewable energy, so as to achieve the purpose of reducing the carbon emission level of the user-level data center.

[0105] S1 can be specifically implemented as follows: the user-level data center uses electricity generated by renewable energy to run non-real-time tasks, which can greatly reduce the carbon emission level of the user-level data center. The electricity generated by renewable energy used by the user-level data center generally comes from the electricity produced by wind / photovoltaic / hydroelectric generators built by the user-level data center or purchased from the region. The electricity generated by renewable energy can be connected to the user-level data center through energy storage equipment for power supply, or it can be directly connected to the user-level data center in conjunction with a renewable energy power generation power prediction model to reduce the impact of the strong volatility of renewable energy generation on the operation of non-real-time tasks in the user-level data center.

[0106] Exemplarily, the power consumption of each non-real-time task in S2 is proportional to the computing resources occupied by the task, where the computing resources are the number of CPU (Central Processing Unit) / GPU (Graphics Processing Unit) cores of the processor cluster occupied by the task.

[0107] Depending on the non-real-time tasks received by a single user-level data center, the power consumption required to run each non-real-time task is also different. At the same time, the power generation power of renewable energy has a time waveform.

[0108] S3 may include: when the power generation power of the connected renewable energy source has sufficient surplus, corresponding to the power consumption required to run the non-real-time tasks connected to the user layer data center, the connected non-real-time tasks are powered and run using the power generation power of the renewable energy source. The total power consumption of these non-real-time tasks matches the surplus power of the electric energy generated by the connected renewable energy source to the highest degree.

[0109] S4 can be implemented as follows: when the total power of the electricity generated by the connected renewable energy is insufficient, the user-layer data center will suspend the non-real-time tasks that have not matched the power consumption or have not been completed in the waiting queue in order from small to large according to the power required to run, until the total power consumption of the running tasks matches the total power of the electricity generated by the renewable energy, thereby realizing the task migration of non-real-time tasks and time scale migration, and reducing the impact of the volatility of the electricity generated by renewable energy on the processing of non-real-time tasks.

[0110] The waiting queue is implemented by sorting the non-real-time tasks that are not running, and at the same time, the waiting time of each non-real-time task suspended in the waiting queue is accumulated. The waiting time of each non-real-time task is calculated from the time the corresponding non-real-time task is allocated to the user-level data center.

[0111] The waiting time threshold is predefined according to the importance and urgency of each non-real-time task. The waiting time thresholds of different non-real-time tasks can be set to the same value or can vary with their respective importance and urgency.

[0112] The task space scale migration strategy in S5 is a method for spatial scale migration of computing loads. It is used to migrate non-real-time tasks whose waiting time exceeds the waiting time threshold in a certain user-level data center to other user-level data centers with access to renewable energy for processing, thereby meeting the low-carbon emission requirements of running tasks.

[0113] S5 can be implemented as follows: for non-real-time tasks that cannot be completed by using renewable energy power generation and power supply in the initial user-layer data center, when the cumulative waiting time reaches the threshold, first carry out regional-level task space scale migration, where the region refers to other user-layer data centers within the province; if there is no user-layer data center that meets the migration task at the regional level, expand the search set to user-layer data centers outside the province, and carry out overall-level task space scale migration, and the overall-level user-layer data center can be implemented as a national user-layer data center.

[0114] By obtaining non-real-time tasks with matching power consumption according to the power generation power of renewable energy connected to the user-layer data center, and using the electric energy generated by renewable energy to run the matching non-real-time tasks, the low-carbon requirements of the user-layer data center during operation are met; at the same time, the non-matching non-real-time tasks are suspended to the waiting queue, and the non-real-time tasks in the waiting queue that reach the waiting time threshold are migrated to the user-layer data center at the regional level or the overall level by using the task space scale migration, and are powered and operated by the electric energy of renewable energy connected to the user-layer data center at the regional level or the overall level, thereby ensuring the full absorption of renewable energy power generation and avoiding the waste of electric energy generated by renewable energy. At the same time, the electric energy generated by renewable energy is used to process tasks, which can reduce the carbon emission level caused by processing tasks.

[0115] By reducing the carbon emissions from the daily operation of data centers, while meeting the huge computing power needs of society's digital transformation, it helps achieve the country's "dual carbon" strategic goals; by matching the computing load of data centers and the output of renewable energy power generation in both time and space, thereby taking into account both the high task processing performance and low carbon emissions of data centers, the matching and synergistic complementarity of data center computing power and renewable energy green electricity is achieved, taking into account both the task processing needs and low carbon emissions of data centers.

[0116] like Figure 2As shown, in some embodiments, in order to ensure that the user-level data center is at a lower carbon emission level during the migration of non-real-time tasks, S5 proposed by the present invention includes:

[0117] S51, taking a non-real-time task whose waiting time exceeds a waiting time threshold as a migration task;

[0118] S52, determining computing resources required to run the migration task according to the power consumption of the migration task;

[0119] S53, obtaining, based on the computing resources and using a traversal search algorithm, a plurality of candidate user-layer data centers that are capable of running the migration task in the current area or other areas;

[0120] S54, determining the carbon emissions required to complete the task space scale migration with each candidate user layer data center according to the migration path with each candidate user layer data center;

[0121] S55, determining, in each carbon emission amount, a candidate user layer data center with a minimum carbon emission value as a target user layer data center;

[0122] S56, migrating the migration task to the target user layer data center.

[0123] The power consumption of the migration task in S52 is implemented as the power generation power required to complete the operation of the migration task, and the computing resources of the migration task are the number of CPU / GPU cores of the processor cluster occupied during the operation of the migration task.

[0124] The constraints of the migration task are:

[0125] The power consumption of the migration task is ≤ the remaining green power of the user-level data center.

[0126] The computing resources required by the migration task are ≤ the remaining computing resources of the user-level data center.

[0127] The constraints of the migration task are used to filter user-layer data centers that meet the requirements for running the migration task in the current region or other regions, and use the filtered user-layer data centers as candidate user-layer data centers to facilitate the acquisition of the carbon emissions required for task space migration between each candidate user-layer data center.

[0128] In S53, the search set is determined to be all available user-layer data centers in the current area, and a traversal search algorithm is used to first search in the user-layer data centers at the regional level. If there is no user-layer data center that meets the conditions at the regional level, the search set is expanded to user-layer data centers in other regions to carry out task space migration at the overall level.

[0129] Therefore, S53 includes:

[0130] Based on the computing resources, a traversal search algorithm is used to determine whether there is a user-level data center in the current area that meets the requirements for running the migration task:

[0131] If yes, taking the user layer data center in the current area that meets the requirements for running the migration task as a candidate user layer data center;

[0132] If not, a user layer data center that meets the migration task is obtained in other regions, and the user layer data center that meets the migration task in other regions is used as a candidate user layer data center.

[0133] By determining the search set as all available user-layer data centers in the current area, a traversal search algorithm is used to first search in the user-layer data centers at the regional level. If there is no user-layer data center that meets the constraints at the regional level, the search set is expanded to user-layer data centers in other regions, and user-layer data centers that meet the constraints are searched in other regions.

[0134] In some embodiments, S54 includes:

[0135] According to the total amount of data of the migration task, the available transmission rate of the optical network in the migration path of each alternative user layer data center, the characteristics of the optical communication equipment and the number of routes in the transmission path are comprehensively considered to determine the carbon emissions required to complete the spatial scale migration of the task with each alternative user layer data center.

[0136] The available transmission rate of the optical network in the migration task of the alternative data center corresponds to the transmission speed required to complete the transmission migration task. The characteristics of optical communication equipment are the correspondence between transmission speed and transmission distance, and the routing in the transmission path is the equipment to ensure the optical signal strength of the migration task during the migration process. The longer the migration path, the more routing is required. Similarly, the corresponding carbon emissions are also higher.

[0137] By migrating non-real-time tasks, we can improve the matching and synergistic complementarity between data center computing load and renewable energy power, increase the proportion of renewable energy supply in data centers, optimize the energy consumption structure of data centers, and reduce carbon emissions from data centers with less time and economic costs; at the same time, it can also promote the consumption of renewable energy and improve the utilization rate of renewable energy green electricity.

[0138] In some embodiments, after suspending the non-real-time task that does not match the generated power to a waiting queue and obtaining the waiting time, the method further includes:

[0139] If there is at least one non-real-time task whose waiting time exceeds the waiting time threshold and has not been migrated to the user-layer data center at the regional level or the overall level, the task space scale migration of the at least one non-real-time task is stopped, and the at least one non-real-time task is converted into a real-time task, the converted real-time task is powered by the power grid in the current area, and the converted real-time task is run.

[0140] When it is impossible to migrate non-real-time tasks using task space scale migration, that is, it is impossible to power the non-real-time tasks by using the electricity generated by renewable energy in other user-level data centers, and the waiting time of the non-real-time tasks reaches the threshold, it indicates that the corresponding non-real-time tasks need to be processed urgently. In the above situation, the user data center stops the task space scale migration of the corresponding non-real-time tasks, and at the same time changes the task type of the non-real-time tasks to real-time type, and the data center is powered by the external power grid in the area until completion.

[0141] In some embodiments, after using the renewable energy to power the matching non-real-time tasks, the method further includes:

[0142] Determine whether there is any suspended non-real-time task in the waiting queue;

[0143] If it does not exist, then receive and run the non-real-time tasks migrated by other user-layer data centers through the task space scale, clear the accumulated waiting time of the non-real-time tasks migrated by other user-layer data centers, and shorten the waiting time threshold of the non-real-time tasks migrated by other user-layer data centers;

[0144] When the corresponding non-real-time task is suspended in the waiting queue again, the time scale migration and the above-mentioned space scale migration method are cyclically executed according to the shortened waiting time threshold until the non-real-time task is completed.

[0145] When there are no pending non-real-time tasks in the waiting queue of any user-layer data center, it indicates that the power generation power of the renewable energy connected to the corresponding user-layer data center may have a surplus, that is, it meets the constraints of running non-real-time tasks migrated from other user-layer data centers. Therefore, by utilizing the electric energy of the connected renewable energy to run the received non-real-time tasks migrated from other user-layer data centers, it is convenient to timely absorb the power generation power provided by renewable energy generation, which can reduce the probability of wasting excess electric energy generated by renewable energy.

[0146] Shortening the waiting time threshold of non-real-time tasks of other user-layer data center migration is used to prevent the cumulative waiting time of the corresponding non-real-time tasks from being too long. The waiting time threshold of non-real-time tasks can be shortened to half, or the waiting time threshold of non-real-time tasks can be adjusted according to the importance and urgency of the task. The reduction amount of the cumulative waiting time and the threshold reduction amount are adjusted. In this embodiment, the specific implementation is to shorten the waiting time threshold to half.

[0147] Since real-time tasks are non-interruptible, they cannot be added to the waiting queue and cannot migrate the task space scale. Therefore, real-time tasks have poor ability to utilize renewable energy power generation.

[0148] In some embodiments, in order to improve the convenience of using the electric energy generated by renewable energy for real-time tasks, after using the renewable energy to power the matching non-real-time tasks, the method further includes:

[0149] For the connected real-time task, obtaining the power consumption of running the connected real-time task;

[0150] Determine whether there is any suspended non-real-time task in the waiting queue;

[0151] If it is determined that there is no such a power source, the remaining power generated by the connected renewable energy source is obtained;

[0152] If the surplus power generated by the renewable energy matches the power consumption of the connected real-time task, the renewable energy is used to power the independent server, and the independent server runs the connected real-time task.

[0153] Based on the existing server equipment in the user-level data center, a server cluster is divided to independently process real-time tasks. The power supply system of the cluster is independent of the power supply system of other non-real-time task processing clusters. When all other non-real-time tasks in the current user-level data center have been run, and there is still surplus power generated by renewable energy, the surplus power of renewable energy will be used to run the corresponding real-time tasks, thereby ensuring that the power generated by renewable energy is fully absorbed and reducing the carbon emissions of the user-level data center.

[0154] Embodiment 2:

[0155] Reference Figure 3 The present invention based on the same inventive concept also proposes a multi-level coordinated control system based on the dynamic load of the data center, the system includes multiple user-level data centers, multiple regional-level data centers and an overall-level data center; each of the regional-level data centers is communicatively connected to multiple user-level data centers, and each of the regional-level data centers is connected to the overall-level data center;

[0156] Each user layer data center is used to execute a multi-level coordinated control method of a central dynamic load as described in the above technical solution;

[0157] The regional data center is used to transfer the remaining computing resources of other user data centers in the region and the surplus power of the connected renewable energy to the user data center to be used for task space scale migration;

[0158] The overall layer data center is used to transfer the remaining computing resources of other user layer data centers in other areas and the surplus power generation power of the connected renewable energy sources to the user layer data center to be migrated to the task space scale through the corresponding regional layer data center.

[0159] On the basis of considering the time migration of computing load within a single data center, comprehensively considering factors such as carbon emissions from task migration and transmission in optical communication networks, surplus power of renewable energy and green electricity within and outside the province, data processing tasks are gradually migrated to data centers with abundant green electricity within / outside the province, thereby achieving computing power-power matching over a large spatial scale. This can provide key methods and tools for the construction of green and low-carbon data centers and the consumption of renewable energy, and thus contribute to the realization of the country's "dual carbon" and "Eastern Data and Western Computing" strategic goals.

[0160] Reference Figure 4 Optionally, the user layer data center includes:

[0161] The power data acquisition module is used to obtain the power side parameters of the data center connected to the current user layer;

[0162] A computing resource acquisition module is used to obtain computing resources of the current user layer data center;

[0163] A data center information collection module, used to communicate with a target user layer data center receiving the non-real-time task when performing task space scale migration on the non-real-time task;

[0164] The task migration and tracking module is connected to the regional layer data center or the overall layer data center for communication, and is used to migrate the task space scale of non-real-time tasks and track the running status of the non-real-time tasks.

[0165] The data center information collection module specifically includes a list of candidate user-layer data centers within and outside the province that can accept migration tasks. When performing task space-scale migration, the data center information collection module can communicate with the candidate user-layer data center to obtain the power / computing power information of the candidate user-layer data center and other information such as the available transmission rate of the optical network and the number of routes in the transmission path. The power information is the power generation of renewable energy connected to the user-layer data center, which is used to determine whether the migrated task meets the low-carbon operation requirements; the computing power information is used to match the total amount of data of the migrated task, which corresponds to the memory required to run the migrated task; the available transmission rate of the optical network and the number of routes in the transmission path are used to calculate the carbon emissions generated during the task space-scale migration process.

[0166] The task migration and tracking module is used to perform user-level time migration and regional / overall-level space migration methods based on the data information in the above modules, and carry out multi-level computing power-power coordinated regulation. Furthermore, the task migration and tracking module can also be responsible for monitoring the energy supply control of renewable energy power generation of server clusters running real-time tasks.

[0167] In some embodiments, the power side parameters received by the power data acquisition module include:

[0168] Green electricity / non-green electricity supply power, cumulative power supply, consumed power, cumulative consumed electricity, green electricity supply ratio, wind / solar / hydro ratio in green electricity, non-green electricity coal / gas / nuclear ratio, green electricity utilization rate and green electricity in energy storage equipment.

[0169] In some embodiments, the computing resources received by the computing resource acquisition module include:

[0170] Task type, number of tasks, task status, task data volume, task computing resource usage, energy consumption per unit computing resource, cumulative running time, cumulative waiting time, waiting time threshold, task power consumption, cumulative power consumption, proportion of green power in task power consumption, data center path for task space migration, total computing resources and idle computing resources in the data center.

[0171] In some embodiments, the user layer data center further includes:

[0172] The parameter statistics module is used to count the carbon emission information of the current user layer data center, and the carbon emission information includes the carbon emission of the current user layer data center, the carbon emission generated by the task space scale migration and / or the green electricity surplus power.

[0173] By setting up the parameter statistics module, the parameter statistics module can count other parameters required by the system based on the existing data information of other modules, such as data center carbon emissions, carbon emissions generated by task space migration, and green electricity surplus power.

[0174] The parameter statistics module can calculate the power / computing power and other parameters in each province, the whole country and a custom area based on the data information of each single user-level data center, such as the overall green electricity utilization rate and the proportion of green electricity supply in the province or other provinces.

[0175] Embodiment 3:

[0176] The present invention based on the same inventive concept also provides a multi-level coordinated control device based on the dynamic load of a data center, comprising:

[0177] The task power consumption acquisition module is used to acquire the power generation power of the connected renewable energy and the power consumption of each connected non-real-time task;

[0178] A task matching module, configured to determine, in each of the non-real-time tasks, a non-real-time task that matches the generated power based on the generated power of the renewable energy and the power consumption of each non-real-time task;

[0179] A task processing module, used to utilize the renewable energy to perform power supply processing for matching non-real-time tasks;

[0180] A task waiting module is used to suspend non-real-time tasks that do not match the generated power to a waiting queue and obtain the waiting time;

[0181] The task migration module is used to migrate the non-real-time tasks whose waiting time exceeds the waiting time threshold to the user-level data center at the regional level or the overall level as migration tasks by using the task space scale migration strategy.

[0182] Optionally, the task migration module includes:

[0183] A migration task determination unit, used for determining a non-real-time task whose waiting time exceeds a waiting time threshold as a migration task;

[0184] A computing resource and power consumption determination unit, configured to determine the computing resources required to run the migration task according to the power consumption of the migration task;

[0185] A target data center search unit, configured to obtain, based on the computing resources and using a traversal search algorithm, a plurality of candidate user-layer data centers that meet the requirements for running the migration task in the current area or other areas;

[0186] A carbon emission determination unit, used to determine the carbon emission required to complete the task space scale migration with each candidate user layer data center according to the migration path with each candidate user layer data center;

[0187] A target data center determination unit is used to determine, in each carbon emission amount, an alternative user layer data center with a minimum carbon emission value as the target user layer data center;

[0188] The migration implementation unit is used to migrate the migration task to the target user layer data center.

[0189] Optionally, the carbon emission determination unit is specifically used to:

[0190] According to the total amount of data of the migration task, the available transmission rate of the optical network in the migration path of each alternative user layer data center, the characteristics of the optical communication equipment and the number of routes in the transmission path are comprehensively considered to determine the carbon emissions required to complete the spatial scale migration of the task with each alternative user layer data center.

[0191] Optionally, the target data center search unit is specifically used to:

[0192] Based on the computing resources, a traversal search algorithm is used to determine whether there is a user-level data center in the current area that meets the requirements for running the migration task:

[0193] If yes, taking the user layer data center in the current area that meets the requirements for running the migration task as a candidate user layer data center;

[0194] If not, a user layer data center that meets the migration task is obtained in other regions, and the user layer data center that meets the migration task in other regions is used as a candidate user layer data center.

[0195] Optionally, a task conversion module is also included, which is specifically used to:

[0196] When there is at least one non-real-time task whose waiting time exceeds the waiting time threshold and has not been migrated to the user-layer data center at the regional level or the overall level, the task space scale migration of the at least one non-real-time task is stopped, and the at least one non-real-time task is converted into a real-time task, the converted real-time task is powered by the power grid in the current area, and the converted real-time task is run.

[0197] Optionally, a task receiving module is also included, which is specifically used to:

[0198] Determine whether there is any suspended non-real-time task in the waiting queue;

[0199] If it does not exist, then receive and run the non-real-time tasks migrated from other user-layer data centers through the task space scale, clear the accumulated waiting time of the non-real-time tasks migrated from other user-layer data centers, and shorten the waiting time threshold of the non-real-time tasks migrated from other user-layer data centers.

[0200] Optionally, a power supply switching module is also included, which is specifically used for:

[0201] For the connected real-time task, obtaining the power consumption of running the connected real-time task;

[0202] Determine whether there is any suspended non-real-time task in the waiting queue;

[0203] If it is determined that there is no such a power source, the remaining power generated by the connected renewable energy source is obtained;

[0204] If the surplus power generated by the renewable energy matches the power consumption of the connected real-time task, the renewable energy is used to power the independent server, and the independent server runs the connected real-time task.

[0205] Embodiment 4:

[0206] like Figure 5 As shown, the present invention also provides an electronic device, which may be a computer device, a single-chip device, an intelligent mobile device, etc. The electronic device in this embodiment may include a processor, a memory, a transceiver component, etc. The memory, the processor, and the transceiver component are connected via a bus; the memory may be used to store an execution program, and an exemplary execution program may include instructions; the processor is used to execute the instructions stored in the memory. The memory may also be used to store data, which may be called and / or modified when the instructions are executed.

[0207] The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. It is the computing core and control core of the terminal, which is suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions in a storage medium to implement the corresponding method flow or corresponding functions, so as to implement the steps of a multi-level collaborative control method based on the dynamic load of a data center in the above-mentioned embodiment.

[0208] Embodiment 5:

[0209] Based on the same inventive concept, the present invention also provides a readable storage medium, specifically an electronic device readable storage medium (Memory), which is a memory device in an electronic device for storing programs and data. It can be understood that the storage medium here can include both built-in storage media in electronic devices and, of course, extended storage media supported by electronic devices. The storage medium provides a storage space that stores the operating system of the terminal. In addition, one or more instructions suitable for being loaded and executed by a processor are also stored in the storage space, and these instructions can be one or more execution programs (including program codes). It should be noted that the storage medium here can be a high-speed RAM memory or a non-volatile memory, such as at least one disk storage. The processor loads and executes one or more instructions stored in the storage medium, which can implement the steps of a multi-level collaborative control method based on the dynamic load of a data center in the above embodiment.

[0210] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0211] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0212] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1A function specified in one or more boxes.

[0213] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0214] The above are merely embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention are included in the scope of the claims of the present invention to be approved.

Claims

1. A multi-level collaborative control method based on dynamic load of data center, characterized in that: include: Based on each user-level data center, obtain the power generation of the connected renewable energy and the power consumption of each connected non-real-time task; Based on the power generation power of the renewable energy source and the power consumption of each non-real-time task, determining a non-real-time task matching the power generation power in each non-real-time task; Utilizing the renewable energy to power matching non-real-time tasks; Suspend the non-real-time tasks that do not match the generated power to the waiting queue, and obtain the waiting time; If there are non-real-time tasks whose waiting time exceeds the waiting time threshold, the task space scale migration strategy is used to migrate the non-real-time tasks whose waiting time exceeds the waiting time threshold as migration tasks to the user-level data center at the regional level or the overall level.

2. The method according to claim 1, characterized in that The task space scale migration strategy is used to migrate non-real-time tasks whose waiting time exceeds the waiting time threshold as migration tasks to a user-level data center at the regional level or the overall level, including: Non-real-time tasks whose waiting time exceeds the waiting time threshold are considered as migration tasks; Determining computing resources required to run the migration task according to the power consumption of the migration task; According to the computing resources, using a traversal search algorithm, a plurality of candidate user-layer data centers that meet the requirements for running the migration task are obtained in the current area or other areas; Determine, according to the migration path with each candidate user layer data center, the carbon emissions required to complete the task space scale migration with each candidate user layer data center; In each carbon emission amount, determine the candidate user layer data center with the minimum carbon emission amount as the target user layer data center; Migrate the migration task to the target user layer data center.

3. The method according to claim 2, characterized in that The step of determining the carbon emissions required to complete the task space scale migration with each candidate user layer data center according to the migration path with each candidate user layer data center includes: According to the total amount of data of the migration task, the available transmission rate of the optical network in the migration path of each alternative user layer data center, the characteristics of the optical communication equipment and the number of routes in the transmission path are comprehensively considered to determine the carbon emissions required to complete the spatial scale migration of the task with each alternative user layer data center.

4. The method according to claim 2, characterized in that The step of obtaining, based on the computing resources, a plurality of candidate user-layer data centers that meet the requirements for running the migration task in the current area or other areas by using a traversal search algorithm includes: Based on the computing resources, a traversal search algorithm is used to determine whether there is a user-level data center in the current area that meets the requirements for running the migration task: If yes, taking the user layer data center in the current area that meets the requirements for running the migration task as a candidate user layer data center; If not, a user layer data center that meets the migration task is obtained in other regions, and the user layer data center that meets the migration task in other regions is used as a candidate user layer data center.

5. The method according to claim 1, characterized in that After suspending the non-real-time tasks that do not match the generated power to the waiting queue and obtaining the waiting time, the method further includes: If there is at least one non-real-time task whose waiting time exceeds the waiting time threshold and has not been migrated to the user-layer data center at the regional level or the overall level, the task space scale migration of the at least one non-real-time task is stopped, and the at least one non-real-time task is converted into a real-time task, the converted real-time task is powered by the power grid in the current area, and the converted real-time task is run.

6. The method according to claim 1, characterized in that After the renewable energy is used to power the matching non-real-time tasks, the method further includes: Determine whether there is any suspended non-real-time task in the waiting queue; If it does not exist, then receive and run the non-real-time tasks migrated from other user-layer data centers through the task space scale, clear the accumulated waiting time of the non-real-time tasks migrated from other user-layer data centers, and shorten the waiting time threshold of the non-real-time tasks migrated from other user-layer data centers.

7. The method according to claim 1, characterized in that After the renewable energy is used to power the matching non-real-time tasks, the method further includes: For the connected real-time task, obtaining the power consumption of running the connected real-time task; Determine whether there is any suspended non-real-time task in the waiting queue; If it is determined that there is no such a power source, the remaining power generated by the connected renewable energy source is obtained; If the surplus power generated by the renewable energy matches the power consumption of the connected real-time task, the renewable energy is used to power the independent server, and the independent server runs the connected real-time task.

8. A multi-level collaborative control system based on dynamic load of data center, characterized in that: The system includes a plurality of user layer data centers, a plurality of regional layer data centers and an overall layer data center; each of the regional layer data centers is communicatively connected to a plurality of the user layer data centers, and each of the regional layer data centers is connected to the overall layer data center; Each user layer data center is used to perform the steps of a multi-level coordinated control method of a central dynamic load as described in any one of claims 1-7; The regional data center is used to transfer the remaining computing resources of other user data centers in the region and the surplus power of the connected renewable energy to the user data center to be migrated to the task space scale; The overall layer data center is used to transfer the remaining computing resources of other user layer data centers in other areas and the surplus power generation power of the connected renewable energy sources to the user layer data center to be migrated to the task space scale through the corresponding regional layer data center.

9. The system according to claim 8, characterized in that The user layer data center includes: The power data acquisition module is used to obtain the power side parameters of the data center connected to the current user layer; A computing resource acquisition module is used to obtain computing resources of the current user layer data center; A data center information collection module, used to communicate with a target user layer data center receiving the non-real-time task when performing task space scale migration on the non-real-time task; The task migration and tracking module is connected to the regional layer data center or the overall layer data center for performing task space scale migration on non-real-time tasks and tracking the running status of the non-real-time tasks.

10. The system according to claim 9, characterized in that The power side parameters received by the power data acquisition module include: Green electricity / non-green electricity supply power, cumulative power supply, consumed power, cumulative consumed electricity, green electricity supply ratio, wind / solar / hydro ratio in green electricity, non-green electricity coal / gas / nuclear ratio, green electricity utilization rate and green electricity in energy storage equipment.

11. The system according to claim 9, characterized in that The computing resources received by the computing resource acquisition module include: Task type, number of tasks, task status, task data volume, task computing resource usage, energy consumption per unit computing resource, cumulative running time, cumulative waiting time, waiting time threshold, task power consumption, cumulative power consumption, proportion of green power in task power consumption, data center path for task space migration, total computing resources and idle computing resources in the data center.

12. The system according to claim 9, characterized in that The user layer data center also includes: The parameter statistics module is used to count the carbon emission information of the current user layer data center, and the carbon emission information includes the carbon emission of the current user layer data center, the carbon emission generated by the task space scale migration and / or the green electricity surplus power.

13. A multi-level collaborative control device based on dynamic load of data center, characterized in that: include: The task power consumption acquisition module is used to acquire the power generation power of the connected renewable energy and the power consumption of each connected non-real-time task; A task matching module, configured to determine, in each of the non-real-time tasks, a non-real-time task that matches the generated power based on the generated power of the renewable energy and the power consumption of each non-real-time task; A task processing module, used to utilize the renewable energy to perform power supply processing for matching non-real-time tasks; A task waiting module is used to suspend non-real-time tasks that do not match the generated power to a waiting queue and obtain the waiting time; The task migration module is used to migrate the non-real-time tasks whose waiting time exceeds the waiting time threshold to the user-level data center at the regional level or the overall level as migration tasks by using the task space scale migration strategy.

14. An electronic device, characterized in that: include: at least one processor and memory; The memory and the processor are connected via a bus; The memory is used to store one or more programs; When the one or more programs are executed by the at least one processor, a multi-level collaborative control method based on dynamic load of a data center as described in any one of claims 1 to 7 is implemented.

15. A readable storage medium, characterized in that: An execution program is stored thereon, and when the execution program is executed, the multi-level collaborative control method based on the dynamic load of the data center as described in any one of claims 1 to 7 is implemented.

Images