Server power-on method and device, electronic equipment and storage medium

Abstract

The application provides a power-on method and device of a server, electronic equipment and a storage medium, and relates to the technical field of data processing. The method comprises the following steps: constructing a server power-on sequence through an ant colony algorithm according to the amount of pheromone of the time difference of reaching a power peak value in the start-up process of each server in a server cluster to be powered on; updating the amount of pheromone of the time difference of reaching a power peak value in the start-up process of each server in the server power-on sequence according to the total time of reaching a power consumption peak value of each server in the server power-on sequence; continuously constructing the server power-on sequence through the ant colony algorithm according to the updated amount of pheromone until a first preset condition is met, and obtaining a target server power-on sequence.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a method, apparatus, electronic device, and storage medium for powering on a server. Background Technology

[0002] With the continuous development of digital technology, the demand for servers is increasing. Server maintenance personnel usually power on servers by issuing power-on commands. When there are many servers in a data center, maintenance personnel issue power-on commands to a batch of servers to power them on. However, powering on multiple servers at the same time is very likely to cause a "power surge" phenomenon, which can damage the circuit.

[0003] Therefore, how to effectively power on servers has become an urgent problem to be solved in the industry. Summary of the Invention

[0004] This invention provides a server power-on method, apparatus, electronic device, and storage medium to solve the problem of how to effectively power on a server in the prior art.

[0005] This invention provides a method for powering on a server, comprising: Based on the pheromone content of the time difference between the time difference when each server in the server cluster reaches its power peak during the power-on process, the power-on sequence of the servers is constructed using the ant colony algorithm. Based on the total time for each server to reach its peak power consumption during the staggered power-up process corresponding to the power-up sequence of each server, update the pheromone content of the time difference between the power-up times of each server during the power-up process of the server power-up sequence. Based on the updated pheromone levels, the ant colony algorithm is used to continue constructing the server power-on sequence until the first preset condition is met, thus obtaining the target server power-on sequence.

[0006] According to a server power-on method provided by the present invention, the step of constructing the server power-on sequence using an ant colony algorithm based on the initial pheromone content of the time difference between the time difference in the power-up process of each server in the server cluster to be powered on includes: If the number of servers in the constructed server power-on sequence is less than a first preset threshold, the ants in the ant colony algorithm calculate the state transition probability based on heuristic information and pheromone quantity, and select the next server to be powered on based on the state transition probability. The next server to be powered on is written into the server power-on sequence in order, and the relationship between the number of servers in the server power-on sequence and the first preset threshold is further compared. If the number of servers in the server power-on sequence is still less than the first preset threshold, the next server to be powered on will continue to be selected according to the ant colony algorithm. If the number of servers in the server power-on sequence is greater than or equal to the first preset threshold, each ant in the ant colony algorithm completes the traversal of the servers in the cluster to be powered on, thus completing the construction of the server power-on sequence.

[0007] According to a server power-on method provided by the present invention, the step of updating the pheromone content of the time difference between the power-up times of each server during the power-on process in the server power-on sequence, based on the total time for the servers to reach their peak power consumption during staggered power-up according to the power-on sequence of each server, includes: The first pheromone increment is determined based on the total time it takes for the optimal server to reach its peak power consumption during off-peak hours. The total pheromone increment is determined based on the first pheromone increment and the second pheromone increment left by each ant during the construction of the power-on sequence on the server. Based on the total pheromone increment, update the pheromone amount of each server in the power-on sequence to the time difference between reaching peak power during the power-on process, and obtain the updated pheromone amount.

[0008] According to a server power-on method provided by the present invention, the step of continuing to construct the server power-on sequence based on the updated pheromone content using the ant colony algorithm until a first preset condition is met to obtain the target server power-on sequence includes: In the ant colony algorithm, the ants select the next server to be powered on based on the updated pheromone levels and heuristic information, in order to establish the server power-on sequence. Based on the reconstructed server power-on sequence, continue to update the pheromone content of each server in the power-on sequence during the power-on process to reach the power peak, until the first preset condition is met, then stop updating, and obtain multiple sets of server power-on sequences; The target power-on sequence of servers in multiple sets of server power-on sequences is determined based on the total time it takes for servers to reach their peak power consumption during off-peak periods.

[0009] According to an embodiment of the present invention, a server power-on method is provided, wherein the method for calculating the state transition probability specifically includes: ; In the ant colony algorithm, the ant k (k=1, 2, ..., n) uses a tabu list during its movement. finish k To record currently powered-on servers, using To record the servers that are allowed to boot up next. ={S-finish k ｝ ; yij(x) The amount of information remaining at each time difference. This provides heuristic information for the time differences.

[0010] According to an embodiment of the present invention, a server power-on method is provided, wherein the calculation method for the first pheromone increment includes: ; in, Indicates the known optimal power-on sequence of servers. The total time it takes for all servers to reach their peak power consumption during off-peak hours. L It is a positive number.

[0011] According to an embodiment of the present invention, a server power-on method is provided, wherein the method for calculating the total pheromone increment includes: ; ; ; in, Indicates the first k The total time it takes for all servers to reach their peak power consumption during this server startup process (using only one ant). This represents the time difference in this iteration ( i , j The increase in pheromones on the surface, h It is a coefficient. （1-h） express x arrive x+ d The pheromone evaporation rate at any given time.

[0012] The present invention also provides a server power-on device, comprising: The module is used to construct the power-on sequence of servers based on the pheromone content of the time difference between the time difference when each server in the power-on cluster reaches its power peak during the power-on process, using the ant colony algorithm. The update module is used to update the pheromone content of the time difference between the power peak times of each server during the power-on process in the power-on sequence of each server, based on the total time for the servers to reach their peak power consumption during the staggered power-up process. The power-on module is used to continue constructing the server power-on sequence based on the updated pheromone content using the ant colony algorithm until the first preset condition is met, thereby obtaining the target server power-on sequence.

[0013] According to a server power-on device provided by the present invention, the device is further configured to: If the number of servers in the constructed server power-on sequence is less than a first preset threshold, the ants in the ant colony algorithm calculate the state transition probability based on heuristic information and pheromone quantity, and select the next server to be powered on based on the state transition probability. The next server to be powered on is written into the server power-on sequence in order, and the relationship between the number of servers in the server power-on sequence and the first preset threshold is further compared. If the number of servers in the server power-on sequence is still less than the first preset threshold, the next server to be powered on will continue to be selected according to the ant colony algorithm. If the number of servers in the server power-on sequence is greater than or equal to the first preset threshold, each ant in the ant colony algorithm completes the traversal of the servers in the cluster to be powered on, thus completing the construction of the server power-on sequence.

[0014] According to a server power-on device provided by the present invention, the device is further configured to: The first pheromone increment is determined based on the total time it takes for the optimal server to reach its peak power consumption during off-peak hours. The total pheromone increment is determined based on the first pheromone increment and the second pheromone increment left by each ant during the construction of the power-on sequence on the server. Based on the total pheromone increment, update the pheromone amount of each server in the power-on sequence to the time difference between reaching peak power during the power-on process, and obtain the updated pheromone amount.

[0015] According to a server power-on device provided by the present invention, the device is further configured to: In the ant colony algorithm, the ants select the next server to be powered on based on the updated pheromone levels and heuristic information, in order to establish the server power-on sequence. Based on the reconstructed server power-on sequence, continue to update the pheromone content of each server in the power-on sequence during the power-on process to reach the power peak, until the first preset condition is met, then stop updating, and obtain multiple sets of server power-on sequences; The target power-on sequence of servers in multiple sets of server power-on sequences is determined based on the total time it takes for servers to reach their peak power consumption during off-peak periods.

[0016] According to a server power-on device provided by the present invention, the method for calculating the state transition probability specifically includes: ; In the ant colony algorithm, the ant k (k=1, 2, ..., n) uses a tabu list during its movement. finish k To record currently powered-on servers, using To record the servers that are allowed to boot up next. ={S-finish k ｝ ; yij(x) The amount of information remaining at each time difference. This provides heuristic information for the time differences.

[0017] According to a server power-on device provided by the present invention, the method for calculating the first pheromone increment includes: ; in, Indicates the known optimal power-on sequence of servers. The total time it takes for all servers to reach their peak power consumption during off-peak hours. L It is a positive number.

[0018] According to a server power-on device provided by the present invention, the method for calculating the total pheromone increment includes: ; ; ; in, Indicates the first k The total time it takes for all servers to reach their peak power consumption during this server startup process (using only one ant). This represents the time difference in this iteration ( i , j The increase in pheromones on the surface, h It is a coefficient. （1-h） express x arrive x+ d The pheromone evaporation rate at any given time.

[0019] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the server power-on method as described above.

[0020] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the server power-on method as described above.

[0021] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the server power-on method as described above.

[0022] The server power-on method, apparatus, electronic device, and storage medium provided by this invention can effectively construct the server power-on sequence by using a swarm intelligence optimization algorithm that simulates ant colony foraging behavior. By staggering power-on times, it effectively avoids power surges caused by multiple servers reaching their power peak simultaneously, reducing the energy consumption and operating costs of data centers. Furthermore, it uses time difference pheromone quantity to guide the ant colony algorithm to search for the optimal server power-on sequence, and iteratively updates the pheromone quantity to approximate the optimal solution. This solves the power surge problem that may be caused when powering up data center servers in batches, and optimizes the server startup sequence to achieve fast and balanced power consumption. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0024] Figure 1 This is a schematic diagram of a server off-peak power-on system provided in an embodiment of this application.

[0025] Figure 2 This is a schematic diagram of the method for calculating the time to reach the power peak during the power-on process, provided in an embodiment of this application.

[0026] Figure 3 This is a schematic diagram of the server power-on method provided in an embodiment of this application.

[0027] Figure 4 A schematic diagram of the server off-peak power-on method provided in this application embodiment.

[0028] Figure 5 This is a schematic diagram of the power-on device structure for a server provided in an embodiment of this application.

[0029] Figure 6 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0030] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0031] Ant Colony Optimization (ACO) is a swarm intelligence optimization algorithm that simulates the collaborative process of real ant colonies by studying their collective foraging behavior. ACO has been successfully applied to a range of combinatorial optimization problems, including the Traveling Salesman Problem.

[0032] The Traveling Salesman Problem (TSP) is as follows: Given a set of n cities C = {c1, c2, ..., cn} and the lengths of the travel paths between the cities dij (1 ≤ i ≤ n, 1 ≤ j ≤ n, i ≠ j), a traveling salesman starts from one of the cities, visits each city exactly once, and finally returns to the starting city. The goal of the TSP is to find a closed loop with the shortest possible path length.

[0033] The staggered power-on of data center servers can be transformed into a combinatorial optimization problem. Given a set of m servers S = {s1, s2, ..., sm} and the time difference tij between the servers reaching their peak power during the power-on process (1≤i≤m, 1≤j≤m, i≠j), the problem is to start powering on a certain server, complete the power-on of all servers, and achieve the shortest power-on time for all servers while staggering the peak power of the servers.

[0034] Figure 1 This is a schematic diagram of a server off-peak power-on system provided in an embodiment of this application, as shown below. Figure 1 As shown, a data center typically has multiple server racks, with each rack usually housing multiple servers. One server in the data center can be selected as a control server. The control server's baseboard controller can be used to aggregate data from the data center servers and perform ant colony algorithm calculations. A control workstation is optional. The control workstation connects to all servers in the data center via a network and can also be used to aggregate data from the data center servers and perform ant colony algorithm calculations.

[0035] Figure 2 A flowchart illustrating the method for calculating the time to reach peak power during the power-on process, as provided in this application embodiment, includes: The control server or control workstation sends a query command to the baseboard management controller of all servers to initiate the calculation of peak power time.

[0036] Check if each server has ever been powered on. If the server has been powered on, recalculate the historical power-on time; if it has not been powered on, recalculate the time for the same server model.

[0037] For servers with a history of power-on, calculate the average time to reach peak power after historical power-on. Send the calculation result to the control server or workstation for time correction to avoid time overlap with other servers. If time overlap is found, add a random time accurate to milliseconds and record it to prevent future repetition.

[0038] If the server has not been booted, set the historical boot correction time to 0.

[0039] For the current server, query the average time it takes for the same model to reach peak power upon startup. Check and correct the time to avoid duplication with the already calculated time, also adding a random time accurate to milliseconds, and record it.

[0040] If the server has been powered on before, set the power-on correction time for the same model to 0.

[0041] The formula for calculating the time ti for each server to reach its peak power during startup is as follows: (1)

[0042] The formula for calculating the time difference tij between servers reaching peak power upon startup is shown in equation (2): (2)

[0043] This calculation process allows for the precise calculation of the time it takes for each server to reach its peak power consumption upon startup, as well as the time difference between servers. This facilitates off-peak power-on for servers, optimizes power management in data centers, reduces surge risks, and improves overall energy efficiency and system stability.

[0044] Figure 3 This is a schematic diagram of the server power-on method provided in the embodiments of this application, such as... Figure 3 As shown, it includes: Step 310: Based on the pheromone content of the time difference between the power peak times of each server in the power-on cluster, construct the power-on sequence of the servers using the ant colony algorithm. In this embodiment, before the ant colony algorithm begins, it is first necessary to initialize the time difference pheromone content of each server in the power-on server cluster to reach its power peak. This pheromone content can be estimated based on historical data, simulation results, or experience, representing the time difference in the arrival of the power peak between different servers during the power-on process.

[0045] The ant colony algorithm is launched, where each ant represents a search path for the possible power-on order of servers. At the start of the algorithm, each ant randomly selects a server as its starting point.

[0046] Each ant determines which server to power up next based on its current pheromone levels and heuristic information (such as time difference). Paths with higher pheromone levels represent the power-up order that the ants previously considered optimal during their search, while heuristic information guides the ants to choose servers that can further optimize the off-peak power-up effect.

[0047] When making decisions, ants calculate the probability of moving from their current server to another server. This probability is determined by both the amount of pheromone and heuristic information. Typically, this probability calculation formula takes into account the relative importance of pheromones and heuristic information.

[0048] In this embodiment, each ant constructs a complete power-on sequence for the server step by step according to a probabilistic selection mechanism. Once an ant completes a full traversal, it has constructed a possible power-on sequence.

[0049] Step 320: Based on the total time for each server to reach its peak power consumption during the power-on process corresponding to the power-on sequence of each server, update the pheromone content of the time difference between the power-on times of each server during the power-on process. In this embodiment, after constructing a server power-on sequence, each ant calculates the total time it takes for all servers in that sequence to reach their peak power consumption. This typically involves evaluating the power curve of each server at startup and calculating the peak power consumption and its time distribution throughout the startup process. Based on the calculated total staggered time, the quality of each power-on sequence is evaluated. This evaluation forms the basis for pheromone updates, because a shorter total time indicates that the power-on sequence is more likely to be close to the optimal solution.

[0050] In this application, firstly, the amount of pheromones that reach the power peak during the power-up process of all servers is volatilized, that is, the amount of all pheromones is reduced, so as to avoid the search space solidification caused by the rapid accumulation of pheromones.

[0051] Then, based on the total time of staggered power-on sequence found by each ant, the pheromone content of the corresponding time difference is enhanced. Specifically, for the time difference between each pair of servers, if the time difference results in a better staggered effect (i.e., a shorter total time) in a certain power-on sequence, the pheromone content of that time difference is increased.

[0052] By continuously iterating the above process, the distribution of pheromones gradually tilts towards a better power-on sequence for the server, thereby guiding the ants in subsequent iterations to be more likely to find a shorter total time for peak shaving.

[0053] Step 330: Based on the updated pheromone levels, continue to construct the server power-on sequence using the ant colony algorithm until the first preset condition is met, and obtain the target server power-on sequence.

[0054] In this embodiment, in each iteration, each ant begins a new round of power-on sequence construction based on the updated pheromone quantity in step 120. The updated pheromone quantity reflects the effective power-on sequences discovered in previous iterations; therefore, the updated pheromone quantity will guide the ants to be more inclined to choose those power-on paths that have previously proven to be effective.

[0055] Each ant, based on the updated pheromone levels and heuristic information (such as the time difference between servers reaching peak power), gradually constructs a new power-on sequence for the servers through a probabilistic selection mechanism.

[0056] After each ant constructs a power-on sequence, it calculates the total time it takes for all servers under that sequence to reach their peak power consumption. This total time is a key metric for evaluating the quality of the new power-on sequence.

[0057] Based on the quality of the newly constructed power-on sequence, the pheromone levels are updated again. This step further strengthens the pheromones of power-on paths that result in shorter total off-peak times, while weakening the pheromones of less effective paths.

[0058] After each iteration, check if the maximum number of iterations has been reached. Once the termination condition is met, the algorithm terminates and outputs the currently found optimal server power-up sequence as the target power-up sequence. This sequence achieves the shortest peak power consumption time for the servers while meeting the requirements for staggered power-up.

[0059] In this embodiment, a swarm intelligence optimization algorithm simulating ant colony foraging behavior can effectively construct the server power-on sequence. By staggering power-on times, it effectively avoids power surges caused by multiple servers reaching their power peak simultaneously, reducing the energy consumption and operating costs of the data center. Furthermore, by using time difference pheromone quantity to guide the ant colony algorithm in searching for the optimal server power-on sequence, and by iteratively updating the pheromone quantity to approximate the optimal solution, it solves the power surge problem that may be caused when data center servers are powered on in batches, and optimizes the server power-on sequence to achieve fast and balanced power consumption.

[0060] Optionally, the step of constructing the server power-on sequence using the ant colony algorithm based on the initial pheromone content of the time difference between the time when each server in the server cluster reaches its power-on peak includes: If the number of servers in the constructed server power-on sequence is less than a first preset threshold, the ants in the ant colony algorithm calculate the state transition probability based on heuristic information and pheromone quantity, and select the next server to be powered on based on the state transition probability. The next server to be powered on is written into the server power-on sequence in order, and the relationship between the number of servers in the server power-on sequence and the first preset threshold is further compared. If the number of servers in the server power-on sequence is still less than the first preset threshold, the next server to be powered on will continue to be selected according to the ant colony algorithm. If the number of servers in the server power-on sequence is greater than or equal to the first preset threshold, each ant in the ant colony algorithm completes the traversal of the servers in the cluster to be powered on, thus completing the construction of the server power-on sequence.

[0061] In this embodiment, ants select the server boot order based on heuristic information and pheromone levels. This process begins with initializing the pheromone levels, which is the time difference between each server reaching its peak power upon boot. These initial pheromone levels provide the ants with preliminary guidance, helping them make selections within the search space.

[0062] As the algorithm progresses, the ant calculates the state transition probability based on the current pheromone level and heuristic information (such as the time difference between servers). This probability determines which server the ant will choose to power on next. Each selection is based on a probabilistic model, where paths with higher pheromone levels are more likely to be selected because they represent solutions considered optimal in previous iterations.

[0063] Ants select the next server based on the calculated state transition probabilities and add it sequentially to the power-on sequence list. As the selection process progresses, this list gradually grows until it reaches or exceeds a preset threshold. This threshold is a crucial parameter that defines the minimum number of servers that should be in the power-on sequence, ensuring its practicality and effectiveness.

[0064] If the number of servers in the power-on sequence has not yet reached this preset threshold, Ant will continue to select the next server to power on until this condition is met. This process is iterative, and each iteration may update the pheromone content to reflect newly discovered better or more feasible power-on sequences.

[0065] When the number of servers in the power-on sequence reaches or exceeds a preset threshold, the ants in the ant colony algorithm complete their traversal of the server cluster to be powered on, thus completing the construction of the server power-on sequence. This power-on sequence aims to achieve staggered power-on of servers to optimize power consumption in the data center and avoid power surges.

[0066] Throughout the process, pheromone updates not only reflect the discoveries made by the ants during their search but also guide subsequent search directions. In this way, the ant colony algorithm can gradually approach the optimal server power-on sequence, thereby achieving efficient and balanced power management and improving the energy efficiency and stability of data centers.

[0067] Optionally, updating the pheromone content for the time difference between the power peak times of each server during the power-on process in the server power-on sequence, based on the total time for the servers to reach their peak power consumption during staggered power-up according to the power-on sequence of each server, includes: The first pheromone increment is determined based on the total time it takes for the optimal server to reach its peak power consumption during off-peak hours. The total pheromone increment is determined based on the first pheromone increment and the second pheromone increment left by each ant during the construction of the power-on sequence on the server. Based on the total pheromone increment, update the pheromone amount of each server in the power-on sequence to the time difference between reaching peak power during the power-on process, and obtain the updated pheromone amount.

[0068] In this embodiment, an optimal total time is determined for the server to reach its peak power consumption. This time is obtained from the power-on sequence constructed by all ants and represents the best performance metric found in the current iteration. Based on this optimal total time, the algorithm calculates the first pheromone increment, which reflects the contribution of the optimal solution to the pheromone update.

[0069] The method for calculating the first pheromone increment includes: ; in, Indicates the known optimal power-on sequence of servers. The total time it takes for all servers to reach their peak power consumption during off-peak hours. L It is a positive number.

[0070] Each ant leaves its own pheromone increment during the process of constructing the server power-on sequence; this is the second pheromone increment. This increment is determined based on the performance of the ant's own traversal path, reflecting the ant's exploration and evaluation of the solution space during the search process.

[0071] The algorithm combines the first pheromone increment with the second pheromone increment left by each ant to determine the total pheromone increment. This total increment forms the basis for the algorithm to update the pheromone level, and it integrates the influence of the global optimal solution and the local discoveries made during the search process of all ants.

[0072] The method for calculating the total pheromone increment includes: ; ; ; in, Indicates the first k The total time it takes for all servers to reach their peak power consumption during this server startup process (using only one ant). This represents the time difference in this iteration ( i , j The pheromone increment on the surface, i.e., the second pheromone increment. h It is a coefficient. （1-h） express x arrive x+d The pheromone evaporation rate at any given time.

[0073] Finally, the algorithm updates the pheromone levels based on the total pheromone increment during the power-on sequence of each server, indicating the time difference between reaching peak power during startup. In subsequent iterations, this guides the ants to more likely select server power-on sequences that contribute to achieving a shorter overall peak-shifting time.

[0074] Optionally, the step of continuing to construct the server power-on sequence using the ant colony algorithm based on the updated pheromone levels until a first preset condition is met to obtain the target server power-on sequence includes: In the ant colony algorithm, the ants select the next server to be powered on based on the updated pheromone levels and heuristic information, in order to establish the server power-on sequence. Based on the reconstructed server power-on sequence, continue to update the pheromone content of each server in the power-on sequence during the power-on process to reach the power peak, until the first preset condition is met, then stop updating, and obtain multiple sets of server power-on sequences; The target power-on sequence of servers in multiple sets of server power-on sequences is determined based on the total time it takes for servers to reach their peak power consumption during off-peak periods.

[0075] In this embodiment, each ant in the ant colony algorithm calculates the state transition probability based on the updated pheromone levels and heuristic information (such as the time difference between servers reaching peak power). The state transition probability guides the ants to select the next server to be powered on, thus establishing the server power-on sequence.

[0076] Ants select servers based on calculated probabilities and write the selected servers into a power-on sequence list in order. This process is repeated until a complete server power-on sequence is constructed.

[0077] After each server power-up sequence is constructed, the algorithm updates the pheromone level based on the total time it takes for the servers to reach their peak power consumption under that sequence. Updating the pheromone level reflects the performance of the newly discovered power-up sequence and guides subsequent search processes.

[0078] The ants continue to select the next server to power on based on the updated pheromone levels and heuristic information, repeatedly constructing the power-on sequence. Each iteration may generate a new power-on sequence, which is used to update the pheromone levels. The iterative process continues until a first preset condition is met, such as reaching the maximum number of iterations, the pheromone update change being less than a preset threshold, or the quality of the solution not significantly improving in consecutive iterations. Once the preset condition is met, the algorithm stops iterating.

[0079] Among multiple server power-on sequences, the optimal power-on sequence is determined based on the total time it takes for the servers to reach their peak power consumption during off-peak periods. This target power-on sequence is found across all iterations and achieves the shortest peak power consumption time while ensuring off-peak power-on.

[0080] In this embodiment, the ant colony algorithm can effectively optimize the power-on sequence of servers, reduce the peak power consumption of the data center when servers are powered on in batches, and improve overall energy efficiency and system stability. This method achieves effective management of server power-on strategy by dynamically adjusting pheromone levels and continuously optimizing the power-on sequence.

[0081] Figure 4 This is a schematic diagram of the server off-peak power-on method provided in the embodiments of this application, as shown below. Figure 4 As shown, it includes: Initiate the optimization process of staggered server power-on to prepare for iterative search using the ant colony algorithm.

[0082] An initial pheromone level is set for the time difference between the power-on time of each server in the data center and its peak power level. This pheromone level forms the basis for the algorithm's initial search and guides the ants' decisions regarding the server power-on order.

[0083] Before each iteration begins, check if the preset termination condition has been met. In the current iteration... Nc The algorithm stops iterating when the preset value is reached or exceeded.

[0084] Each ant randomly selects a server as a starting point and begins to construct a possible server power-on sequence. This sequence aims to stagger the power-on times of the servers to reduce peak power consumption.

[0085] Ants use a state transition probability model, based on heuristic information and pheromone concentration, to select the next server to be powered on. This process involves probability calculations to determine the likelihood of each server being selected.

[0086] The ants select servers one by one until the entire boot sequence is completed. After selecting each server, the loop parameters are updated, and it is checked whether all servers have been selected.

[0087] For each ant's boot sequence, calculate the total time it takes for the server to reach its peak power consumption under that sequence. This total time is a key metric for evaluating the quality of the boot sequence.

[0088] The pheromone levels are updated based on the startup sequence established by the ants and the corresponding total staggered startup time. This update process involves the volatilization and enhancement of pheromones to reflect newly discovered optimal startup sequences.

[0089] If the termination condition is not met, continue iterating to construct the server power-on order. If the termination condition is met, the algorithm terminates and outputs the final target server power-on order.

[0090] In this embodiment, the ant colony algorithm continuously explores and optimizes the server power-on sequence, gradually reducing the total time for servers to reach peak power consumption during off-peak periods. Each iteration brings the pheromone distribution closer to the optimal solution, thus achieving an efficient server power-on sequence when the termination condition is met. This method not only improves the energy efficiency of data centers but also enhances system stability and response speed.

[0091] The server power-on device provided by the present invention is described below. The server power-on device described below can be referred to in correspondence with the server power-on method described above.

[0092] Figure 5 This is a schematic diagram of the server power-on device structure provided in the embodiments of this application, such as... Figure 5 As shown, it includes: The construction module 510 is used to construct the power-on sequence of the servers based on the pheromone content of the time difference between the power peak times of each server in the power-on cluster. The update module 520 is used to update the pheromone content of the time difference between the power peak times of each server during the power-on process in the power-on sequence of the servers, based on the total time for the servers to reach the peak power consumption during the staggered power-on process corresponding to the power-on sequence of each server. The power-on module 530 is used to continue constructing the server power-on sequence according to the updated pheromone content using the ant colony algorithm until the first preset condition is met, thereby obtaining the target server power-on sequence.

[0093] According to a server power-on device provided by the present invention, the device is further configured to: If the number of servers in the constructed server power-on sequence is less than a first preset threshold, the ants in the ant colony algorithm calculate the state transition probability based on heuristic information and pheromone quantity, and select the next server to be powered on based on the state transition probability. The next server to be powered on is written into the server power-on sequence in order, and the relationship between the number of servers in the server power-on sequence and the first preset threshold is further compared. If the number of servers in the server power-on sequence is still less than the first preset threshold, the next server to be powered on will continue to be selected according to the ant colony algorithm. If the number of servers in the server power-on sequence is greater than or equal to the first preset threshold, each ant in the ant colony algorithm completes the traversal of the servers in the cluster to be powered on, thus completing the construction of the server power-on sequence.

[0094] According to a server power-on device provided by the present invention, the device is further configured to: The first pheromone increment is determined based on the total time it takes for the optimal server to reach its peak power consumption during off-peak hours. The total pheromone increment is determined based on the first pheromone increment and the second pheromone increment left by each ant during the construction of the power-on sequence on the server. Based on the total pheromone increment, update the pheromone amount of each server in the power-on sequence to the time difference between reaching peak power during the power-on process, and obtain the updated pheromone amount.

[0095] According to a server power-on device provided by the present invention, the device is further configured to: In the ant colony algorithm, the ants select the next server to be powered on based on the updated pheromone levels and heuristic information, in order to establish the server power-on sequence. Based on the reconstructed server power-on sequence, continue to update the pheromone content of each server in the power-on sequence during the power-on process to reach the power peak, until the first preset condition is met, then stop updating, and obtain multiple sets of server power-on sequences; The target power-on sequence of servers in multiple sets of server power-on sequences is determined based on the total time it takes for servers to reach their peak power consumption during off-peak periods.

[0096] According to a server power-on device provided by the present invention, the method for calculating the state transition probability specifically includes: ; In the ant colony algorithm, the ant k (k=1, 2, ..., n) uses a tabu list during its movement. finish k To record currently powered-on servers, using To record the servers that are allowed to boot up next. ={S-finish k ｝ ; yij(x) The amount of information remaining at each time difference. This provides heuristic information for the time differences.

[0097] According to a server power-on device provided by the present invention, the method for calculating the first pheromone increment includes: ; in, Indicates the known optimal power-on sequence of servers. The total time it takes for all servers to reach their peak power consumption during off-peak hours. L It is a positive number.

[0098] According to a server power-on device provided by the present invention, the method for calculating the total pheromone increment includes: ; ; ; in, Indicates the first k The total time it takes for all servers to reach their peak power consumption during this server startup process (using only one ant). This represents the time difference in this iteration ( i , j The increase in pheromones on the surface, h It is a coefficient. （1-h） express x arrive x+ d The pheromone evaporation rate at any given time.

[0099] In this embodiment, a swarm intelligence optimization algorithm simulating ant colony foraging behavior can effectively construct the server power-on sequence. By staggering power-on times, it effectively avoids power surges caused by multiple servers reaching their power peak simultaneously, reducing the energy consumption and operating costs of the data center. Furthermore, it uses time difference pheromone quantity to guide the ant colony algorithm to search for the optimal server power-on sequence, and iteratively updates the pheromone quantity to approximate the optimal solution. This solves the power surge problem that may be caused when powering on a batch of data center servers and optimizes the server startup sequence to achieve fast and balanced power consumption.

[0100] Figure 6 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 6As shown, the electronic device may include: a processor 610, a communication interface 620, a memory 630, and a communication bus 640, wherein the processor 610, the communication interface 620, and the memory 630 communicate with each other through the communication bus 640. The processor 610 can call logical instructions in the memory 630 to execute a server power-on method, which includes: constructing the server power-on sequence using an ant colony algorithm based on the pheromone content of the time difference between the power peak times of each server in the server cluster to be powered on. Based on the total time for each server to reach its peak power consumption during the staggered power-up process corresponding to the power-up sequence of each server, update the pheromone content of the time difference between the power-up times of each server during the power-up process of the server power-up sequence. Based on the updated pheromone levels, the ant colony algorithm is used to continue constructing the server power-on sequence until the first preset condition is met, thus obtaining the target server power-on sequence.

[0101] Furthermore, the logical instructions in the aforementioned memory 630 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0102] On the other hand, the present invention also provides a computer program product, the computer program product including a computer program, the computer program being able to be stored on a non-transitory computer-readable storage medium, the computer program being executed by a processor, the computer being able to execute the server power-on method provided by the above methods, the method including: constructing the server power-on sequence by using an ant colony algorithm based on the pheromone content of the time difference between the power peaks reached by each server in the server cluster to be powered on during the power-on process; Based on the total time for each server to reach its peak power consumption during the staggered power-up process corresponding to the power-up sequence of each server, update the pheromone content of the time difference between the power-up times of each server during the power-up process of the server power-up sequence. Based on the updated pheromone levels, the ant colony algorithm is used to continue constructing the server power-on sequence until the first preset condition is met, thus obtaining the target server power-on sequence.

[0103] In another aspect, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to perform the server power-on method provided by the above methods, the method comprising: constructing the server power-on sequence by means of an ant colony algorithm based on the pheromone content of the time difference between the power peaks reached by each server in the power-on process of the server cluster to be powered on; Based on the total time for each server to reach its peak power consumption during the staggered power-up process corresponding to the power-up sequence of each server, update the pheromone content of the time difference between the power-up times of each server during the power-up process of the server power-up sequence. Based on the updated pheromone levels, the ant colony algorithm is used to continue constructing the server power-on sequence until the first preset condition is met, thus obtaining the target server power-on sequence.

[0104] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0105] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0106] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for powering on a server, characterized in that, include: Based on the pheromone content of the time difference between the time difference when each server in the server cluster reaches its power peak during the power-on process, the power-on sequence of the servers is constructed using the ant colony algorithm. Based on the total time for each server to reach its peak power during the staggered power-up process, update the pheromone content of the time difference between the power-up times of each server in the power-up sequence. Based on the updated pheromone levels, the ant colony algorithm is used to continue constructing the server power-on sequence until the first preset condition is met, thus obtaining the target server power-on sequence. The time difference is obtained by subtracting the time it takes for the servers to reach their power peak during startup. The time ti for each server to reach its power peak during startup is ti = to + tm, where to is the historical startup correction time and tm is the startup correction time for the same model. If the server has previously performed a power-on operation, calculate the average time it takes for the server to reach its peak power after historical power-on, and perform time correction to avoid time overlap with other servers. If time overlap is found, add a random time accurate to milliseconds and record it to obtain the historical power-on correction time to, and set the power-on correction time tm of the same model to 0. If the server has not performed a power-on operation, set the historical power-on correction time to 0, query the average time for the same model to reach peak power during power-on, check and correct the time to avoid duplication with the calculated time. If duplication is found, add a random time accurate to milliseconds and record it to obtain the power-on correction time tm for the same model.

2. The server power-on method according to claim 1, characterized in that, The step of constructing the server power-on sequence using the ant colony algorithm, based on the initial pheromone content of the time difference between the time when each server in the server cluster reaches its power-on peak, includes: If the number of servers in the constructed server power-on sequence is less than a first preset threshold, the ants in the ant colony algorithm calculate the state transition probability based on heuristic information and pheromone quantity, and select the next server to be powered on based on the state transition probability. The next server to be powered on is written into the server power-on sequence in order, and the relationship between the number of servers in the server power-on sequence and the first preset threshold is further compared. If the number of servers in the server power-on sequence is still less than the first preset threshold, the next server to be powered on will continue to be selected according to the ant colony algorithm. If the number of servers in the server power-on sequence is greater than or equal to the first preset threshold, each ant in the ant colony algorithm completes the traversal of the servers in the cluster to be powered on, thus completing the construction of the server power-on sequence.

3. The server power-on method according to claim 1, characterized in that, The step of updating the pheromone content of the time difference between the power-up times of each server during the power-up process, based on the total time for each server to reach its peak power according to the power-up sequence, includes: The first pheromone increment is determined based on the total time it takes for the optimal server to reach its power peak during off-peak periods. The total pheromone increment is determined based on the first pheromone increment and the second pheromone increment left by each ant during the construction of the power-on sequence on the server. Based on the total pheromone increment, update the pheromone amount of each server in the power-on sequence to the time difference between reaching peak power during the power-on process, and obtain the updated pheromone amount.

4. The server power-on method according to claim 2, characterized in that, The step of continuing to construct the server power-on sequence using the ant colony algorithm based on the updated pheromone levels until the first preset condition is met, and obtaining the target server power-on sequence, includes: In the ant colony algorithm, the ants select the next server to be powered on based on the updated pheromone levels and heuristic information, in order to establish the server power-on sequence. Based on the reconstructed server power-on sequence, continue to update the pheromone content of each server in the power-on sequence during the power-on process to reach the power peak, until the first preset condition is met, then stop updating, and obtain multiple sets of server power-on sequences; Based on the total time it takes for the servers to reach their peak power during off-peak hours, determine the target power-on sequence for the servers in multiple power-on sequences.

5. The server power-on method according to claim 2, characterized in that, The method for calculating the state transition probability specifically includes: ； In the ant colony algorithm, the ant k (k=1, 2, ..., n) uses a tabu list during its movement. finish k To record currently powered-on servers, using To record the servers that are allowed to boot up next. ={S-finish k ｝ ; yij(x) The amount of information remaining at each time difference. This provides heuristic information for the time differences.

6. The server power-on method according to claim 3, characterized in that, The method for calculating the first pheromone increment includes: ； in, Indicates the known optimal power-on sequence of the server. The total time it takes for all servers to reach peak power consumption during off-peak hours. L It is a positive number.

7. The server power-on method according to claim 6, characterized in that, The method for calculating the total pheromone increment includes: ； ； ； in, Indicates the first k The total time it takes for all servers to reach their peak power during this server startup process (using only one ant). This represents the time difference in this iteration ( i , j The pheromone increment on the surface, i.e., the second pheromone increment. h It is a coefficient. （1-h） express x arrive x+d The pheromone evaporation rate at any given time.

8. A server power-on device, characterized in that, include: The module is used to construct the power-on sequence of servers based on the pheromone content of the time difference between the time difference when each server in the power-on cluster reaches its power peak during the power-on process, using the ant colony algorithm. The update module is used to update the pheromone content of the time difference between the power peak times of each server during the power-on process in the power-on sequence of each server, based on the total time for the servers to reach their peak power during the staggered power-on process. The power-on module is used to continue constructing the server power-on sequence based on the updated pheromone content using the ant colony algorithm until the first preset condition is met, thereby obtaining the target server power-on sequence. The time difference is obtained by subtracting the time it takes for the servers to reach their power peak during startup. The time ti for each server to reach its power peak during startup is ti = to + tm, where to is the historical startup correction time and tm is the startup correction time for the same model. If the server has previously performed a power-on operation, calculate the average time it takes for the server to reach its peak power after historical power-on, and perform time correction to avoid time overlap with other servers. If time overlap is found, add a random time accurate to milliseconds and record it to obtain the historical power-on correction time to, and set the power-on correction time tm of the same model to 0. If the server has not performed a power-on operation, set the historical power-on correction time to 0, query the average time for the same model to reach peak power during power-on, check and correct the time to avoid duplication with the calculated time. If duplication is found, add a random time accurate to milliseconds and record it to obtain the power-on correction time tm for the same model.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the server power-on method as described in any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the server power-on method as described in any one of claims 1 to 7.

Images