A data center-based information storage method and system

Abstract

The present application relates to a data center-based information storage method and system, and to the field of data storage technology. The method includes obtaining the current load value of each server; defining a high-load server and an idle server according to the current load value; determining the demand release load in the high-load server according to the current load value and the upper limit load value, and determining the remaining idle load in the idle server according to the current load value and the upper limit load value; judging whether the demand release load is greater than the sum of the remaining idle loads; if the demand release load is not greater than the sum of the remaining idle loads, determining the demand movement information in the high-load server according to the demand release load, and determining the target server in each idle server according to each demand movement information, and controlling the demand movement information to move to the target server. The present application has the effect of improving the overall operation effect of the server.

Description

Technical Field

[0001] The present application relates to the field of data storage technology, and in particular to an information storage method and system based on a data center. Background Art

[0002] A data center is a global collaborative network of specific devices used to transmit, accelerate, display, calculate, and store data information on the Internet infrastructure. The role of a data center is to centralize data and store, calculate, and exchange it through servers to meet various business needs.

[0003] In the related art, servers can be divided into low-load servers, medium-load servers and high-load servers according to how much data they store. When data information is collected by a data center and needs to be stored in a server, medium-load servers have better storage space and information processing performance, so the data information will be stored in the medium-load servers first. Only when there are no medium-load servers will low-load servers and high-load servers be considered.

[0004] With respect to the above-mentioned related technologies, the inventor believes that when the data information in the server is called, the data information in the server will disappear. When the information in the high-load server has not been called for a long time, the server will be in a high-load state for a long time. At this time, the server will not only experience a decline in service quality, but in serious cases, the server may even crash, resulting in poor overall server performance. There is still room for improvement. Summary of the invention

[0005] In order to improve the overall operation effect of the server, the present application provides an information storage method and system based on a data center.

[0006] In a first aspect, the present application provides an information storage method based on a data center, which adopts the following technical solution:

[0007] An information storage method based on a data center, comprising:

[0008] Get the current load value of each server during off-peak hours;

[0009] The servers whose current load values ​​are greater than the preset upper load value are defined as high-load servers, and the remaining servers are defined as idle servers;

[0010] Performing a difference calculation based on the current load value and the upper limit load value in the high-load server to determine the required load release, and performing a difference calculation based on the current load value and the upper limit load value in the idle server to determine the remaining idle load;

[0011] Determine whether the required release load is greater than the sum of the remaining idle loads;

[0012] If the required release load is greater than the sum of the remaining idle loads, each server is controlled to maintain its original state;

[0013] If the demand release load is not greater than the sum of the remaining idle loads, the demand movement information is determined in the high-load server according to the demand release load, and the target server is determined in each idle server according to each demand movement information, and the demand movement information is controlled to move to the target server.

[0014] Optionally, the step of determining the required movement information according to the required load release in the high-load server includes:

[0015] Obtain the information monomer load of each data information in the high-load server;

[0016] In a high-load server, a random combination is performed according to any number of data information to determine a load information combination, and a summation calculation is performed according to the information monomer load of each data information in the load information combination to determine the information combination load;

[0017] A load information combination whose information combination load is not less than the demand release load is defined as an alternative information combination;

[0018] Determine the remaining idle load with the largest value in the idle servers according to a preset sorting rule, define the remaining idle load as the maximum load, and perform calculation according to the maximum load and a preset weight ratio to determine the single maximum load;

[0019] The alternative information combinations whose information monomer loads are not greater than the monomer maximum load are defined as available information combinations, and a unique preferred combination is determined among the available information combinations, and the data information in the preferred combination is determined as the required mobile information.

[0020] Optionally, the step of determining a unique preferred combination among available information combinations includes:

[0021] Counting the data information in the available information combinations to determine the number of combined information;

[0022] The excess release load is determined by performing a difference calculation based on the information combined load and the required release load;

[0023] Calculation is performed based on the amount of combined information and the excess release load to determine the preferred parameters;

[0024] The preferred parameter with the largest value is determined according to the sorting rule, and the available information combination corresponding to the preferred parameter is determined as the preferred combination.

[0025] Optionally, the step of determining a target server from each idle server according to each demand movement information includes:

[0026] According to the information monomer load of each demand mobile information, the mobile information sequence is determined by sorting from large to small, and the demand mobile information with the largest information monomer load is defined as the head-end mobile information;

[0027] Determine the remaining idle load with the largest value according to the sorting rule, and define the idle server corresponding to the remaining idle load as the head-end server;

[0028] The first-end mobile information is matched with the first-end server to determine the first-end server as the target server of the first-end mobile information, and the first-end mobile information is removed from the mobile information sequence and redefined, and the first-end server is re-determined and bound according to the new remaining idle load, until each required mobile information corresponds to a target server, and the first target solution is determined according to the current matching relationship between the first-end mobile information and the first-end server.

[0029] Optionally, after the first target solution is determined, the data center-based information storage method further includes:

[0030] After the first head-end mobile information is matched with the head-end server, a difference calculation is performed based on the information monomer load of the head-end mobile information and the remaining idle load of the head-end server to determine the theoretical remaining load;

[0031] Determine the demand mobility information in which the information unit load is not greater than the theoretical residual load and the difference with the theoretical residual load is the smallest in the demand mobility information, and match the demand mobility information with the current head-end server until the corresponding demand mobility information cannot be determined and the head-end server is redefined;

[0032] Determine a second target solution according to the matching relationship between the current head-end mobile information and the head-end server;

[0033] Obtaining the available remaining load of the idle server after each demand movement information is moved in the first target solution and the second target solution;

[0034] defining a server whose available remaining load is in a preset medium load interval as a medium load server, and defining a server whose available remaining load is in a preset low load interval as a low load server;

[0035] Counting against the medium load servers to determine the medium load number and counting against the low load servers to determine the low load number;

[0036] A server whose available remaining load is in a preset waste interval is defined as a waste server, and a difference calculation is performed based on the available remaining load of the waste server and the lower limit of the waste interval to determine the resource waste load;

[0037] Calculate the solution selection coefficient based on the number of medium loads, the number of low loads, and the resource waste load;

[0038] A scheme with a larger scheme selection coefficient is selected from the first target scheme and the second target scheme as the target determination scheme, and a target server corresponding to each required mobile information is determined according to the target determination scheme.

[0039] Optionally, if the required release load is greater than the sum of the remaining idle loads, the data center-based information storage method further includes:

[0040] Obtain the possible call probability of each data information in the server;

[0041] The data information whose possible call probability is greater than the preset upper limit probability is defined as mobile information, and the remaining data information is defined as fixed information;

[0042] Calculation is performed based on the demand release load and the preset amplification factor to determine the probabilistic demand load;

[0043] In the high-load server, the overall mobile load is determined by summing up the mobile information;

[0044] Determine whether the overall mobile load is less than the probabilistic demand load;

[0045] If the overall mobile load is not less than the probabilistic demand load, the high-load server is controlled to maintain its original state;

[0046] If the overall moving load is less than the probabilistic demand load, the demand differential load is determined by performing a difference calculation based on the overall moving load and the probabilistic demand load;

[0047] Selecting from the fixed information of the high-load server according to the possible call probability from small to large to determine a fixed mobile combination with a load just greater than the demand difference;

[0048] Calculate the difference of information monomer load based on the fixed information in the fixed-mobile combination and the mobile information in the remaining servers to determine the difference comparison load;

[0049] The difference comparison load with the smallest value is determined according to the sorting rule, and the movement information relative to the difference comparison load is defined as the associated information of the fixed information, and the associated information is controlled to exchange positions with the corresponding movement information.

[0050] Optionally, a step of determining a possible call probability is also included, the step comprising:

[0051] Establishing a detection interval on a preset time axis with the current time point as the end point and a width as the preset detection duration;

[0052] Get the usage data information called within the detection interval;

[0053] Determine usage data attributes and usage storage time based on usage data information;

[0054] Count the data information corresponding to each attribute of the used data to determine the number of attribute information, and calculate the proportion of individual attributes based on the number of each attribute information;

[0055] Obtain the analysis data attributes and analysis storage time of each data information in each server;

[0056] Determine the selection parameters based on the analysis data attributes and the corresponding monomer attribute proportions;

[0057] Delineating a similar interval based on the storage time and the preset similar duration, and determining the correction parameter based on analyzing the relative relationship between the storage time and the similar interval;

[0058] Calculations are performed based on the selection parameters and the correction parameters to determine the possible call probability.

[0059] In a second aspect, the present application provides an information storage system based on a data center, which adopts the following technical solution:

[0060] An information storage system based on a data center, comprising:

[0061] The acquisition module is used to obtain the current load value of each server during non-peak hours;

[0062] A processing module, connected to the acquisition module and the judgment module, for storing and processing information;

[0063] A judgment module, connected with the acquisition module and the processing module, for judging the information;

[0064] The processing module defines a server whose current load value is greater than a preset upper load value as a high-load server, and defines the remaining servers as idle servers;

[0065] The processing module performs a difference calculation based on the current load value and the upper limit load value in the high-load server to determine the required release load, and performs a difference calculation based on the current load value and the upper limit load value in the idle server to determine the remaining idle load;

[0066] The judging module judges whether the required release load is greater than the sum of the remaining idle loads;

[0067] If the judgment module determines that the required release load is greater than the sum of the remaining idle loads, the processing module controls each server to maintain the original state;

[0068] If the judgment module determines that the demand release load is not greater than the sum of the remaining idle loads, the processing module determines the demand movement information in the high-load server according to the demand release load, and determines the target server in each idle server according to each demand movement information, and controls the demand movement information to move to the target server.

[0069] In summary, the present application includes at least one of the following beneficial technical effects:

[0070] Identify high-load servers during off-peak hours and control the transfer of data information in the high-load servers to achieve better overall operation of the corresponding servers;

[0071] The data information transfer plan can be selected according to the status of each server after the transfer to improve the overall operation effect of each server;

[0072] In the case where the redundant data information in the high-load server cannot be transferred, the data information that may be called can be transferred to the high-load server to reduce the storage information of the high-load server, thereby maximizing the overall operation effect of the server. BRIEF DESCRIPTION OF THE DRAWINGS

[0073] Figure 1 It is a flow chart of the information storage method based on data center.

[0074] Figure 2 It is a flow chart of a method for determining demand movement information.

[0075] Figure 3 is a flow chart of a method for determining a preferred combination.

[0076] Figure 4 is a flow chart of a target server determination method.

[0077] Figure 5 It is a flow chart of the information transfer scheme selection method.

[0078] Figure 6 It is a flow chart of the method for replacing the position of data information.

[0079] Figure 7 is a flow chart of a possible invocation of a probability determination method.

[0080] Figure 8 It is a module flow chart of the information storage method based on data center. DETAILED DESCRIPTION

[0081] In order to make the purpose, technical solutions and advantages of this application more clear, the following Figure 1-8 It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0082] The embodiments of the present application are further described in detail below in conjunction with the drawings in the specification.

[0083] The embodiment of the present application discloses an information storage method based on a data center. During non-peak hours, the data information storage status of each server can be analyzed. When there is a high-load server, the storage capacity of the excess data information can be determined, and it can be analyzed whether the external server can accept this part of the data information. When it can be accepted, appropriate data information can be determined to be transferred. When it cannot be accepted, data information with a higher possibility of being called can be controlled to be exchanged with the information in the high-load server, thereby achieving a reduction in the amount of data information in the high-load server, reducing the situation where the server is under high load for a long time, and improving the overall operation effect of the server.

[0084] Reference Figure 1 The method flow of the information storage method based on the data center includes the following steps:

[0085] Step S100: Obtain the current load value of each server during a non-peak period.

[0086] The current load value is the overall capacity value of the data information stored in the server itself during non-peak hours, that is, the load value.

[0087] Step S101: define a server whose current load value is greater than a preset upper limit load value as a high-load server, and define the remaining servers as idle servers.

[0088] The upper limit load value is the minimum load value set by the staff when the service quality will decrease due to a large server load. Defining high-load servers and idle servers can determine the data information storage status of different servers, which is convenient for subsequent analysis.

[0089] Step S102: performing a difference calculation based on the current load value and the upper limit load value in the high-load server to determine the required load release, and performing a difference calculation based on the current load value and the upper limit load value in the idle server to determine the remaining idle load.

[0090] The demand release load is the minimum load value that needs to be reduced when a high-load server is converted into an idle server, which is determined by subtracting the upper limit load value from the current load value of the high-load server; the remaining idle load is the maximum data information load value that the idle server can accommodate without becoming a high-load server, which is determined by subtracting the current load value of the idle server from the upper limit load value.

[0091] Step S103: Determine whether the required release load is greater than the sum of the remaining idle loads.

[0092] The purpose of the judgment is to find out whether the redundant data information in the high-load server can be transferred to convert the high-load server into an idle server.

[0093] Step S1031: If the required release load is greater than the sum of the remaining idle loads, control each server to maintain the original state.

[0094] When the required release load is greater than the sum of the remaining idle loads, it means that the high-load server cannot achieve a better operating effect by transferring the redundant data information. At this time, the original state can be maintained normally.

[0095] Step S1032: If the demand release load is not greater than the sum of the remaining idle loads, determine the demand movement information in the high-load server according to the demand release load, determine the target server in each idle server according to each demand movement information, and control the demand movement information to move to the target server.

[0096] When the demand release load is not greater than the sum of the remaining idle loads, it means that the high-load server can have a better operating effect by transferring the redundant data information. At this time, information control is required; the demand movement information is the data information that needs to be transferred in the high-load server. In one embodiment, the data information in the high-load server can be sorted from front to back according to the storage time of each data information, and the data information can be determined as the demand movement information in sequence according to the sorting until the load of the determined demand movement information is not less than the demand release load. In another embodiment, the method in steps S200-S204 can be used to determine the demand movement information; the target server is each The destination to which the demand mobile information is to be directed after it is moved from the high-load server can be determined in one embodiment by numbering each idle server and comparing the priority demand mobile information with the corresponding numbered idle servers in turn. When the placement requirements can be met, the corresponding idle server is determined as the target server of the demand mobile information. In another embodiment, the method of steps S400-S402 can be used for determination; by controlling the demand mobile information to move to the target server, the load of the high-load server can be lowered, so as to reduce the situation where the server is in a high-load state for a long time and causes a decrease in service quality, thereby improving the overall operation effect of the server.

[0097] Reference Figure 2 The steps of determining the required mobile information according to the required load release in the high-load server include:

[0098] Step S200: Obtaining the information monomer load of each data information in the high-load server.

[0099] The information monomer load is the space load size occupied by a single data information in a high-load server.

[0100] Step S201: randomly combining any number of data information in a high-load server to determine a load information combination, and summing up the information monomer loads of each data information in the load information combination to determine the information combination load.

[0101] The load information combination is a data information combination formed by combining different data information. For example, if there are data information A, B, and C, the corresponding load information combinations include A, B, C, AB, AC, BC, and ABC. The information combination load is the sum of the loads of the data information in the load information combination.

[0102] Step S202: define a load information combination whose information combination load is not less than the required release load as a candidate information combination.

[0103] When the information combination load is not less than the required release load, it means that the data information in the current load information combination can meet the requirements of the high-load server after being fully released. At this time, the corresponding load information combination is defined as an alternative information combination for identification, which is convenient for subsequent analysis.

[0104] Step S203: Determine the remaining idle load with the largest value in the idle servers according to a preset sorting rule, define the remaining idle load as the maximum load, and perform calculation according to the maximum load and a preset weight ratio to determine the single maximum load.

[0105] The sorting rule is a method set by the staff to sort the size of values, such as the bubble method. The sorting rule can be used to determine the idle server that can bear the most groups. At this time, the maximum load is defined to distinguish different remaining idle loads; the weight ratio is a fixed value set by the staff, and its value range is 0-1. The load value can be limited by setting the weight ratio. The maximum load of a single unit is the maximum information single unit load allowed in the combination to reduce the situation where the server cannot subsequently accept data information. For example, the maximum load is 100, and the corresponding weight ratio is 0.5, then the corresponding maximum load of a single unit is 50, which means that the load of the data information to be moved does not exceed 50, thereby facilitating the transfer of the required mobile information.

[0106] Step S204: define candidate information combinations whose information monomer loads are not greater than the monomer maximum load as available information combinations, determine a unique preferred combination among the available information combinations, and determine the data information in the preferred combination as the required mobile information.

[0107] When the information monomer load is not greater than the monomer maximum load, it means that each data information in the alternative information combination has a greater probability of being transferred well. At this time, the available information combination is defined to achieve the distinction between different alternative information combinations; the preferred combination is the combination that requires data information transfer. In one embodiment, a combination can be determined as the preferred combination in the available information combination by random selection. In another embodiment, the method of steps S300-S303 can also be used to determine the preferred combination.

[0108] Reference Figure 3 The step of determining a unique preferred combination from the available information combinations includes:

[0109] Step S300: Count the available information combinations according to the data information to determine the number of combination information.

[0110] The combined information quantity is the total quantity of data information contained in a single available information combination.

[0111] Step S301: Calculate the difference between the information combined load and the required released load to determine the redundant released load.

[0112] The excess release load is the excess value of the determined combined load compared to the load required to be released, and is determined by subtracting the required release load from the information combined load.

[0113] Step S302: Calculate and determine the optimal parameters according to the amount of combined information and the redundant released load.

[0114] The optimal parameter is a parameter that reflects whether the selected combination is reasonable. The larger the value, the more reasonable the determined combination is. The specific calculation formula is: , where δ is the preferred parameter, α and β are fixed parameters set by the staff, M is the number of combined information, and μ is the redundant release load.

[0115] Step S303: determining the preferred parameter with the largest value according to the sorting rule, and determining the available information combination corresponding to the preferred parameter as the preferred combination.

[0116] The preferred parameter with the largest value can be determined through the sorting rules, that is, the available information combination corresponding to the preferred parameter is the most suitable as a whole. At this time, the available information combination is defined as the preferred combination to facilitate subsequent processing.

[0117] Reference Figure 4 The step of determining the target server in each idle server according to each demand mobile information includes:

[0118] Step S400: sorting the information monomer loads of each demand mobile information from large to small to determine the order of the mobile information, and defining the demand mobile information with the largest information monomer load as the head-end mobile information.

[0119] The mobile information sequence is the sequence of the demand mobile information sorted from large to small according to the information monomer load, and the head mobile information is the demand mobile information at the front end of the mobile information sequence.

[0120] Step S401: Determine the remaining idle load with the largest value according to the sorting rule, and define the idle server corresponding to the remaining idle load as the head-end server.

[0121] The sorting rules can be used to determine the remaining idle load with the largest value, that is, the data information load that can be borne by the idle server that is most capable of receiving data information. At this time, it is defined as the head-end server for identification to facilitate subsequent analysis.

[0122] Step S402: Match the head-end mobile information with the head-end server to determine the head-end server as the target server of the head-end mobile information, and remove the head-end mobile information from the mobile information sequence and redefine the head-end mobile information, and redetermine and bind the head-end server based on the new remaining idle load, until each required mobile information corresponds to a target server, and determine the first target solution based on the current matching relationship between the head-end mobile information and the head-end server.

[0123] For example, if there is a mobile information sequence of ABCD, and the determined head-end server is X, then the target server of A is X. After the target server of A is determined, the target server of B is determined. At this time, because the X server has received the data information A, there will be a load change. At this time, it is necessary to continue to determine the head-end server. When the server that continues to be determined is still X, the target server of B is also X. The target servers of the required mobile information can be determined in turn through the above method, and data information with a large load can be processed preferentially, so as to facilitate the determination of the target servers of each required mobile information; at this time, each required mobile information corresponds to a target server, so that a first target plan can be formed, and the system can control the movement of the required mobile information according to the first target plan.

[0124] Reference Figure 5 After the first target solution is determined, the information storage method based on the data center also includes:

[0125] Step S500: after the first head-end mobile information is matched with the head-end server, a difference calculation is performed based on the information monomer load of the head-end mobile information and the remaining idle load of the head-end server to determine the theoretical remaining load.

[0126] The theoretical remaining load is the load remaining for accommodating data information after the head-end server receives the head-end mobile information, and is determined by subtracting the information monomer load of the head-end mobile information from the remaining idle load of the head-end server.

[0127] Step S501: Determine the demand mobility information in which the information unit load is not greater than the theoretical remaining load and the difference with the theoretical remaining load is the smallest, and match the demand mobility information with the current head-end server until the corresponding demand mobility information cannot be determined and the head-end server is redefined.

[0128] For example, there is a mobile information sequence of ABCD, and the corresponding information monomer loads are 5, 4, 2, and 1 respectively. The determined head-end server is X, and the corresponding remaining idle load is 7. After the target server determined by A is X, the remaining idle load of X is 2. At this time, the demand mobile information C can be determined through load analysis, that is, the target server determined by C is also X. At this time, the remaining idle load of X is 0, and no demand mobile information can be accommodated in X. At this time, the definition of the X head-end server is abandoned, and the head-end server is re-determined from the remaining idle servers; through the above method, the server determined as the head-end server can be filled to the upper limit value of the medium load as much as possible, thereby improving the overall utilization of the server.

[0129] Step S502: Determine a second target solution according to the matching relationship between the current head-end mobile information and the head-end server.

[0130] The second target solution is a solution formed after each demand mobile information corresponds to a target server. The first target solution and the second target solution are solutions for determining the target server of the demand mobile information according to two different rules.

[0131] Step S503: Obtain the available remaining load of the idle server after each demand movement information is moved in the first target solution and the second target solution.

[0132] The available remaining load is the load value that each idle server can still bear the data information under the medium load condition after the demand movement information is moved to the corresponding target server.

[0133] Step S504: defining the servers whose available remaining loads are in a preset medium load interval as medium load servers, and defining the servers whose available remaining loads are in a preset low load interval as low load servers.

[0134] The medium load interval is the load interval set by the staff to identify the system under medium load conditions. For example, if the server load is 55% and the corresponding upper load limit is 80%, the available remaining load is 25%. The set medium load interval should include 25%. Similarly, the low load interval is set only to distinguish the server load conditions. The lower limit of the low load interval is the upper limit of the medium load interval. By defining medium load servers and low load servers, the specific conditions of each idle server can be known, which is convenient for subsequent analysis.

[0135] Step S505: Count the medium-load servers to determine the number of medium-load servers, and count the low-load servers to determine the number of low-load servers.

[0136] The medium load number is the total number of medium load servers currently owned, and the low load number is the total number of low load servers currently owned.

[0137] Step S506: define a server whose available remaining load is in a preset waste interval as a waste server, and perform a difference calculation based on the available remaining load of the waste server and the lower limit of the waste interval to determine the resource waste load.

[0138] The waste interval is the load value set by the staff when it is determined that the available remaining load cannot handle most of the data information. For example, if most of the data information is above 1%, the set waste interval is 0-1%. At this time, it is defined as a waste server for subsequent analysis; the resource waste load is the load situation when resource waste occurs. Taking the above example, the corresponding resource waste load is 1%.

[0139] Step S507: Calculate according to the number of medium loads, the number of low loads, and the resource waste load to determine the solution selection coefficient.

[0140] The scheme selection coefficient is a parameter that reflects the rationality of the scheme selection. The larger the coefficient is, the better the selected scheme is. The specific calculation formula is: , where σ is the scheme selection coefficient, γ is the fixed coefficient set by the staff, D is the number of low loads, Z is the number of medium loads, and τ is the sum of resource waste loads.

[0141] Step S508: selecting a solution with a larger solution selection coefficient from the first target solution and the second target solution as the target determination solution, and determining a target server corresponding to each required mobile information according to the target determination solution.

[0142] A scheme with a larger selection coefficient is selected as a target determination scheme, so as to determine the target server of each demand mobile information according to the target determination scheme, thereby facilitating the transfer of the demand mobile information.

[0143] Reference Figure 6 If the required release load is greater than the sum of the remaining idle loads, the data center-based information storage method further includes:

[0144] Step S600: Obtain the possible calling probability of each data information in the server.

[0145] The possible call probability is the probability value that the data information will be called in a short time. The larger the probability value is, the more likely the data information is to disappear in the server. The specific determination method is described in steps S700 to S707.

[0146] Step S601: define data information whose possible call probability is greater than a preset upper limit probability as mobile information, and define the rest of the data information as fixed information.

[0147] The upper limit probability is the minimum probability value set by the staff that the recognized data information will be called within a short period of time. Mobile information and fixed information are defined to distinguish different data information for subsequent analysis.

[0148] Step S602: Calculate according to the demand release load and the preset amplification factor to determine the probabilistic demand load.

[0149] The amplification factor is a fixed value greater than 1 set by the staff according to actual conditions, such as 2. The probabilistic demand load is the minimum load of the data information that may be called, which is determined by multiplying the demand release load by the amplification factor.

[0150] Step S603: performing a summation calculation based on the movement information in the high-load server to determine the overall movement load.

[0151] The overall mobile load is the total load value of the data information determined as mobile information in the high-load server, and is determined by summing up the information monomer loads of each mobile information.

[0152] Step S604: Determine whether the overall mobile load is less than the probabilistic required load.

[0153] The purpose of the judgment is to find out whether the amount of mobile information in the high-load server satisfies the requirement that the server has a high probability of not being under high load.

[0154] Step S6041: If the overall mobile load is not less than the probabilistic demand load, the high-load server is controlled to maintain the original state.

[0155] When the overall mobile load is not less than the probabilistic demand load, it means that the mobile information in the high-load server meets the load requirements. At this time, no additional processing is required, and the server load can be reduced by simply calling the subsequent mobile information.

[0156] Step S6042: If the overall mobile load is less than the probabilistic demand load, a difference calculation is performed based on the overall mobile load and the probabilistic demand load to determine the demand difference load.

[0157] When the overall mobile load is less than the probabilistic demand load, it means that the amount of mobile information in the high-load server cannot meet the requirements and further analysis is needed; the demand difference load is the load of mobile information missing in the high-load server, which is determined by subtracting the overall mobile load from the probabilistic demand load.

[0158] Step S605: Select from the fixed information of the high-load server according to the possible call probability from small to large to determine a fixed mobile combination with a load just greater than the demand difference.

[0159] The fixed mobile combination is a combination in which the load value of the data information in the combination obtained by selecting the fixed information from small to large according to the possible calling probability is just greater than the demand difference load, that is, the combination of data information with the lowest probability of being called in the high-load server and occupying the corresponding load.

[0160] Step S606: Calculate the information monomer load difference based on the fixed information in the fixed-mobile combination and the mobile information in the remaining servers to determine the difference comparison load.

[0161] The difference comparison load is the difference between the information monomer load of a single fixed information in the fixed-mobile combination and the information monomer load of the mobile information in the remaining idle servers, and the difference is an absolute value.

[0162] Step S607: Determine the difference comparison load with the smallest value according to the sorting rule, define the movement information relative to the difference comparison load as the associated information of the fixed information, and control the associated information to exchange positions with the corresponding movement information.

[0163] The sorting rules can be used to determine the comparison load with the smallest difference value, that is, the mobile information determined at this time is closest to the fixed information load in the combination. At this time, it is determined as related information and the position is controlled to be swapped, so that when the load of each server does not change significantly, the data information that is easily called can be in the high-load server, thereby enabling the high-load server to improve service performance after the data information stored in itself is called.

[0164] Reference Figure 7 , and also includes a possible call probability determination step, which includes:

[0165] Step S700: establishing a detection interval on a preset time axis with the current time point as the end point and a width of a preset detection duration.

[0166] The time axis is a coordinate axis formed by the combination of various time points. The detection time is the time set by the staff to analyze the recent data information call situation of the data center. The detection interval is established to facilitate the acquisition of data.

[0167] Step S701: Acquire usage data information called within the detection interval.

[0168] Usage data information refers to data information originally stored in the server that disappears after being called.

[0169] Step S702: Determine usage data attributes and usage storage time according to the usage data information.

[0170] The usage data attribute is the attribute of the usage data information, which may include the label of the information, such as R&D information, preliminary preparation information, etc. The specific attribute distinction scheme is determined in advance by the staff based on the server's own situation; the usage storage time is the time point when the usage data information is stored in the server.

[0171] Step S703: Count the data information corresponding to each usage data attribute to determine the quantity of attribute information, and calculate the proportion of each attribute information according to the quantity of each attribute information.

[0172] The number of attribute information is the number of data information corresponding to each data attribute used, and the single attribute proportion is the ratio of the number of a single attribute to the number of all attributes, which is determined by the number of attribute information being in the sum of the number of attribute information.

[0173] Step S704: Obtain the analysis data attributes and analysis storage time of each data information in each server.

[0174] The analysis data attributes are the attributes of a single data information, and the analysis storage time is the time point when the data information in the server is stored in the server.

[0175] Step S705: Determine selection parameters according to the analyzed data attributes and the corresponding monomer attribute proportions.

[0176] The selection parameter is a parameter value that reflects whether the data information will be called from the attribute level. The larger the parameter, the higher the possibility of being called. Different analysis data information corresponds to different proportions of individual attributes, and the corresponding selection parameters are also different. The relationship between the two is determined by the staff through multiple experiments in advance. It is only necessary to ensure that the larger the proportion of individual attributes corresponding to the analysis data attributes, the larger the selection parameter.

[0177] Step S706: Delimiting a similar interval according to the storage time and the preset similar duration, and determining a correction parameter according to analyzing the relative relationship between the storage time and the similar interval.

[0178] The similar duration is the error duration allowed when the recognized data information is stored simultaneously, which is set by the staff. The similar interval is an interval with the storage time as the center point and the width as the similar duration. When the analysis storage time is within the similar interval, it means that the data information and the called information are from the same batch of stored information. At this time, the possibility of the information being called is greater than that of information not from the same batch. Therefore, different correction parameters can be obtained according to whether the analysis storage time is within the similar interval. The correction parameter is the parameter value of whether the data information will be called from the data storage level. The parameter value is only 2 and 1. 2 represents that the analysis storage time is within the similar interval, and 1 represents that the analysis storage time is not within the similar interval.

[0179] Step S707: Calculate based on the selected parameters and the modified parameters to determine the possible call probability.

[0180] The probability of possible calls can be obtained by multiplying the selection parameter by the correction parameter.

[0181] Reference Figure 8 Based on the same inventive concept, an embodiment of the present invention provides an information storage system based on a data center, including:

[0182] The acquisition module is used to obtain the current load value of each server during non-peak hours;

[0183] A processing module, connected to the acquisition module and the judgment module, for storing and processing information;

[0184] A judgment module, connected with the acquisition module and the processing module, for judging the information;

[0185] The processing module defines a server whose current load value is greater than a preset upper load value as a high-load server, and defines the remaining servers as idle servers;

[0186] The processing module performs a difference calculation based on the current load value and the upper limit load value in the high-load server to determine the required release load, and performs a difference calculation based on the current load value and the upper limit load value in the idle server to determine the remaining idle load;

[0187] The judging module judges whether the required release load is greater than the sum of the remaining idle loads;

[0188] If the judgment module determines that the required release load is greater than the sum of the remaining idle loads, the processing module controls each server to maintain the original state;

[0189] If the judging module judges that the demand release load is not greater than the sum of the remaining idle loads, the processing module determines the demand movement information in the high-load server according to the demand release load, determines the target server in each idle server according to each demand movement information, and controls the demand movement information to move to the target server;

[0190] A demand movement information determination module is used to determine more appropriate demand movement information to control the transfer of corresponding data information;

[0191] A preferred combination determination module, used to determine a unique preferred combination from multiple available combinations;

[0192] A target server determination module, used to determine the target server corresponding to each required mobile information;

[0193] A transfer scheme selection module is used to determine a selection scheme for a target server corresponding to the required mobile information;

[0194] The data information exchange module exchanges the location of data information in different servers in case that data information cannot be completely transferred;

[0195] The possible call probability determination module is used to determine the possible call probability of each data information.

[0196] Those skilled in the art can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional modules is used as an example. In actual applications, the above-mentioned functions can be assigned to different functional modules as needed, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. The specific working process of the system, device and unit described above can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.

Claims

1. A data center-based information storage method, characterized in that: include: Get the current load value of each server during off-peak hours; The servers whose current load values ​​are greater than the preset upper load value are defined as high-load servers, and the remaining servers are defined as idle servers; Performing a difference calculation based on the current load value and the upper limit load value in the high-load server to determine the required load release, and performing a difference calculation based on the current load value and the upper limit load value in the idle server to determine the remaining idle load; Determine whether the required release load is greater than the sum of the remaining idle loads; If the required release load is greater than the sum of the remaining idle loads, each server is controlled to maintain its original state; If the demand release load is not greater than the sum of the remaining idle loads, then determine the demand movement information in the high-load server according to the demand release load, determine the target server in each idle server according to each demand movement information, and control the demand movement information to move to the target server; The steps of determining the required movement information according to the demand for load release in the high-load server include: Obtain the information monomer load of each data information in the high-load server; In a high-load server, a random combination is performed according to any number of data information to determine a load information combination, and a summation calculation is performed according to the information monomer load of each data information in the load information combination to determine the information combination load; A load information combination whose information combination load is not less than the demand release load is defined as an alternative information combination; Determine the remaining idle load with the largest value in the idle servers according to a preset sorting rule, define the remaining idle load as the maximum load, and perform calculation according to the maximum load and a preset weight ratio to determine the single maximum load; The alternative information combinations whose information monomer loads are not greater than the monomer maximum load are defined as available information combinations, and a unique preferred combination is determined among the available information combinations, and the data information in the preferred combination is determined as the required mobile information.

2. The data center-based information storage method according to claim 1, characterized in that: The steps of determining a unique preferred combination among available information combinations include: Counting the data information in the available information combinations to determine the number of combined information; The excess release load is determined by performing a difference calculation based on the information combined load and the required release load; Calculation is performed based on the amount of combined information and the excess release load to determine the preferred parameters; The preferred parameter with the largest value is determined according to the sorting rule, and the available information combination corresponding to the preferred parameter is determined as the preferred combination.

3. The data center-based information storage method according to claim 1, characterized in that: The step of determining a target server from each idle server according to each demand movement information includes: According to the information monomer load of each demand mobile information, the mobile information sequence is determined by sorting from large to small, and the demand mobile information with the largest information monomer load is defined as the head-end mobile information; Determine the remaining idle load with the largest value according to the sorting rule, and define the idle server corresponding to the remaining idle load as the head-end server; The first-end mobile information is matched with the first-end server to determine the first-end server as the target server of the first-end mobile information, and the first-end mobile information is removed from the mobile information sequence and redefined, and the first-end server is re-determined and bound according to the new remaining idle load, until each required mobile information corresponds to a target server, and the first target solution is determined according to the current matching relationship between the first-end mobile information and the first-end server.

4. The data center-based information storage method according to claim 3, characterized in that: After the first target solution is determined, the information storage method based on the data center also includes: After the first head-end mobile information is matched with the head-end server, a difference calculation is performed based on the information monomer load of the head-end mobile information and the remaining idle load of the head-end server to determine the theoretical remaining load; Determine the demand mobility information in which the information unit load is not greater than the theoretical residual load and the difference with the theoretical residual load is the smallest in the demand mobility information, and match the demand mobility information with the current head-end server until the corresponding demand mobility information cannot be determined and the head-end server is redefined; Determine a second target solution according to the matching relationship between the current head-end mobile information and the head-end server; Obtaining the available remaining load of the idle server after each demand movement information is moved in the first target solution and the second target solution; defining a server whose available remaining load is in a preset medium load interval as a medium load server, and defining a server whose available remaining load is in a preset low load interval as a low load server; Counting against the medium load servers to determine the medium load number and counting against the low load servers to determine the low load number; A server whose available remaining load is in a preset waste interval is defined as a waste server, and a difference calculation is performed based on the available remaining load of the waste server and the lower limit of the waste interval to determine the resource waste load; Calculate the solution selection coefficient based on the number of medium loads, the number of low loads, and the resource waste load; A scheme with a larger scheme selection coefficient is selected from the first target scheme and the second target scheme as the target determination scheme, and a target server corresponding to each required mobile information is determined according to the target determination scheme.

5. The data center-based information storage method according to claim 1, characterized in that: If the required release load is greater than the sum of the remaining idle loads, the data center-based information storage method further includes: Obtain the possible call probability of each data information in the server; The data information whose possible call probability is greater than the preset upper limit probability is defined as mobile information, and the remaining data information is defined as fixed information; Calculation is performed based on the demand release load and the preset amplification factor to determine the probabilistic demand load; In the high-load server, the overall mobile load is determined by summing up the mobile information; Determine whether the overall mobile load is less than the probabilistic demand load; If the overall mobile load is not less than the probabilistic demand load, the high-load server is controlled to maintain its original state; If the overall moving load is less than the probabilistic demand load, the demand differential load is determined by performing a difference calculation based on the overall moving load and the probabilistic demand load; Selecting from the fixed information of the high-load server according to the possible call probability from small to large to determine a fixed mobile combination with a load just greater than the demand difference; Calculate the difference in load of information monomers based on the fixed information in the fixed-mobile combination and the mobile information in the remaining servers to determine the difference comparison load; The difference comparison load with the smallest value is determined according to the sorting rule, and the movement information relative to the difference comparison load is defined as the associated information of the fixed information, and the associated information is controlled to exchange positions with the corresponding movement information.

6. The data center-based information storage method according to claim 5, characterized in that: The step of determining the probability of possible invocation is also included, the step comprising: Establishing a detection interval on a preset time axis with the current time point as the end point and a width as the preset detection duration; Get the usage data information called within the detection interval; Determine usage data attributes and usage storage time based on usage data information; Count the data information corresponding to each usage data attribute to determine the number of attribute information, and calculate the proportion of individual attributes based on the number of each attribute information; Obtain the analysis data attributes and analysis storage time of each data information in each server; Determine the selection parameters based on the analysis data attributes and the corresponding monomer attribute proportions; Delineating a similar interval based on the storage time and the preset similar duration, and determining the correction parameter based on analyzing the relative relationship between the storage time and the similar interval; Calculations are performed based on the selection parameters and the correction parameters to determine the possible call probability.

7. An information storage system based on a data center, characterized in that: include: The acquisition module is used to obtain the current load value of each server during non-peak hours; A processing module, connected to the acquisition module and the judgment module, for storing and processing information; A judgment module, connected with the acquisition module and the processing module, for judging the information; The processing module defines a server whose current load value is greater than a preset upper load value as a high-load server, and defines the remaining servers as idle servers; The processing module performs a difference calculation based on the current load value and the upper limit load value in the high-load server to determine the required release load, and performs a difference calculation based on the current load value and the upper limit load value in the idle server to determine the remaining idle load; The judging module judges whether the required release load is greater than the sum of the remaining idle loads; If the judgment module determines that the required release load is greater than the sum of the remaining idle loads, the processing module controls each server to maintain the original state; If the judging module judges that the demand release load is not greater than the sum of the remaining idle loads, the processing module determines the demand movement information in the high-load server according to the demand release load, determines the target server in each idle server according to each demand movement information, and controls the demand movement information to move to the target server; The steps of determining the required movement information according to the demand for load release in the high-load server include: The acquisition module acquires the information monomer load of each data information in the high-load server; The processing module randomly combines any number of data information in the high-load server to determine a load information combination, and performs summation calculation based on the information monomer load of each data information in the load information combination to determine the information combination load; The processing module defines a load information combination whose information combination load is not less than the required release load as an alternative information combination; The processing module determines the remaining idle load with the largest value in the idle servers according to the preset sorting rules, defines the remaining idle load as the maximum load, and performs calculation according to the maximum load and the preset weight ratio to determine the single maximum load; The processing module defines candidate information combinations whose information monomer loads are not greater than the monomer maximum load as available information combinations, determines a unique preferred combination among the available information combinations, and determines the data information in the preferred combination as the required mobile information.

Images