Cloud platform capacity expansion and contraction method, device, equipment and medium

Abstract

The invention discloses a cloud platform capacity expansion and contraction method and device, equipment and a medium, and relates to the technical field of computers, and the method comprises the steps: determining a trend coefficient and a load fluctuation rate according to each target resource data of each virtual machine in a cloud platform and the size of a sliding window; determining a prediction error rate and a service priority, and adjusting the size of the sliding window according to the trend coefficient, the load fluctuation rate and the prediction error rate to obtain the adjusted size of the sliding window; the prediction error rate represents a deviation percentage between a historical prediction value and an actual value of the target resource data; determining each resource demand prediction value according to each target resource data and the adjusted size of the sliding window, and judging a to-be-executed operation of the cloud platform through a preset threshold value determined according to the service priority and each resource demand prediction value; the operation to be executed comprises a capacity expansion operation and a capacity reduction operation; and executing corresponding capacity expansion and shrinkage operation on the cloud platform according to the resource demand predicted value based on the judgment result. Elastic capacity expansion and shrinkage are achieved, and therefore the resource utilization rate is increased.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of computer technology, in particular to a cloud platform capacity expansion and contraction method, device, equipment and medium. BACKGROUND

[0002] With the rapid development of cloud computing technology, more and more enterprises and individuals deploy applications on cloud platforms. Cloud platform elastic capacity expansion and contraction methods are mainly divided into two categories: threshold-based methods and prediction-based methods. The threshold-based method sets a threshold for resource usage, and triggers capacity expansion when the actual resource usage exceeds the threshold, and triggers capacity contraction when the threshold is below. This method is simple and easy to implement, but has a reaction lag, often causing performance bottlenecks or resource waste when triggering capacity expansion and contraction.

[0003] The prediction-based method analyzes historical resource usage data to predict future resource demand and perform capacity expansion and contraction operations in advance. Common prediction algorithms include time series analysis, machine learning, etc. However, there are the following limitations:

[0004] Single parameter adjustment dimension: the sliding window size (N) is adjusted only according to the resource data fluctuation rate, which cannot adapt to the trend changes of the load (such as continuous growth or decline); the capacity expansion and contraction threshold (α, β) is in a fixed range, which cannot distinguish the differentiated needs of core and non-core businesses; the margin coefficient (γ) lacks a correlation mechanism with the prediction error, which is prone to insufficient capacity expansion or resource waste.

[0005] Insufficient capacity expansion and contraction strategy refinement: capacity expansion is based only on a static margin coefficient, without considering the load growth acceleration, making it difficult to respond to sudden traffic peaks; capacity contraction does not distinguish between core and non-core resource nodes, which may affect the stability of critical services; the frequency limit mechanism is fixed and cannot be dynamically adjusted according to the load fluctuation intensity, resulting in response lag or excessive system disturbance.

[0006] Therefore, how to combine dynamic parameter adjustment and refined strategy to achieve elastic capacity expansion and contraction is a problem that needs to be solved by those skilled in the art. SUMMARY

[0007] The purpose of the embodiments of the present application is to provide a cloud platform capacity expansion and contraction method, device, equipment and medium, which can combine dynamic parameter adjustment and refined strategy to achieve elastic capacity expansion and contraction, thereby improving resource utilization and reducing costs. The specific scheme is as follows:

[0008] In a first aspect, a cloud platform capacity expansion and contraction method is disclosed, comprising:

[0009] determine a trend coefficient and a load fluctuation rate according to the target resource data of each virtual machine in the cloud platform and the sliding window size; the target resource data comprises a central processing unit utilization rate, a memory usage, a disk input / output rate and a network bandwidth occupancy rate; the trend coefficient is a coefficient representing a change trend of a simple moving average value;

[0010] determine a prediction error rate and a service priority, adjust the sliding window size according to the trend coefficient, the load fluctuation rate and the prediction error rate, and obtain an adjusted sliding window size; the prediction error rate represents a deviation percentage between a historical prediction value and an actual value of the target resource data;

[0011] determine a resource demand prediction value according to each target resource data and the adjusted sliding window size, and determine a preset threshold according to the service priority; and determine a to-be-executed operation of the cloud platform according to the resource demand prediction value and the preset threshold; the to-be-executed operation comprises an expansion operation and a contraction operation;

[0012] perform a corresponding expansion / contraction operation on the cloud platform according to the resource demand prediction value based on a corresponding determination result.

[0013] Optionally, the method for determining the trend coefficient and the load fluctuation rate according to the target resource data of each virtual machine in the cloud platform and the sliding window size comprises the following steps.

[0014] determine a simple moving average value of each target resource data according to the sliding window size;

[0015] determine a first difference value between a first simple moving average value and a last simple moving average value in a target number of continuous simple moving average values;

[0016] determine a first product between a sampling period, the sliding window size and a target value; the target value is 2;

[0017] determine the trend coefficient corresponding to each target resource data according to a ratio between the first difference value and the first product;

[0018] determine the load fluctuation rate based on a ratio between a standard deviation and a mean value of each target resource data;

[0019] Correspondingly, the method for determining the service priority comprises the following steps.

[0020] determine the service priority according to a service priority label configured in advance through an application programming interface.

[0021] Optionally, the method for adjusting the sliding window size according to the trend coefficient, the load fluctuation rate and the prediction error rate, and obtaining the adjusted sliding window size comprises the following steps.

[0022] calculate an average trend coefficient according to the trend coefficient corresponding to each target resource data;

[0023] calculate an average load fluctuation rate according to the load fluctuation rates respectively corresponding to the target resource data;

[0024] calculate an average prediction error rate according to the prediction error rates respectively corresponding to the target resource data;

[0025] if the average load fluctuation rate is greater than a first threshold value, reduce the size of the sliding window, and if the average load fluctuation rate is less than a second threshold value, increase the size of the sliding window, to obtain a first sliding window size;

[0026] if the average trend coefficient is greater than a third threshold value, reduce the first sliding window size, and if the average trend coefficient is less than a fourth threshold value, increase the first sliding window size, to obtain a second sliding window size;

[0027] if the average prediction error rate for a continuous preset number of times is greater than a fifth threshold value, reduce the second sliding window size, and if the average prediction error rate for a continuous preset number of times is less than a sixth threshold value, increase the second sliding window size, to obtain an adjusted sliding window size.

[0028] Optionally, determining the resource demand prediction values of each target resource data according to the target resource data and the adjusted sliding window size, comprises:

[0029] determining the sum of the target resource data at each time point;

[0030] determining the resource demand prediction value as the ratio between the sum and the adjusted sliding window size.

[0031] Optionally, judging the to-be-executed operation of the cloud platform according to the preset threshold values determined according to the business priority and the resource demand prediction values, comprises:

[0032] performing weighted summation according to the resource demand prediction values respectively corresponding to each target resource data and the weight values, to obtain a comprehensive load;

[0033] if the resource demand prediction value or the comprehensive load of any target resource data is greater than a first preset threshold value determined according to the business priority, determining that the to-be-executed operation of the cloud platform is an expansion operation;

[0034] if the resource demand prediction values and the comprehensive loads of each target resource data are less than a second preset threshold value determined according to the business priority, determining that the to-be-executed operation of the cloud platform is a contraction operation.

[0035] Optionally, performing the corresponding expansion or contraction operation on the cloud platform according to the resource demand prediction values based on the corresponding judgment results, comprises:

[0036] If the to-be-executed operation of the cloud platform is an expansion operation, a load growth acceleration is determined according to a ratio between a second difference value and a sampling period; the second difference value is a difference between simple moving average line values of any two continuous target resource data;

[0037] A margin coefficient is determined based on the load growth acceleration;

[0038] A pre-increment is determined by a product between the resource demand prediction value, the load growth acceleration and a preset proportion coefficient;

[0039] A second product between the resource demand prediction value and the margin coefficient is determined;

[0040] A third difference value between the second product and the current capacity is determined, and a target sum between the third difference value and the pre-increment is determined;

[0041] An expansion capacity is determined by a maximum value between the target sum and a preset minimum expansion unit corresponding to the target resource data;

[0042] A maximum expansion capacity in the expansion capacities corresponding to the target resource data is determined as a target expansion capacity;

[0043] A corresponding expansion operation is executed on the cloud platform based on the target expansion capacity;

[0044] If the to-be-executed operation of the cloud platform is a contraction operation, a contraction capacity is determined according to the resource demand prediction value;

[0045] A minimum contraction capacity in the contraction capacities corresponding to the target resource data is determined as a target contraction capacity;

[0046] The core nodes are marked according to the service call chain; the core nodes are nodes bearing databases, gateways and authentication services;

[0047] A corresponding contraction operation is executed on the non-core nodes in the cloud platform according to the target contraction capacity.

[0048] Optionally, before the cloud platform is executed with a corresponding expansion or contraction operation according to the resource demand prediction value based on the corresponding judgment result, the method further includes:

[0049] A maximum expansion or contraction times in a preset unit time is determined according to the load fluctuation rate, so that the cloud platform is executed with a corresponding expansion or contraction operation according to the resource demand prediction value based on the corresponding judgment result according to the maximum expansion or contraction times.

[0050] In a second aspect, the application discloses a cloud platform expansion and contraction device, which includes:

[0051] a coefficient and volatility determination module configured to determine a trend coefficient and a load volatility according to target resource data of each virtual machine in the cloud platform and a sliding window size, the target resource data including central processing unit utilization, memory usage, disk input / output rate and network bandwidth occupancy, and the trend coefficient being a coefficient representing a change trend of a simple moving average value;

[0052] a sliding window size adjustment module configured to determine a prediction error rate and a service priority, and adjust the sliding window size according to the trend coefficient, the load volatility and the prediction error rate to obtain an adjusted sliding window size, the prediction error rate representing a deviation percentage between a historical prediction value and an actual value of the target resource data;

[0053] a determination module configured to determine each resource demand prediction value according to each target resource data and the adjusted sliding window size, and determine a to-be-executed operation of the cloud platform by comparing each resource demand prediction value with a preset threshold determined according to the service priority, the to-be-executed operation including an expansion operation and a contraction operation;

[0054] an expansion / contraction operation execution module configured to execute a corresponding expansion / contraction operation on the cloud platform according to the resource demand prediction value based on a corresponding determination result.

[0055] In a third aspect, the present application discloses an electronic device, comprising:

[0056] a memory configured to store a computer program;

[0057] a processor configured to execute the computer program to implement the cloud platform expansion / contraction method as described above.

[0058] In a fourth aspect, the present application discloses a computer readable storage medium, the computer readable storage medium storing a computer program, and the computer program is executed by a processor to implement the cloud platform expansion / contraction method as described above.

[0059] In the application, firstly, trend coefficients and load fluctuation rates are determined according to target resource data of each virtual machine in the cloud platform and a sliding window size; the target resource data includes central processing unit utilization, memory usage, disk input / output rate and network bandwidth occupancy rate; the trend coefficient is a coefficient for representing the change trend of a simple moving average value; a prediction error rate and a service priority are determined, the sliding window size is adjusted according to the trend coefficient, the load fluctuation rate and the prediction error rate, and an adjusted sliding window size is obtained; the prediction error rate represents the deviation percentage between a historical prediction value and an actual value of the target resource data; each resource demand prediction value is determined according to each target resource data and the adjusted sliding window size; the cloud platform is determined to perform an operation to be executed according to a preset threshold value determined according to the service priority and each resource demand prediction value; the operation to be executed includes a capacity expansion operation and a capacity reduction operation; and corresponding capacity expansion and reduction operations are performed on the cloud platform according to the resource demand prediction value based on the corresponding determination result.

[0060] Beneficial effects: The application adjusts the sliding window size in multiple dimensions through the fluctuation rate, the trend coefficient and the error rate, realizes intelligent adaptation of the sliding window size, and solves the technical problem that a fixed window is difficult to balance prediction accuracy and real-time performance. The differentiated threshold strategy is constructed based on the service priority, the accurate matching of the resource management strategy and the service characteristics is realized, and the defect that the traditional fixed threshold value cannot meet the diversified business needs is overcome. Then, through the multi-dimensional parameter optimization, the load feature driven decision logic and the business-aware resource scheduling, the intelligent management level of the cloud platform is improved, the elastic capacity expansion and reduction can be realized by combining dynamic parameter adjustment and fine-grained strategy, so as to improve the resource utilization rate and reduce the cost. BRIEF DESCRIPTION OF DRAWINGS

[0061] In order to more clearly illustrate the embodiments of the application, the drawings needed in the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the application, and other drawings can be obtained by those skilled in the art without creative labor.

[0062] Figure 1 A cloud platform capacity expansion and reduction method flow chart is provided for the embodiments of the application.

[0063] Figure 2 A specific cloud platform capacity expansion and reduction method flow chart is provided for the embodiments of the application.

[0064] Figure 3 A cloud platform capacity expansion and reduction device structure schematic diagram is provided for the embodiments of the application.

[0065] Figure 4 An electronic equipment structure diagram is provided for the embodiments of the application. DETAILED DESCRIPTION

[0066] The technical solutions in the embodiments of the present application will be apparently and completely described below with the accompanying drawings of the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments of the present application, all the other embodiments obtained by a person of ordinary skill in the art without creative work fall within the protection scope of the present application.

[0067] The terms of “include” and “have” and any variations of “include” and “have” in the specification and the above drawings of the present application are intended to cover the inclusions without the exclusions. For example, a process, a method, a system, a product, or an apparatus including a series of steps or units is not limited to the listed steps or units, but can include the steps or units not listed.

[0068] In order for the person in the technical field to better understand the present application, the present application will be further described below with the accompanying drawings and specific embodiments.

[0069] The prediction-based method predicts future resource demand by analyzing historical resource usage data, and performs scaling operation in advance. Common prediction algorithms include time series analysis, machine learning, etc. However, there are the following limitations: single dimension of parameter adjustment: the sliding window size (N) is only adjusted according to the resource data fluctuation rate, and cannot adapt to the trend changes of the load (such as continuous growth or decline); the scaling threshold (α, β) is in a fixed range, and cannot distinguish the differentiated demand of core and non-core business; the margin coefficient (γ) lacks a correlation mechanism with the prediction error, and is easy to cause insufficient scaling or resource waste. The scaling strategy is not refined enough: the scaling is only based on the static margin coefficient, and the load growth acceleration is not considered, which is difficult to respond to sudden traffic peaks; the scaling does not distinguish core and non-core resource nodes, which may affect the stability of key services; the frequency limitation mechanism is fixed, and cannot be dynamically adjusted according to the load fluctuation intensity, resulting in response lag or excessive system disturbance. In order to solve the above technical problems, the present application discloses a cloud platform scaling method, device, equipment and medium, which can realize elastic scaling by combining dynamic parameter adjustment and refined strategy, thereby improving resource utilization and reducing cost.

[0070] Referring to Figure 1 The embodiment of the present application provides a cloud platform scaling method, which comprises the following steps:

[0071] Step S11, determining a trend coefficient and a load fluctuation rate according to target resource data of each virtual machine in the cloud platform and a sliding window size; the target resource data comprises a central processing unit utilization rate, a memory usage, a disk input / output rate and a network bandwidth occupancy rate; the trend coefficient is a coefficient representing the change trend of a simple moving average line value.

[0072] In the embodiment of the present application, firstly, key performance indicators of each virtual machine in the cloud platform are collected, including but not limited to CPU (Central Processing Unit) utilization, memory usage, disk I / O (input / output) rate, network bandwidth occupancy rate, and the sampling period Δt can be dynamically adjusted in the range of 1-60 seconds (default 5-30 seconds). Then, the simple moving average line value of each target resource data is determined according to the sliding window size; the first difference value between the first simple moving average line value and the last simple moving average line value in the target number of continuous simple moving average line values is determined; the first product between the sampling period, the sliding window size and the target value is determined; the target value is 2; the trend coefficient corresponding to each target resource data is determined according to the ratio between the first difference value and the first product; and the load fluctuation rate is determined based on the ratio between the standard deviation and the mean value of each target resource data. Specifically, the formula of the change slope of the SMA (Simple Moving Average) values of the continuous 3 sliding windows is: trend coefficient k = (M1-M3) / (2×Δt×N). Wherein, N is the sliding window size; Δt is the sampling period; M1 is the first simple moving average line value in the 3 sliding windows; M3 is the last simple moving average line value in the 3 sliding windows; and the ratio between the standard deviation and the mean value of the data in the current window is the load fluctuation rate. It should be noted that the SMA, trend coefficient (k), fluctuation rate (σ) and prediction error rate (e) of CPU, memory, disk I / O rate and network bandwidth occupancy rate need to be calculated respectively. -

[0073] In step S12, the prediction error rate and the service priority are determined, the sliding window size is adjusted according to the trend coefficient, the load fluctuation rate and the prediction error rate, and the adjusted sliding window size is obtained; and the prediction error rate represents the percentage deviation between the historical prediction value and the actual value of the target resource data.

[0074] ​​​In the embodiment of the present application, the prediction error rate is the percentage of deviation of the historical prediction value from the actual value, and the formula is: error rate e = |predicted value-actual value| / actual value*100%. The service priority is determined according to the service priority label configured in advance through the application programming interface. That is, according to the service importance, the services are divided into core services (such as payment systems and transaction platforms) and non-core services (such as log analysis and backup services), and the label can be dynamically configured through the API (Application Programming Interface) interface. Then the collected data and characteristic values are stored in a time series database (such as InfluxDB and TimescaleDB) in time sequence, which provides a basis for subsequent parameter adjustment and decision-making.

[0075] Then, when adjusting the sliding window size, the average trend coefficient is calculated according to the trend coefficient corresponding to each target resource data, the average load fluctuation rate is calculated according to the load fluctuation rate corresponding to each target resource data, and the average prediction error rate is calculated according to the prediction error rate corresponding to each target resource data. If the average load fluctuation rate is greater than the first threshold value, the sliding window size is reduced, and if the average load fluctuation rate is less than the second threshold value, the sliding window size is increased, to obtain the first sliding window size. If the average trend coefficient is greater than the third threshold value, the first sliding window size is reduced, and if the average trend coefficient is less than the fourth threshold value, the first sliding window size is increased, to obtain the second sliding window size. If the average prediction error rate of a continuous preset number of times is greater than the fifth threshold value, the second sliding window size is reduced, and if the average prediction error rate of a continuous preset number of times is less than the sixth threshold value, the second sliding window size is increased, to obtain the adjusted sliding window size. Specifically, the initial value of the sliding window size N is 20 (the value range is 5-100), which is dynamically adjusted based on the following multi-dimensional conditions: basic adjustment logic: when the fluctuation rate σ>20%, N is reduced by 10%-20% (to increase the sensitivity to the latest data); when σ<10%, N is increased by 10%-20% (to smooth the noise interference). Trend adaptation adjustment: if the trend coefficient k of a continuous 5 windows is greater than 0.05 (rapid load rise), N is additionally reduced by 10% regardless of whether the fluctuation rate reaches or not; if k<-0.05 (rapid load decline), N is additionally increased by 10%. Error feedback adjustment: when the prediction error rate e is greater than 25% for 3 consecutive times, N is reduced by 5%-10% (to improve the prediction real-time performance); when e is less than 5%, N is increased by 5%-10% (to improve the prediction stability). And the adjusted N value needs to satisfy 5≤N≤100, and the single adjustment amplitude is not more than 30%, to avoid the parameter from fluctuating sharply.

[0076] In this way, the limitation of adjusting the window size according to the traditional single volatility is broken, and the sliding window size (N) is intelligently adapted by creatively combining multi-dimensional factors such as load trend and prediction error rate. When the load rapidly rises (k>0.05) or the prediction error is too large (e>25%), the window is automatically reduced to enhance the response sensitivity; when the load is stable or the prediction is accurate, the window is increased to smooth the noise, thereby solving the technical problem that the fixed window is difficult to balance the prediction accuracy and real-time performance.

[0077] In step S13, the resource demand prediction value of each target resource data is determined according to the adjusted sliding window size, and the preset threshold value determined according to the business priority and the resource demand prediction value are used to determine the to-be-executed operation of the cloud platform; the to-be-executed operation includes the expansion operation and the contraction operation.

[0078] In the embodiment of the application, when the resource demand prediction value of each target resource data is determined according to the adjusted sliding window size, the sum of the target resource data at each time point is determined; and the ratio between the sum and the adjusted sliding window size is determined as the resource demand prediction value. In this process, the simple moving average value is used as the prediction basis, and the formula is: SMA(t)=(x(t-N+1)+x(t-N+2)+...+x(t)) / N; wherein x(t-i) is the resource data at the t-i time point, SMA(t) is used as the resource demand prediction value of the T time period (T=N*Δt) after the t time point, and N is the adjusted sliding window size.

[0079] Then, the preset threshold value is determined according to the business priority, the core business: α=60%-70% (expansion in advance to ensure high availability), β=40%-50% (conservative contraction to reduce service risk); non-core business: α=80%-85% (improve resource utilization), β=20%-30% (aggressive contraction to reduce waste); period dynamic adjustment: the α of the core business is reduced by 5%-10% in the peak period (such as 9:00-18:00), and the β of the non-core business is reduced by 5%-10% in the trough period (such as 0:00-6:00). In this way, the differentiated threshold (α, β) strategy is constructed based on the business priority, the core business adopts the conservative threshold (α=60%-70%) to ensure high availability, and the non-core business adopts the aggressive threshold (α=80%-85%) to improve resource utilization. At the same time, the threshold is dynamically adjusted combined with the period characteristics, the accurate matching of the resource management strategy and the business characteristics is realized, and the defect that the traditional fixed threshold cannot meet the diversified business demand is overcome.

[0080] Then, the resource demand prediction value of each target resource data is weighted and summed according to the corresponding weight value to obtain a comprehensive load; if the resource demand prediction value or the comprehensive load of any target resource data is greater than a first preset threshold value determined according to the service priority, it is determined that the to-be-executed operation of the cloud platform is an expansion operation; if the resource demand prediction value and the comprehensive load of each target resource data are less than a second preset threshold value determined according to the service priority, it is determined that the to-be-executed operation of the cloud platform is a contraction operation. Otherwise, no expansion or contraction operation is performed. In the embodiment of the application, the SMA, the trend coefficient (k), the volatility rate (σ) and the prediction error rate (e) of the CPU, the memory and the like are calculated in parallel, and then the prediction values of multiple indexes are aggregated into a comprehensive load score S according to the weight strategy read from the configuration center.

[0081] In a specific embodiment, it is assumed that not only CPU data but also memory data (unit: %) is collected in the same sampling period. The current window CPU data is [20%, 21%, 22%, 23%, 24%, 25%, 26%, 27%, 28%, 29%, 31%, 31%, 32%, 33%, 34%, 35%, 36%, 37%, 38%, 39%]; the SMA of the CPU utilization rate itself is: = (20% + 21% + 22% +... + 39%) / 20 = 35%. The current window memory usage sequence is [65%, 66%, 67%, 65%, 68%, 69%, 70%, 68%, 71%, 70%, 72%, 73%, 71%, 74%, 75%, 73%, 76%, 75%, 77%, 78%]; the memory = 71%. The trend coefficient k_m = 0.03% / s (assuming that the rising trend is more violent than that of the CPU). The load volatility rate σ_m = 3.5% (the memory usage grows very smoothly). The CPU index: = 35%, k_c = 0.02% / s, σ_c = 8.26%; the memory index: = 71%, k m = 0.03% / s, σ m = 3.5%. Define weights for different services in the configuration center. For example: core service (virtual machine operation): CPU intensive, weight W c = 0.7, memory weight W m = 0.3; non-core service (log analysis): memory intensive, weight W c = 0.4, memory weight W m = 0.6. The calculation formula is: S = (SMA c * W c + SMA m * W m) * 100%. Finally, S = (35% * 0.7) + (71% * 0.3) = 24.5% + 21.3% = 45.8%. The comprehensive load is 45.8%, which has not reached the core service expansion threshold α (60%). Make decisions based on the comprehensive load score S and the independent threshold of any index. This is an "or" logic, which ensures that any critical resource shortage can trigger expansion. Arbitration logic:

[0082] IF S>=α OR SMA_c>=α OR SMA_m>=α_m (memory independent threshold) - trigger expansion; (if any of S, CPU usage, memory usage reaches the expansion threshold, expansion is triggered);

[0083] IF S<=β AND SMA_c<=β AND SMA_m<=β_m - trigger contraction (contraction must meet all indicators, safer); (only if S, CPU usage, memory usage all reach the contraction threshold, contraction is triggered).

[0084] Therefore, although the comprehensive load S = 45.8% < 60%, the memory index SMA m = 71% has exceeded the core service's independent memory expansion threshold (assuming α m = 70%). Finally, expansion is triggered. Because it meets the condition of SMA m >= α m.

[0085] Step S14, based on the corresponding judgment result, perform corresponding expansion and contraction operations on the cloud platform according to the resource demand prediction value.

[0086] In this embodiment of the invention, if the operation to be performed on the cloud platform is a capacity expansion operation, the load growth acceleration is determined based on the ratio between the second difference and the sampling period; the second difference is the difference between the simple moving averages of any two consecutive target resource data; a margin coefficient is determined based on the load growth acceleration; a pre-increment is determined by multiplying the resource demand forecast, the load growth acceleration, and a preset proportional coefficient; a second product is determined between the resource demand forecast and the margin coefficient; a third difference is determined between the second product and the current capacity, and a target sum is determined between the third difference and the pre-increment; the target sum is then compared with the target resource data. The expansion capacity is determined by the maximum value among the corresponding preset minimum expansion units; the largest expansion capacity among the expansion capacities corresponding to each target resource data is determined as the target expansion capacity; the corresponding expansion operation is performed on the cloud platform based on the target expansion capacity; if the operation to be performed on the cloud platform is a shrinkage operation, the shrinkage capacity is determined based on the predicted resource demand; the smallest shrinkage capacity among the shrinkage capacities corresponding to each target resource data is determined as the target shrinkage capacity; core nodes are marked according to the service call chain; core nodes are nodes that carry database, gateway, and authentication services; the corresponding shrinkage operation is performed on non-core nodes in the cloud platform based on the target shrinkage capacity. Specifically, the expansion strategy is as follows: growth acceleration calculation: load growth acceleration a = ( ) / Δt, reflecting the rate of load change; dynamic margin coefficient: γ=1.1+0.05×a (when a≥0), γ [1.1, 1.5], the greater the acceleration, the higher the margin coefficient; Pre-increment addition: if a>0 and increases for two consecutive windows, the pre-increment = predicted value × a × 0.1 (the proportional coefficient 0.1 can be dynamically adjusted); Expansion formula: Expansion resource amount = max((predicted value × γ - current capacity) + pre-increment, minimum expansion unit). CPU requirement: Expansion resource amount _c = max(( ×γ_c-100%)+Pre-increment_c,Minimum expansion unit_c);Memory requirement:Expansion resource amount_m=max(( (×γ_m-100%) + pre-increment_m, minimum expansion unit_m). The final expansion capacity is the greater of the CPU requirement and the memory requirement as the final number of instances to be expanded. Because expansion is performed on a per-virtual machine or Pod basis, all resource demands must be met. In a specific implementation, 2 virtual machines need to be expanded in terms of CPU. In terms of memory: assuming the same logic is used (γ_m=1.1, pre-increment_m=5%), it is found that 3 virtual machines need to be expanded to meet the memory requirement. The system will ultimately take the maximum value and perform the operation of expanding 3 virtual machines, thereby simultaneously meeting the growth requirements of CPU and memory.

[0087] And the shrinkage strategy is: based on service call chain data, mark the nodes of the bearer database, gateway and authentication service as core nodes, and preferentially shrink non-core nodes; safety margin setting: the capacity after shrinkage ≥ predicted value × 1.05 (reserve 5% safety margin); dependency check: check resource dependency before shrinkage (such as whether the shrinkage of a node leads to disk I / O overload of associated nodes), and if there is a risk, the shrinkage is suspended. During the shrinkage process, shrinkage is performed according to the minimum shrinkage amount, such as cpu calculating that 2 virtual machines need to be shrunk, and memory needs to be shrunk by 2 virtual machines, and shrinkage is performed according to 2 virtual machines to meet business requirements.

[0088] In addition, in the cloud platform shrinkage scenario, in addition to preferentially shrinking non-critical nodes, nodes with CPU utilization below 20% and memory idle rate above 60% can also be preferentially shrunk; or automatically trigger shrinkage during the business low period (such as 2-5 am), predict the load cycle through time series analysis; or set podDisruptionBudget for critical pods to ensure that core services are not affected during shrinkage; or adopt batch shrinkage (such as not more than 10% of the total number of nodes each time), and cooperate with health checks to avoid the snowball effect.

[0089] In the embodiment of the application, before the cloud platform is executed according to the resource demand prediction value based on the corresponding judgment result, the maximum expansion and shrinkage times in a preset unit time are determined according to the load fluctuation rate, so that the cloud platform is executed according to the resource demand prediction value based on the corresponding judgment result. In a specific embodiment, the maximum expansion and shrinkage times in a unit time (default 10 minutes) are dynamically adjusted according to the fluctuation rate σ: σ>30% (intense fluctuation): allow 5 operations, quickly respond to user demand; 10%≤σ≤30% (moderate fluctuation): allow 3 operations; σ<10% (steady state): allow 1 operation, reduce system disturbance. And the application can also store historical data and operation records, and optimize parameters such as γ, pre-increment ratio through machine learning model. This intelligent frequency limiting mechanism reduces invalid operations in the steady state by 70%, and improves the timeliness of response in intense fluctuation by 50%.

[0090] In this way, the expansion link innovatively introduces the concept of load growth acceleration (a), realizes the forward-looking response to the rising trend of load through dynamic margin coefficient (γ=1.1+0.05×a) and pre-increment mechanism; the shrinkage link first identifies core nodes and sets safety margin (≥ predicted value × 1.05), realizes accurate release of resources under the premise of ensuring service stability, and solves the problem of response lag of traditional expansion and shrinkage strategy to load change trend and easy to cause service risk.

[0091] Beneficial effects: The application realizes intelligent adaptation of the sliding window size by adjusting the sliding window size in multiple dimensions of volatility, trend coefficient and error rate, solves the technical problem that the fixed window is difficult to balance the prediction accuracy and real-time performance. Based on the business priority, a differentiated threshold strategy is constructed to realize accurate matching of resource management strategy and business characteristics, and overcome the defects that the traditional fixed threshold cannot meet the diversified business needs. Then, through multi-dimensional parameter optimization, load feature driven decision logic and business aware resource scheduling, the intelligent management level of the cloud platform is improved, the elastic expansion and contraction can be realized by combining dynamic parameter adjustment and fine-grained strategy, so as to improve the resource utilization and reduce the cost.

[0092] Based on the above embodiment, the application discloses a cloud platform expansion and contraction method, which can realize elastic expansion and contraction. Next, the cloud platform expansion and contraction method will be described in detail according to specific embodiments.

[0093] Referring to Figure 2 The application discloses a specific cloud platform expansion and contraction method. The scene parameters in a specific embodiment are:

[0094] Business type: virtual machine operation (core business), log analysis (non-core business);

[0095] Sampling period Δt=5 seconds, initial window size N=20;

[0096] Core business α=60%, β=40%; Non-core business α=85%, β=20%;

[0097] Minimum expansion unit=2 virtual machines, minimum contraction unit=2GB memory, cpu=2%.

[0098] The data acquisition module collects the memory usage of each virtual machine every 5 seconds to form time series data. For example, the time series of CPU utilization of a certain virtual machine is as follows, and the calculation is as follows:

[0099] The current window CPU data is [20%, 21%, 22%, 23%, 24%, 25%, 26%, 27%, 28%, 29%, 31%, 31%, 32%, 33%, 34%, 35%, 36%, 37%, 38%, 39%]; The SMA of CPU utilization itself is: =(20%+21%+22%+...+39%) / 20=35%. The data of the first two windows is [10%, 10%, 11%, 11%, 12%, 12%, 13%, 13%, 14%, 14%, 15%, 16%, 16%, 17%, 17%, 18%, 18%, 19%, 19%, 20%]; =(10%+10%+11%+...+20%) / 20=31%; trend coefficient k=(35%-31%) / (2x5x20)=0.02% / s; volatility μ=34% (mean value in window) Window data: 30%, 31%, 31%, 32%, 33%, 34%, 35%, 36%, 37%, 38%; prediction error rate e=8%.

[0100] Previous 3 predicted values: 28%, 30%, 32%; corresponding actual values: 30%, 31%, 33%. e1=|28-30| / 30x100%≈6.67%, e2=|30-31| / 31x100%≈3.23%, e3=|32-33| / 33x100%≈3.03%. Current error rate: e=(6.67%+3.23%+3.03%) / 3≈4.31%.

[0101] Sliding window adjustment: trend coefficient k=0.02% / s>0.01% / s (load rising); volatility σ=8.26%<10% (smooth fluctuation); error rate e=4.31%<5% (prediction accurate). Since the trend is rising, but the volatility is low and the prediction is accurate, N increases from 20 to 22 (increased by 10%), balancing smoothness and trend response.

[0102] SMA predicted value is 62% of current capacity (sum of CPU utilization data at 22 time points / N), which exceeds the core business expansion threshold α=60%, triggering expansion (this is a single resource data example, if it is multiple resource data, the calculation method of other resource data is the same as the above calculation process, and when judging whether to trigger expansion, a comprehensive load S needs to be calculated according to the predicted value of all resource data, and then the relationship between the predicted value of all resource data and the expansion threshold needs to be judged. If S, CPU usage, memory usage, etc. Any one of the resource data reaches the expansion threshold, expansion is triggered).

[0103] When expanding, the growth acceleration a=(62%-58%) / 5=0.8% / s; the margin coefficient γ=1.1+0.05x0.8=1.14; the pre-increment=62% x 0.8 x 0.1=4.96%; the expansion resource amount=max((62% x 1.14-100%)+4.96%, 2 virtual machines)=2 virtual machines. In this process, real-time monitoring is carried out, and during the active peak period (volatility σ=35%>30%), 5 expansions are allowed within 10 minutes, and the system completes 2 expansions according to the real-time load to ensure service stability.

[0104] When the capacity is reduced, the load decreases (k=-0.015% / s); the current server is identified as a non-core node, and the capacity after the reduction is = predicted value x 1.05 = 25% x 1.05 = 26.25%. (Similarly, when the capacity is reduced, the predicted values of all resource data need to be calculated to obtain a comprehensive load S, and then the relationship between the predicted values of all resource data and the expansion threshold needs to be determined. If S, CPU usage, memory usage and other resource data all reach the expansion threshold, the reduction is triggered). Since the load fluctuation rate σ = 5% < 10%, the reduction is only performed once within 10 minutes.

[0105] Beneficial effects: Establish operation frequency limit rules linked with load fluctuation rate (σ), relax the limit to 5 times / 10 minutes to quickly respond in the case of severe fluctuations (σ>30%), and tighten the limit to 1 time / 10 minutes to reduce disturbance in the case of stable state (σ<10%). This mechanism balances system response speed and stability, avoiding the lack of adaptability of fixed frequency limits in complex load scenarios. At the same time, the size of the sliding window is adjusted in multiple dimensions through fluctuation rate, trend coefficient and error rate, reducing the prediction error rate by 30%-50%; the dynamic threshold adapts to business priority and time period characteristics, improving the availability of core business by more than 20%, significantly improving parameter adaptability. The expansion combined with the growth acceleration and pre-amount mechanism improves the response speed to sudden loads by 40%; the reduction through core node protection and safety margin setting reduces service interruption rate by 60%, thereby optimizing the expansion and reduction accuracy.

[0106] Referring to Figure 3 The cloud platform expansion and reduction device provided by the embodiment of the present application comprises:

[0107] The coefficient and fluctuation rate determination module 11 is configured to determine a trend coefficient and a load fluctuation rate according to target resource data of each virtual machine in the cloud platform and a sliding window size; the target resource data comprises a central processing unit utilization rate, a memory usage, a disk input / output rate and a network bandwidth occupancy rate; the trend coefficient is a coefficient representing a change trend of a simple moving average value;

[0108] The sliding window size adjustment module 12 is configured to determine a prediction error rate and a business priority, adjust the sliding window size according to the trend coefficient, the load fluctuation rate and the prediction error rate, and obtain an adjusted sliding window size; the prediction error rate represents a deviation percentage between a historical prediction value and an actual value of the target resource data;

[0109] The determination module 13 is configured to determine resource demand prediction values according to each target resource data and the adjusted sliding window size, and determine a to-be-executed operation of the cloud platform by a preset threshold determined according to the business priority and the resource demand prediction values; the to-be-executed operation comprises an expansion operation and a reduction operation;

[0110] The scaling operation execution module 14 is configured to perform a corresponding scaling operation on the cloud platform according to the resource demand prediction value based on the corresponding judgment result.

[0111] Since the embodiments of the device part correspond to the above-mentioned embodiments, the embodiments of the device part are described with reference to the embodiments of the method part, which will not be repeated here.

[0112] Beneficial effects: The application adjusts the size of the sliding window in multiple dimensions through the volatility rate, the trend coefficient and the error rate, realizes intelligent adaptation of the size of the sliding window, and solves the technical problem that a fixed window is difficult to balance the prediction accuracy and the real-time performance. The application constructs a differentiated threshold strategy based on the business priority, realizes accurate matching of the resource management strategy and the business characteristics, and overcomes the defect that the traditional fixed threshold cannot meet the diversified business demands. Furthermore, through multi-dimensional parameter optimization, decision logic driven by load characteristics and resource scheduling perceived by the business, the intelligent management level of the cloud platform is improved, elastic scaling can be realized by combining dynamic parameter adjustment and fine-grained strategy, so as to improve the resource utilization rate and reduce the cost.

[0113] Further, the application also discloses an electronic device, Figure 4 is an electronic device structure diagram shown according to an exemplary embodiment, the content in the figure cannot be considered as any limitation on the use range of the application. The electronic device can specifically include: at least one processor 21, at least one memory 22, a power supply 23, a communication interface 24, an input / output interface 25 and a communication bus 26. The memory 22 is used to store a computer program, the computer program is loaded and executed by the processor 21 to realize the related steps in the cloud platform scaling method disclosed in any of the preceding embodiments. In addition, the electronic device in the embodiment can be an electronic computer.

[0114] In the embodiment, the power supply 23 is used to provide working voltage for each hardware device on the electronic device; the communication interface 24 can create a data transmission channel between the electronic device and the external device, and the communication protocol followed by the communication interface 24 can be any communication protocol applicable to the technical solution of the application, which will not be limited specifically; the input / output interface 25 is used to obtain external input data or output data to the outside, and the specific interface type can be selected according to the specific application needs, which will not be limited specifically.

[0115] In addition, the memory 22 as a carrier for storing resources can be a read-only memory, a random access memory, a magnetic disk or an optical disk, and the resources stored thereon can include an operating system 221, a computer program 222 and the like, and the storage mode can be temporary storage or permanent storage.

[0116] The operating system 221 is configured to manage and control each hardware device on the electronic device and the computer program 222, which can be Windows Server, Netware, Unix, Linux, etc. The computer program 222 can further include computer programs capable of performing other specific work in addition to the computer program capable of performing the cloud platform scaling method disclosed in any of the preceding embodiments by the electronic device.

[0117] Further, the present application also discloses a computer readable storage medium for storing a computer program; wherein the computer program is executed by a processor to implement the cloud platform scaling method disclosed in the preceding embodiments. The specific steps of the method can refer to the corresponding contents disclosed in the preceding embodiments, which will not be repeated here.

[0118] Further, the present application also discloses a computer program product comprising computer programs / instructions; wherein the computer programs / instructions are executed by a processor to implement the cloud platform scaling method disclosed in the preceding embodiments. The specific steps of the method can refer to the corresponding contents disclosed in the preceding embodiments, which will not be repeated here.

[0119] The embodiments in the specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts of each embodiment can be referred to each other. For the device disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple, and the relevant parts can refer to the method part.

[0120] The skilled person can further realize that the units and algorithm steps of the examples described in combination with the embodiments disclosed herein can be realized by electronic hardware, computer software or a combination of both. In order to clearly show the interchangeability of hardware and software, the components and steps of each example have been described in the above description. Whether the functions are realized in hardware or software depends on the specific application and design constraints of the technical solution. The skilled person can use different methods to realize the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0121] The steps of the method or algorithm described in combination with the embodiments disclosed herein can be directly implemented by hardware, software modules executed by a processor, or a combination of both. The software modules can be placed in random access memory (RAM), memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disks, removable disks, CD-ROMs, or any other form of storage medium known in the art.

[0122] Finally, it needs to be pointed out that in this document, relational terms such as first and second and the like can only be intended to distinguish one entity or operation from another entity or operation without necessarily requiring or implying any actual such relationship or order between such entities or operations. Moreover, the terms "comprising", "including", or any other variant thereof are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can also include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without more limitations, an element defined by the statement "comprising a" does not exclude the existence of additional identical elements in the process, method, article, or apparatus including the stated element.

[0123] The above detailed description of the technical solutions provided by the present application has been provided, and the principles and implementation modes of the present application have been described by applying specific examples. The above description of the examples is only applicable to help understand the method and core idea of the present application; meanwhile, for those skilled in the art, according to the idea of the present application, the specific implementation modes and application ranges will have changes, and the above description of the content of the specification should not be understood as a limitation of the present application.

Claims

1. A method for scaling up and down a cloud platform, characterized in that, include: The trend coefficient and load volatility are determined based on the target resource data and sliding window size of each virtual machine in the cloud platform. The target resource data includes CPU utilization, memory usage, disk I / O rate, and network bandwidth utilization. The trend coefficient is a coefficient that characterizes the changing trend of the simple moving average line value; Determine the prediction error rate and business priority, and adjust the sliding window size according to the trend coefficient, the load volatility, and the prediction error rate to obtain the adjusted sliding window size; the prediction error rate represents the percentage deviation between the historical predicted value and the actual value of the target resource data; Based on the target resource data and the adjusted sliding window size, the predicted value of each resource demand is determined. The cloud platform's pending operations are determined by a preset threshold determined by the business priority and the predicted value of each resource demand. The pending operations include expansion operations and reduction operations. Based on the corresponding judgment results, the cloud platform is subjected to corresponding scaling up or down operations according to the predicted resource demand.

2. The cloud platform scaling method according to claim 1, characterized in that, The determination of trend coefficients and load volatility based on target resource data and sliding window size of each virtual machine in the cloud platform includes: The simple moving average value of each target resource data is determined based on the sliding window size. Determine the first difference between the first and last simple moving average values ​​in a target number of consecutive simple moving average values; Determine the first product between the sampling period, the sliding window size, and the target value; the target value is 2. The trend coefficient corresponding to each target resource data is determined based on the ratio between the first difference and the first product. The load volatility is determined based on the ratio between the standard deviation and the mean of each of the target resource data. Accordingly, business priorities are determined, including: Business priorities are determined based on business priority labels pre-configured via the application programming interface.

3. The cloud platform scaling method according to claim 1, characterized in that, The step of adjusting the sliding window size based on the trend coefficient, the load volatility, and the prediction error rate to obtain the adjusted sliding window size includes: Calculate the average trend coefficient based on the trend coefficient corresponding to each of the target resource data; Calculate the average load volatility based on the load volatility corresponding to each of the target resource data; Calculate the average prediction error rate based on the prediction error rate corresponding to each of the target resource data; If the average load volatility is greater than a first threshold, the sliding window size is reduced; if the average load volatility is less than a second threshold, the sliding window size is increased to obtain the first sliding window size. If the average trend coefficient is greater than the third threshold, the size of the first sliding window is reduced; if the average trend coefficient is less than the fourth threshold, the size of the first sliding window is increased to obtain the size of the second sliding window. If the average prediction error rate for a consecutive preset number of times is greater than the fifth threshold, then the size of the second sliding window is reduced; if the average prediction error rate for a consecutive preset number of times is less than the sixth threshold, then the size of the second sliding window is increased to obtain the adjusted sliding window size.

4. The cloud platform scaling method according to claim 1, characterized in that, The step of determining the predicted value of each resource demand based on the target resource data and the adjusted sliding window size includes: Determine the sum of the target resource data at each time point; The ratio between the sum and the adjusted sliding window size is determined as the predicted resource demand value.

5. The cloud platform scaling method according to claim 1, characterized in that, The step of determining the pending operations of the cloud platform based on a preset threshold determined by the business priority and the predicted values ​​of each resource demand includes: The comprehensive load is obtained by weighting and summing the predicted resource demand values ​​and weight values ​​corresponding to each of the target resource data. If the predicted resource demand value of any of the target resource data or the overall load is greater than the first preset threshold determined according to the business priority, then the operation to be executed by the cloud platform is determined to be an expansion operation. If the predicted resource demand value of each of the target resource data and the overall load are both less than the second preset threshold determined according to the business priority, then the operation to be executed by the cloud platform is determined to be a scaling-down operation.

6. The cloud platform scaling method according to claim 5, characterized in that, The step of performing corresponding scaling operations on the cloud platform based on the corresponding judgment results and the predicted resource demand includes: If the operation to be performed on the cloud platform is a capacity expansion operation, the load growth acceleration is determined based on the ratio between the second difference and the sampling period; the second difference is the difference between the simple moving average values ​​of any two consecutive target resource data. The margin coefficient is determined based on the aforementioned load growth acceleration. The pre-increment is determined by multiplying the predicted resource demand, the load growth acceleration, and a preset proportional coefficient. Determine the second product between the predicted resource demand and the margin coefficient; Determine a third difference between the second product and the current capacity, and determine a target sum between the third difference and the pre-increment; The expansion capacity is determined by the maximum value between the target and the preset minimum expansion unit corresponding to the target resource data; The largest expansion capacity among the expansion capacities corresponding to each of the target resource data is determined as the target expansion capacity; Based on the target capacity expansion, perform corresponding capacity expansion operations on the cloud platform; If the operation to be performed on the cloud platform is a scaling-down operation, the scaling-down capacity is determined based on the predicted resource demand. The smallest reduction capacity among the reduction capacities corresponding to each of the target resource data is determined as the target reduction capacity; Core nodes are marked according to the service call chain; the core nodes are those that carry database, gateway, and authentication services. Perform corresponding scaling-down operations on non-core nodes in the cloud platform according to the target scaling-down.

7. The cloud platform scaling method according to any one of claims 1 to 6, characterized in that, Before performing corresponding scaling operations on the cloud platform based on the resource demand forecast value according to the corresponding judgment result, the method further includes: The maximum number of scaling up and down operations per unit time is determined based on the load volatility, so that the cloud platform can perform corresponding scaling up and down operations based on the resource demand forecast value according to the corresponding judgment result based on the maximum number of scaling up and down operations.

8. A cloud platform scaling device, characterized in that, include: The coefficient and volatility determination module is used to determine the trend coefficient and load volatility based on the target resource data of each virtual machine in the cloud platform and the sliding window size; The target resource data includes CPU utilization, memory usage, disk I / O rate, and network bandwidth utilization. The trend coefficient is a coefficient that characterizes the changing trend of the simple moving average line value; The sliding window size adjustment module is used to determine the prediction error rate and business priority, and adjust the sliding window size according to the trend coefficient, the load volatility rate and the prediction error rate to obtain the adjusted sliding window size; the prediction error rate represents the percentage deviation between the historical predicted value and the actual value of the target resource data; The judgment module is used to determine the predicted value of each resource demand based on the target resource data and the adjusted sliding window size, and to determine the pending operations of the cloud platform based on the preset threshold determined by the business priority and the predicted value of each resource demand; the pending operations include expansion operations and shrinkage operations; The scaling up / down operation execution module is used to perform corresponding scaling up / down operations on the cloud platform based on the corresponding judgment results and the predicted resource demand values.

9. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor for executing the computer program to implement the steps of the cloud platform scaling method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the cloud platform scaling method as described in any one of claims 1 to 7.

Images