CN112269811A - IT capacity prediction method and system based on traffic - Google Patents
IT capacity prediction method and system based on traffic Download PDFInfo
- Publication number
- CN112269811A CN112269811A CN202011092741.9A CN202011092741A CN112269811A CN 112269811 A CN112269811 A CN 112269811A CN 202011092741 A CN202011092741 A CN 202011092741A CN 112269811 A CN112269811 A CN 112269811A
- Authority
- CN
- China
- Prior art keywords
- capacity
- data
- traffic
- module
- predicted
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Withdrawn
Links
- 238000000034 method Methods 0.000 title claims abstract description 32
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 73
- 238000012216 screening Methods 0.000 claims abstract description 27
- 238000004364 calculation method Methods 0.000 claims abstract description 20
- 238000004140 cleaning Methods 0.000 claims abstract description 19
- 238000012417 linear regression Methods 0.000 claims description 27
- 238000012545 processing Methods 0.000 claims description 8
- 238000013473 artificial intelligence Methods 0.000 claims description 7
- 230000001502 supplementing effect Effects 0.000 claims description 7
- 238000004880 explosion Methods 0.000 claims description 6
- PCHJSUWPFVWCPO-UHFFFAOYSA-N gold Chemical compound [Au] PCHJSUWPFVWCPO-UHFFFAOYSA-N 0.000 claims description 6
- 239000010931 gold Substances 0.000 claims description 6
- 229910052737 gold Inorganic materials 0.000 claims description 6
- 238000003066 decision tree Methods 0.000 claims description 4
- 238000007637 random forest analysis Methods 0.000 claims description 3
- 239000000126 substance Substances 0.000 claims description 2
- 238000013277 forecasting method Methods 0.000 claims 1
- 238000013480 data collection Methods 0.000 abstract description 3
- 238000005311 autocorrelation function Methods 0.000 description 12
- 238000012360 testing method Methods 0.000 description 7
- 230000006870 function Effects 0.000 description 4
- 238000007726 management method Methods 0.000 description 4
- 239000011159 matrix material Substances 0.000 description 4
- 206010035664 Pneumonia Diseases 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 3
- 230000007423 decrease Effects 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 238000007689 inspection Methods 0.000 description 3
- 230000010355 oscillation Effects 0.000 description 3
- 206010037660 Pyrexia Diseases 0.000 description 2
- 238000013016 damping Methods 0.000 description 2
- 230000001419 dependent effect Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000009499 grossing Methods 0.000 description 2
- 230000007774 longterm Effects 0.000 description 2
- 238000002360 preparation method Methods 0.000 description 2
- 238000013468 resource allocation Methods 0.000 description 2
- 230000006641 stabilisation Effects 0.000 description 2
- 238000011105 stabilization Methods 0.000 description 2
- 208000035473 Communicable disease Diseases 0.000 description 1
- 238000003648 Ljung–Box test Methods 0.000 description 1
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000015556 catabolic process Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000006731 degradation reaction Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 230000009977 dual effect Effects 0.000 description 1
- 208000015181 infectious disease Diseases 0.000 description 1
- 210000003127 knee Anatomy 0.000 description 1
- 230000004807 localization Effects 0.000 description 1
- 210000004072 lung Anatomy 0.000 description 1
- 238000001744 unit root test Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
- G06F16/2458—Special types of queries, e.g. statistical queries, fuzzy queries or distributed queries
- G06F16/2465—Query processing support for facilitating data mining operations in structured databases
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/21—Design, administration or maintenance of databases
- G06F16/215—Improving data quality; Data cleansing, e.g. de-duplication, removing invalid entries or correcting typographical errors
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
- G06F16/2458—Special types of queries, e.g. statistical queries, fuzzy queries or distributed queries
- G06F16/2477—Temporal data queries
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
Abstract
The invention provides an IT capacity prediction method and system based on traffic, wherein the prediction method and system comprises the steps of collecting traffic data through a traffic data collection module; selecting traffic data and time points through a scene screening module; collecting IT capacity resource data through an IT capacity data collection module; cleaning the data through a data cleaning and noise reduction module; predicting future traffic through a traffic prediction module and establishing an IT capacity resource consumption model; and calculating IT capacity prediction consumption data through an IT capacity resource consumption module, and performing corresponding hardware resource configuration according to the calculation data. The invention adopts an intelligent algorithm to predict the traffic trend of future IT capacity, and predicts the IT capacity requirement in a certain period of time in the future by modeling the relation between the traffic and the IT capacity resource consumption, thereby preparing sufficient resources for the IT capacity, avoiding service interruption caused by the exhaustion of IT capacity resources, and improving the IT capacity and the utilization rate of the resources.
Description
Technical Field
The invention relates to the technical field of IT (information technology) capacity resources, in particular to an IT capacity prediction method and an IT capacity prediction system based on traffic.
Background
The business of modern enterprises is usually concentrated on the IT system, and each business to be done by a client finally needs to consume IT resources, so the business volume is closely related to the consumption of the IT system. Under the condition that the service volume is continuously increased, new resource requirements are provided for the capacity of the IT system, the service volume is caused by resource consumption, and the reasonable prediction of the IT capacity can be carried out only if the increase of the service volume is clearly predicted.
The IT department of the existing enterprise generally prepares sufficient resources for future business volume increase by means of IT capacity management, and avoids business interruption caused by exhaustion of IT capacity resources. IT is common practice to monitor the consumption of IT capacity resources and build a trend model using long-term historical data to predict IT resource consumption over a future period of time. For example, patent applications of capacity prediction device based on ARIMA model and control method thereof, and patent applications of early warning system and method for resource utilization rate of virtual machine in cloud environment, etc. However, the above methods all have a problem that the future traffic increase rate cannot be predicted from the traffic point of view. If a business department needs to launch a sales promotion item at the twenty-one shopping festival, the sales promotion item will cause the business volume to increase suddenly by times at a certain time point, which causes impact on IT capacity resources, and further causes the business volume to break in a peak period, which causes irreparable loss to enterprises. But these are not included in past historical data, in which case the prediction of long-term growth trends would not be able to accurately assess and predict such growth trends.
Disclosure of Invention
The invention provides a technical scheme of an IT capacity prediction method and system based on business volume, aiming at solving the problem that the IT capacity demand cannot be predicted from the perspective of the business volume in the existing IT capacity management technology, and the technical scheme comprises the following technical characteristics in two aspects:
the invention provides an IT capacity prediction method based on traffic, which is characterized by comprising the following steps:
s1: acquiring service volume data; and collecting historical traffic data of different services.
S2: screening a service scene; screening out the service volume data and the time points which are required by meeting scene conditions according to the predicted service scene; the predicted business scene can be the scenes of a bank counter, an internet bank or a mobile phone client and the like, and can also be the number of passengers in the transportation industry and the number of patients in outpatient service or emergency treatment of a hospital; the time points can be the time points such as the release date of bank financing products, the date of the explosion of the gold or stock market, the date of the promotion of the Taobao double eleven network, the traffic and transportation peak such as national day, spring festival and the like, or the outbreak peak of flu in autumn and winter.
S3: collecting IT capacity data; historical data such as IT application systems, computer servers, network equipment, storage equipment, cluster capacity and the like related to the service are collected.
S4: screening an IT capacity scene; and screening data in a stable time period range among the application version, the system software version and the IT resource configuration.
S5: cleaning and denoising the data; cleaning all collected traffic data and IT capacity data, supplementing vacancy values, removing noise points and accurately calculating the data.
S6: predicting future traffic by adopting an intelligent algorithm; and predicting the size of the IT capacity according to the predicted future traffic volume or the size of the future traffic volume, and performing corresponding hardware configuration.
S7: establishing an IT capacity resource consumption model by adopting an intelligent algorithm; the established IT capacity resource consumption model is used for describing the relation between the business volume and the IT capacity resource consumption and calculating the IT capacity use condition of hardware resources consumed by different business volumes running on the system.
The intelligent algorithm comprises an artificial intelligence algorithm, a big data algorithm, a time series algorithm, a linear regression algorithm, a local weighted linear regression algorithm, a polynomial regression algorithm, a GBDT gradient boosting decision tree or a random forest algorithm and the like.
Furthermore, the intelligent algorithm comprises an artificial intelligence algorithm, a big data algorithm, a time series algorithm, a linear regression algorithm and a local weighted linear regression algorithm.
S8: calculating IT capacity predicted consumption data; calculating the predicted IT capacity consumption data using the predicted future traffic data and the IT capacity resource consumption model; the calculation index of the IT capacity prediction consumption data is as follows:
CPU utilization and memory utilization of the server capacity;
a storage capacity usage of the storage capacity;
bandwidth and latency of the network capacity;
the cluster capacity refers to the utilization rate of a CPU (central processing unit) of a server or a virtual machine in a cluster, the utilization rate of a memory of the server or the virtual machine in the cluster, and the storage utilization rate of the server or the virtual machine in the cluster.
Another aspect of the present invention provides a traffic-based IT capacity prediction system, comprising:
a service volume data acquisition module: the module collects historical traffic data of different services and historical data of capacities of an IT application system, a computer server, network equipment, storage equipment, a cluster and the like related to the services.
A service scene screening module: according to the predicted service scene, the service volume data and the time points required by meeting scene conditions can be screened out; wherein the content of the first and second substances,
the business scene can be the scenes of a bank counter, an internet bank or a mobile phone client and the like, and can also be the number of passengers in the transportation industry and the number of patients in outpatient service or emergency treatment of a hospital; the time points can be the time points such as the release date of bank financing products, the date of the explosion of the gold or stock market, the date of the promotion of the Taobao double eleven network, the traffic and transportation peak such as national day, spring festival and the like, or the outbreak peak of flu in autumn and winter.
An IT capacity data acquisition module; the module collects server resource data consumed by the background IT system, storage resource data, network resource data, cluster capacity resources and other historical data.
An IT capacity scene screening module; the module screens data within a time period during which the application version, the system software version, and the IT resource configuration remain stable.
A data cleaning and noise reduction module: and the module is used for cleaning all the acquired traffic data and IT capacity data by supplementing vacancy values and removing noise points.
A traffic prediction module: the module performs calculations that predict future traffic data.
An IT capacity resource consumption modeling module: the module is used for establishing an IT capacity resource consumption model, the model describes the relation between the business volume and the IT capacity resource consumption, and the IT capacity resource consumption model is used for calculating the IT capacity use condition of hardware resources consumed by different business volumes running on the system.
An IT capacity resource consumption prediction module: the module calculates the predicted IT capacity consumption data based on the predicted future traffic data and the IT capacity resource consumption model. Wherein the IT capacity includes: server capacity, storage capacity, network capacity, cluster capacity, etc.
The invention carries out corresponding reasonable hardware resource allocation according to the IT capacity predicted consumption data calculated by the IT capacity resource consumption module.
The technical scheme of the invention is that an artificial intelligence algorithm is adopted to predict the trend of the traffic in a period of time in the future, and the relationship between the IT capacity traffic and the resource consumption is modeled. The predicted future traffic is used for predicting the IT capacity requirement in a future period of time through a traffic and resource consumption model, so that sufficient resources are prepared for the IT capacity, and service interruption caused by exhaustion of IT resources is avoided. The invention explains IT capacity and resource consumption from the perspective of business, on one hand, the invention can communicate with business departments more easily, and on the other hand, the invention can position and analyze capacity problems. The increase in IT capacity may be due to increased traffic or performance problems with IT applications or system software. The invention can estimate the IT capacity change caused by service growth and can identify the IT capacity change caused by an IT application program or system software. The prediction of the IT capacity business volume is more reasonable and accurate in actual operation, and the IT capacity and the resource utilization rate are improved.
Drawings
FIG. 1 is a block flow diagram of the steps of the prediction method of the present invention;
FIG. 2 is a functional block diagram of the prediction system of the present invention;
FIG. 3 shows the traffic from 4 to 12 points in a past day;
FIG. 4 is the utilization of CPU resources for the traffic of FIG. 3;
FIG. 5 is an autocorrelation diagram of a time series of traffic of embodiment 1;
FIG. 6 is a partial autocorrelation graph of the time series of example 1 traffic;
FIG. 7 is predicted traffic volume for the future 3 hours;
fig. 8 is an IT capacity resource consumption model in the case where the super parameter K is 0.1;
fig. 9 is an IT capacity resource consumption model in the case where the super parameter K is 10;
FIG. 10 shows an IT capacity resource consumption model for the case where the hyperparameter K is 100;
FIG. 11 is a predicted required IT capacity.
Detailed Description
The following detailed description of specific embodiments of the invention refers to the accompanying drawings and the exemplary embodiments. The present invention may be variously modified and may include various embodiments. There is shown in the drawings embodiments which are presently preferred. These preferred embodiments do not limit other embodiments of the invention.
Referring to fig. 1, the present invention provides a traffic-based IT capacity prediction method, including the following steps:
s1: acquiring service volume data; and collecting historical traffic data of different services.
S2: screening a service scene; and screening out the traffic data and the time points required by meeting scene conditions according to the predicted service scene.
The module displays the property quantity characteristics of different services according to different IT capacity prediction targets, so that a user can select services and time points meeting conditions. For example, the predicted business scenario can be divided into different channels such as a bank counter, an internet bank or a mobile phone client. The number of passengers in the transportation industry, the number of patients in outpatient service or emergency service in hospitals and the like can be used. If the user wants to test whether the IT capacity of the counter channel can be supported during the peak period when the liabilities are issued, the user only needs to select the historical business volume of the counter channel when the liabilities are issued, and the business volume of other channels can be ignored.
The time point can be a working day, a release day of bank financing products, a gold or stock market fire explosion date, a Taobao Shuangelen network promotion date and other special dates, can also be a traffic and transportation peak such as national day, spring festival and the like, or a disease peak period of flu in autumn and winter, and the like, and the user can select the service and the date meeting the conditions.
S3, acquiring IT capacity data; historical data consumed by a background IT system, namely historical data of capacity resources such as IT application systems, computer servers, network equipment, storage equipment and the like related to services are collected.
S4: screening an IT capacity scene; and screening data in a stable time period range among the application system version, the system software version and the IT resource configuration.
The consumption of IT capacity is related to the traffic, but the consumption of IT capacity divided by the consumption of IT capacity of a single service is only related to the version of the IT application system and the underlying IT resources. Therefore, the embodiment screens the data of the version of the application system and the underlying IT resources within the unchanged time period range, and the data is used as the basis for analyzing the positioning and analysis of the IT capacity resources.
S5: and the data cleaning and noise reduction are realized by supplementing vacancy values to all the acquired traffic data and IT capacity data and cleaning to remove noise points so as to accurately calculate the data by adopting an intelligent algorithm.
The intelligent algorithm comprises an artificial intelligence algorithm, a big data algorithm, a time sequence algorithm, a linear regression algorithm, a local weighted linear regression algorithm, a polynomial regression algorithm, a GBDT Gradient Boosting Decision Tree algorithm (Gradient Boosting Decision Tree), a random forest and other algorithms.
S6: predicting the traffic data by adopting an intelligent algorithm; the embodiment adopts a time series algorithm to calculate the predicted traffic data.
S7: an intelligent algorithm is adopted to establish an IT capacity resource consumption model which is used for describing the relation between the business volume and the IT capacity resource consumption and calculating the use condition of the IT capacity of the hardware resource consumed by different business volumes running on the system. In the embodiment, a local weighted linear regression algorithm is adopted to perform modeling calculation of the IT capacity resource consumption model.
S8: and after the future service volume data are predicted and the IT capacity resource consumption model is established, calculating the predicted consumption data of the IT capacity through the future service volume data and the IT capacity resource consumption model.
The calculation index of the IT capacity predicted consumption data is as follows:
CPU utilization and memory utilization of server capacity; storage capacity usage of storage capacity; and predicting the IT capacity and setting corresponding hardware according to the predicted traffic data size and the predicted delay of the network capacity. The cluster capacity refers to the utilization rate of a CPU (central processing unit) of a server or a virtual machine in a cluster, the utilization rate of a memory of the server or the virtual machine in the cluster, and the storage utilization rate of the server or the virtual machine in the cluster.
The applicant believes that any business that runs on an information system requires consumption of hardware resources, and typically the hardware resources consumed by a business are considered to be the unit IT cost of the business. The unit transaction cost for completing the basic operation of the same service is invariable, and the unit hardware capacity consumed by the same service is basically constant even if the same service is operated on different hardware at different time. By the constant relation, the IT capacity of different traffic can be calculated by utilizing the CPU resource consumption of a single service. Therefore, the embodiment of the invention can perform the calculation of the model for predicting the traffic data and the IT capacity resource consumption by adopting a common artificial intelligence algorithm. If the time series algorithm is adopted to calculate the predicted traffic data, the linear regression algorithm and the local weighted linear regression algorithm are adopted to model and calculate the IT capacity resource consumption model.
The formula for modeling calculation by using the local weighted linear regression algorithm in this embodiment is as follows:
wherein S is historical traffic, R is historical IT resource consumption, and W is a weight matrix;
this matrix has only diagonal elements:
where k is a hyperparameter describing the rate at which the degree of significance decreases as the distance between the known sample and the predicted sample increases; s(i)Is the historical traffic and s' is the predicted traffic.
Where latency refers to the time it takes to make the last round trip from the network, i.e., the sum of all latencies. Latency is one of the indicators for predicting IT capacity. For example, IT is first predicted how large the transaction amount will be in the future 1 month, for example, IT is predicted that 500 transactions will be in the future 1 month, but the current system can only perform 100 transactions, and then if the system IT capacity needs to support 500 transactions, the network system needs to be expanded, and the network system needs to be reset or configured, so that the network system can have a network system that satisfies the IT capacity of 500 transactions in the future 1 month. Such as taking measures of increasing storage space, increasing servers, etc.
For another example, the utilization rate of the cluster capacity is increased, that is, the CPU utilization rate of the servers or virtual machines in the cluster is increased, the memory utilization rate of the servers or virtual machines in the cluster is increased, and the storage utilization rate of the servers or virtual machines in the cluster is increased.
Referring to fig. 2, the traffic-based IT capacity prediction system of the present embodiment includes the following modules and functions:
a service volume data acquisition module: the module collects historical traffic data of different services, namely historical data of capacities of an IT application system, a computer server, network equipment, storage equipment, a cluster and the like related to the services.
A service scene screening module: the user can screen out the service volume data and the time points required by meeting the scene conditions according to the predicted service scene. The business scene can be the bank transaction number of a bank counter, an internet bank or a mobile phone client, the passenger number of the transportation industry, or the patient number of the outpatient service of a hospital or an emergency; the time points can be working days, release days of financial products, gold or stock market fire explosion dates, Taobao Shuangelen network promotion dates and other time points, are traffic and transportation peaks such as national celebration, spring festival and the like, and can also be the outbreak peak period of flu in autumn and winter.
An IT capacity data acquisition module: the module collects server resource data consumed by a background IT system, storage resource data, network resource data, cluster capacity resources and other historical data.
An IT capacity scene screening module: the module screens data that remains within a stable time period between the application version, the system software version, and the IT resource configuration. The predicted result is affected by the upgrade of the application version, the system software (including operating system, middleware, database, etc.) version, or the configuration of the IT resources including the change of the hardware server model and the storage server model.
A data cleaning and noise reduction module: the module carries out cleaning of supplementing vacancy values and removing noise points on all collected traffic data and IT capacity data so as to ensure that accurate calculation can be carried out on the data when the calculation of future traffic is predicted.
A traffic prediction module: the module will use a time series algorithm to perform the calculation of the predicted future traffic data.
An IT capacity resource consumption modeling module: the module is used for establishing an IT capacity resource consumption model, describing the relation between the business volume and the IT capacity resource consumption, and calculating the use condition of the IT capacity of the hardware resource consumed by different business volumes running on the system. In the embodiment, a local weighted linear regression algorithm is adopted to perform modeling calculation of the IT capacity resource consumption model.
An IT capacity resource consumption prediction module: the module predicts hardware resources required to be consumed by running future traffic data on a system, such as server capacity, and stores resource data, such as storage disk capacity, according to an IT capacity resource consumption model; and calculating predicted consumption data of the IT capacity according to the network resource data such as bandwidth and time delay, and consumption data of cluster capacity resources and the like, and performing corresponding hardware resource configuration according to the predicted consumption data of the IT capacity calculated by the IT capacity resource consumption module, such as configuration of hardware resources such as servers or storage disks and the like.
Example 1
Referring to fig. 1 and 2, the enterprise predicts the traffic and IT resource consumption of the next day by using the traffic and IT resource consumption of the past day, and performs the following steps:
s1, collecting the historical traffic data of the service during the commissioning, as shown in fig. 3, the traffic from 4 o 'clock to 12 o' clock in the past day. Specific values refer to table 1 below:
| time of history | Historical traffic/pen transactions |
| 4:11 | 470 |
| 4:12 | 429 |
| 4:13 | 471 |
| 4:14 | 464 |
| 4:15 | 499 |
| … | … |
S2, screening a service scene; since the task is to evaluate future traffic, no special scenarios such as promotions etc. are involved, and therefore no screening is needed.
S3, collecting IT capacity data during test operation; the method comprises the steps of collecting data such as server resources, storage resources and network resources consumed by a background IT system.
In this embodiment, only the percentage of CPU resource usage of the mysql database is selected as an example, as shown in FIG. 4. Specific values refer to table 2 below:
and step S4, IT capacity scene screening. In the time period of the embodiment, the application version, the system software version and the IT resource configuration are not changed, because the IT capacity does not need to be screened;
s5, cleaning and denoising all data collected in the S2 business scene and S4 IT capacity scene screening step; so as to facilitate the following intelligent algorithm to perform accurate calculation. Since the values shown in the above tables are all present and have no significant noise, the present embodiment does not require cleaning and noise reduction.
And step S6, traffic prediction is carried out by using an ARIMA time series algorithm, and the traffic S' of the next day in the future is predicted.
The ARIMA Model is called an Autoregressive Integrated Moving Average Model (ARIMA), and is a famous Time-series prediction method proposed by bosch (Box) and Jenkins (Jenkins) in the beginning of the 70 s, so the ARIMA Model is also called a Box-Jenkins Model and a bosch-Jenkins method. Wherein ARIMA (p, d, q) is called a differential autoregressive moving average model, AR is autoregressive, and p is an autoregressive term; MA is the moving average, q is the number of terms of the moving average, and d is the number of differences made when the time series becomes stationary. The ARIMA model is a model established by converting a non-stationary time sequence into a stationary time sequence and then regressing a dependent variable only on a hysteresis value of the dependent variable and a current value and a hysteresis value of a random error term. The ARIMA model includes a moving average process (MA), an autoregressive process (AR), an autoregressive moving average process (ARMA), and an ARIMA process depending on whether the original sequence is stationary and the part involved in the regression.
Prediction algorithm process of ARIMA:
1) checking time series
Firstly, whether a time sequence is a pure random sequence (also called a white noise sequence) is identified according to an LB test (Ljung-Box test), and if the time sequence is the pure random sequence, an auto. (if the p-value of LB test is less than 0.05, it is not a purely random sequence).
② identifying the stationarity of the time sequence according to ADF Test (unit root Test). The smoothness of the p-value is judged by ADF inspection, and if the p-value is less than 0.05, the time sequence is considered to be smooth.
2) And smoothing the non-stationary time series data by d times of difference until the P value of the processed time series is less than 0.05 after the processed time series is subjected to ADF inspection.
3) After the stabilization, the prediction model is determined and the values of the correlation parameters are estimated by whether the autocorrelation function and the partial autocorrelation function are truncated or trailing, as shown in the relationship between the model and the function in table 3 below.
Table 3 is a model and function relationship table:
| model (model) | AR(p) | MA(q) | ARMA(p,q) |
| Autocorrelation function | Tailing, exponential damping or oscillation | Finite length, truncation (q order) | Trailing, exponentially decaying, or oscillating of order q |
| Partial autocorrelation function | Finite length, truncation (p order) | Tailing, exponential damping or oscillation | p-order tailing, exponential decay, or oscillation |
If the partial autocorrelation function is truncated and the autocorrelation function is trailing, an AR model (ARIMA (p, d,0)) is established;
if the partial autocorrelation function is trailing and the autocorrelation function is truncated, establishing an MA model (ARIMA (0, d, q));
and establishing an ARIMA model (ARIMA (p, d, q)) if the partial autocorrelation function and the autocorrelation function are both trailing.
4) After the model is determined and p and q parameters are obtained through prediction, different ARIMA (p, d and q) models are synthesized according to the parameters, the AIC values of the ARIMA (p, d and q) models are compared according to an AIC rule (akachi pool information criterion), and the model with the minimum AIC value is taken as a final model.
5) And (5) checking whether the residual sequence of the final model is a pure random sequence, and if so, determining that the model is qualified. Otherwise, adjusting the number p of autoregressive terms and the number q of moving average terms until a qualified ARIMA prediction model is obtained.
6) And (3) taking the data acquired by the traffic data acquisition module as prediction sample data, predicting by using the ARIMA (p, d, q) model tested in the step 5, and generating traffic prediction data of 3 hours in the next day.
According to the ARIMA algorithm:
1) the time series was examined, and the results are shown in table 4:
| ADF test results | Test Statistic | -4.15823 |
| P value of t statistic | p-value | 0.000775 |
| Number of delay steps | #Lags Used | 24 |
| Number of observed values | Number of Observations Used | 1403 |
| Critical ADF verified values at 99% confidence intervals | Critical Value(1％) | -3.43502 |
| Critical ADF verified values at 95% confidence intervals | Critical Value(5％) | -2.8636 |
| Critical ADF checked values at 90% confidence interval | Critical Value(10％) | -2.56787 |
The time series is considered to be smooth if the p-value is less than 0.05. From table 4 it can be seen that the p value is 0.000775, so this embodiment satisfies the condition that the ARIMA algorithm must be smooth for the time series of traffic.
2) And carrying out smoothing processing on the non-stationary time series data by d times of difference. Until the P value of the processed time series is less than 0.05 after ADF inspection. Since the time series is smooth, the difference processing is not required.
3) After the stabilization processing, judging a prediction model and estimating the value of the related parameter according to whether the autocorrelation function and the partial autocorrelation function are truncated or trailing.
First the autocorrelation graphs of the stationary time series are examined as shown in fig. 5 and the partial autocorrelation graphs are examined as shown in fig. 6. The auto-correlation and the partial phase coefficient are both found to have the characteristic of tailing, and both have obvious first-order correlation, so the parameters p and q of the model are set to be 1 and 1 respectively. Data fitting was performed using the ARIMA model below. No ARIMA (time series of traffic data, order ═ 1,0,1)) is used for fitting here. As shown in fig. 7, the solid line on the left of the time axis represents historical traffic, and the dashed line on the right represents traffic predicted by the model ARIMA at 3 hours on the next day. The predicted traffic values are shown with reference to table 5 below:
| time of day in the future | Predicted traffic/pen transactions |
| 3:06 | 477.85 |
| 3:07 | 484.78 |
| 3:08 | 485.54 |
| 3:09 | 487.96 |
| 3:10 | 489.42 |
And S7, establishing an IT capacity resource consumption model by adopting a local weighted linear regression algorithm, and modeling the relation between the traffic and the IT capacity resource consumption.
For a good system, under light pressure, the resource usage increases linearly with increasing traffic, levels off after entering the heavy pressure region, but drops sharply after reaching the knee region. The nonlinear relationship cannot be well described by adopting a linear model, a straight line cannot be well fitted with all data points, and even a large error exists. In order to solve the problem of establishing a linear model in a nonlinear model, when predicting the value of a certain point, the inventor selects a point close to the point instead of all points to perform linear regression. Based on which a locally weighted regression algorithm is used.
Local Weighted Linear Regression (localization Weighted Linear Regression) is based on the idea of a non-parametric learning algorithm, so that the selection of characteristics is better, each point near a predicted point is given a certain weight, and common Linear Regression is performed on the basis of a wavelength function, so that accurate fitting of the near points can be realized, and the contribution of points far away can be ignored. Namely, the weight of the near point is large, the weight of the far point is small, k is a wavelength parameter, the speed of the weight descending along with the distance is controlled, the larger the descending is, the smaller the descending is, the more accurate the descending is, and the overfitting problem can occur.
The formula using locally weighted linear regression is as follows:
The weight matrix W has only diagonal elements,
where k is a hyperparameter describing the rate at which the degree of significance decreases as the distance of the known sample from the predicted sample increases. s(i)Is the historical traffic and s' is the predicted traffic. The core code of the locally weighted linear regression is as follows:
in this embodiment, the relationship between traffic and mysql CPU resource consumption is modeled. The modeled data is derived from the traffic volume at step S1 and mysql CPU resource usage at step S3 as described above, with reference to table 6 below.
| Time of history | Traffic/pen transactions | mysql CPU resource usage/percentage |
| 4:11 | 470 | 42.43 |
| 4:12 | 429 | 43.25 |
| 4:13 | 471 | 43 |
| 4:14 | 464 | 44.17 |
| … | … | … |
And respectively taking values of 0.1,10 and 100 for the hyper-parameter k, then calculating through the core code of the local weighted linear regression, and drawing the prediction result. As shown in fig. 8, 9 and 10 for the plots between the predicted and actual values for the locally weighted linear regression model for the cases where K is 0.1,10 and 100, respectively. As can be seen from the figure, when k is 0.1, the model fluctuates frequently and cannot be predicted well. When k is 100, the model becomes linear, but IT cannot describe the gradual saturation of IT resource consumption when mysql CPU exceeds 50% of the weight pressure. Only when k is 10, the relation between the traffic and the mysql CPU resource consumption can be accurately expressed. So, the required IT capacity resource consumption model is shown in fig. 9 when k is 10.
S8, predicting IT capacity consumption by using an IT capacity resource consumption model; the predicted traffic volume S' calculated in step S6 is substituted into the model of S7, and the required IT capacity can be predicted. As shown in table 7.
| A certain time of day in the future | Predicted traffic/penTrading | Predicted mysql cpu resource usage/percentage |
| 3:06 | 477.85 | 46.02916 |
| 3:07 | 484.78 | 46.67083 |
| 3:08 | 485.54 | 46.74164 |
| 3:09 | 487.96 | 46.96503 |
| 3:10 | 489.42 | 47.10001 |
FIG. 11 is a graphical representation of historical cpu usage and predicted cpu usage for mysql. The solid line to the left of the time axis is historical mysql cpu resource usage and the dashed line to the right is predicted future mysql cpu resource usage.
Through the steps, the IT system capacity management personnel of the business can clearly know the business volume of several hours in the future of the business and the needed IT capacity, so that the preparation of the sufficient hardware resource allocation is carried out.
Example 2
In this example, 1440 pieces of data were obtained at 24 hours per 1 day, and the tables could not be laid down, and thus the following tables were expressed by using variables.
The embodiment predicts the traffic data and the IT capacity resource demand during the double eleven shopping wild season, and adopts the following steps:
and S1, collecting and counting historical service volume data of different services through a service volume data collection module. The banking business can be divided into modes of a network outlet counter, a network bank, a mobile phone app client, an Automatic Teller Machine (ATM) and the like according to the accessed channels, and the business volume data S of the 4 channels are respectively collectedbank、Snetwork、Smobile、SATMAs shown in table 8. Traffic data was collected every 1 hour interval, from 0 to 23, for a total of 24 x 4 samples per day.
And S2, screening out the service and time meeting the requirements of various scenes through a scene screening module. The feature of the traffics such as the double eleven shopping festivals is that the peak value of the transaction is highest from 0 a.m., and then gradually decreases. This special traffic variation is not present at other times. Therefore, the twenty-one situation needs to be proposed as a special scene, and historical data of the annual traffic change of the twenty-one situation, such as traffic transaction data during the twenty-one period in the last 3 years, is screened out. Since the transaction volume at 0 am is very low for most other dates than the twenty-one term. Therefore, if the screening of the traffic and the time is not performed, the calculation of the prediction algorithm performed later is influenced, so that the transaction amount of the final prediction result at 0 point in the morning is also low, and the prediction of the IT capacity traffic is unreasonable and inaccurate. IT will also result in inaccurate predictions of IT capacity, reduced utilization of IT capacity and resources, and even interruption of traffic transactions due to exhaustion of IT capacity and resources. The present embodiment selects traffic for the time period of 11 months 1 to 11 months 11 days per year as the data input.
TABLE 8
And S3, acquiring server resources, storage resources, network resources and the like consumed by the background IT system through the IT capacity historical data acquisition module, and acquiring and recording IT capacities consumed by different channels. IT capacity may include server resources, network resources, storage resources, etc. consumed by all applications behind the channel. In this embodiment, taking the CPU capacity of the server as an example, the CPU usage is collected and recorded as follows through a command such as top of the linux system or a task manager of the windows system, as shown in table 9.
TABLE 9
And S4, screening IT capacity scenes. In the time period of this embodiment, in 12 months in 2017, the hardware device is updated, the CPU of the old model is replaced with the CPU of the new model, and the utilization rate data of the CPU of the old model has no referential significance in consideration of the difference in processing capacity between the CPUs of the new and old models. The data in 2017 would then be filtered out and the following process would only process traffic and IT capacity data in 2018 and 2019.
And S5, performing cleaning for supplementing vacancy values and removing noise points on all the acquired traffic data and IT capacity data through a data cleaning and noise reduction module, and processing vacancy values and abnormal values in the acquired data to ensure that the data is effective in prediction calculation.
And S6, predicting the traffic data by adopting an intelligent algorithm through a traffic prediction module.
The embodiment adopts a time sequence algorithm and a linear regression algorithm, and adopts a historical mean value method to predict the traffic data. And predicting the traffic of different channels on the next twenty-one day by using a time series algorithm, and predicting the traffic of the twenty-one day in 2020 by using an ARIMA model.
The traffic prediction data of each channel of the twenty-one year in the future 2020 is generated by taking the data of the twenty-one year in 2018 and 2019, which are acquired by the traffic data acquisition module, as prediction sample data through calculation of an ARIMA algorithm, as shown in the following table 10.
TABLE 102020 year Dual 11 traffic forecast data
S7, establishing an IT capacity resource consumption model, establishing the IT capacity resource consumption model for the relationship between the IT capacity utilization rate and the traffic through a local weighted linear regression algorithm, and calculating the unit IT capacity cost of the hardware resources required to be consumed by the operation of the traffic on a system;
in this embodiment, a local weighted linear regression algorithm is selected to establish the IT capacity resource consumption model.
And S8, predicting the consumption of the IT capacity by using the IT capacity resource consumption model. The predicted CPU usage is obtained by substituting the traffic prediction data from Table 11 at step 6 into the IT capacity resource consumption model at step 7, as shown in Table 11 below.
TABLE 11 predicted CPU usage
Through the steps, the IT system capacity management personnel of the service can clearly know the service volume of the new service in a certain day in the future and the required IT capacity, so that the preparation is made and the hardware resources are reasonably configured.
Example 3
The embodiment is that the IT system is used for dealing with infectious disease outbreak in a certain period, such as whether medical facilities of medical institutions in areas where fever patients are concentrated are fully prepared during a new crown epidemic situation, so as to ensure that the patients can be effectively treated during the whole epidemic situation. If a large number of fever patients appear in hospital outpatient service and need to shoot lung CT images to investigate new coronary pneumonia, the consumption of IT equipment resources related to CT is greatly increased, and the existing IT equipment related to CT in hospital can not meet the requirements far away. At this time, the medical administration department can predict the number of future infected persons by the calculation and prediction method according to the embodiment of the present invention based on the number of confirmed persons of pneumonia caused by new coronary pneumonia that has been confirmed in the past one week or several days, and can urgently purchase or allocate CT-related equipment of hospitals in other areas to support hospitals in epidemic situation outbreak areas by calculating and predicting the number of CT equipment required.
Example 4
The present embodiment is a location and analysis of IT capacity resource problems due to increased traffic. The increase in IT capacity may be due to increased traffic and may be caused by performance problems with IT applications or system software. The embodiment can identify the IT capacity change caused by the IT application program or system software by calculating and predicting the IT capacity change data caused by the increase of the traffic. For example, when 10 versions of an application program are produced on a certain day, the IT resource utilization rate is high because of an error of one application program. When the IT capacity algorithm prediction is carried out on the 10 application programs, if the IT resource consumption of one application program is obviously beyond the predicted value, the application program is indicated to have defects or errors. Therefore, the calculation and prediction method of the embodiment and the above embodiments can be used for quickly performing positioning analysis and emergency treatment, and solving the problem of IT capacity.
The innovation of the present invention is to predict future IT capacity from a traffic point of view, whereas the prior art point of view is to predict future IT capacity only from past IT capacity. But the increase in IT capacity may be due to increased changes in traffic and may also be due to performance degradation issues with the application system. Therefore, the invention adopts an intelligent algorithm to predict the future traffic trend, predicts the IT capacity requirement of a certain period of time in the future through a model of traffic and resource consumption by modeling the relationship between the IT capacity traffic and the resource consumption and thereby prepares sufficient resources for the IT capacity and avoids service interruption caused by the exhaustion of the IT capacity and the resources. The invention can predict the IT capacity business volume more reasonably and accurately in actual operation, thereby improving the IT capacity and the utilization rate of resources.
The above-described embodiments of the present invention are merely exemplary, and various changes may be made in these embodiments without departing from the design principle of the present invention, and these changes are also to be considered as the scope of protection of the present invention and its equivalents.
Claims (10)
1. An IT capacity prediction method based on traffic is characterized by comprising the following steps:
s1: acquiring service volume data; collecting historical traffic data of different services;
s2: screening a service scene; screening out the service volume data and the time points which are required by meeting scene conditions according to the predicted service scene;
s3: collecting IT capacity data; collecting historical data such as IT application systems, computer servers, network equipment, storage equipment, cluster capacity and the like related to services;
s4: screening an IT capacity scene; screening data in a stable time period range among the application version, the system software version and the IT resource configuration;
s5: cleaning and denoising the data;
s6: predicting future traffic by adopting an intelligent algorithm;
s7: establishing an IT capacity resource consumption model by adopting an intelligent algorithm;
s8: calculating IT capacity predicted consumption data; calculating the IT capacity predicted consumption data using the predicted future traffic data and the IT capacity resource consumption model.
2. The prediction method according to claim 1,
the predicted business scene can be the scenes of a bank counter, an internet bank or a mobile phone client and the like, and can also be the number of passengers in the transportation industry and the number of patients in outpatient service or emergency treatment of a hospital;
the time points can be the time points such as the release date of bank financing products, the date of the explosion of the gold or stock market, the date of the promotion of the Taobao double eleven network, the traffic and transportation peak such as national day, spring festival and the like, or the outbreak peak of flu in autumn and winter.
3. The prediction method of claim 1, wherein the IT capacity is predicted according to the predicted future traffic volume or the predicted future traffic volume, and corresponding hardware configuration is performed.
4. The prediction method according to claim 1, wherein the intelligent algorithm comprises an artificial intelligence algorithm, a big data algorithm, a time series algorithm and a linear regression algorithm, a locally weighted linear regression algorithm, a polynomial regression algorithm, a GBDT gradient boosting decision tree or a random forest algorithm, etc.
5. The prediction method according to claim 1 or 4, wherein the intelligent algorithm comprises an artificial intelligence algorithm, a big data algorithm, a time series algorithm, a linear regression algorithm, a locally weighted linear regression algorithm.
6. The forecasting method of claim 1 or 4, wherein the established IT capacity resource consumption model is used for describing the relationship between the traffic and the IT capacity resource consumption, and for calculating the IT capacity usage of hardware resources consumed by different traffic running on the system.
7. The prediction method of claim 1, wherein the calculated indicators of the IT capacity predicted consumption data are as follows:
CPU utilization and memory utilization of the server capacity;
a storage capacity usage of the storage capacity;
bandwidth and latency of the network capacity;
the cluster capacity refers to the utilization rate of a CPU (central processing unit) of a server or a virtual machine in a cluster, the utilization rate of a memory of the server or the virtual machine in the cluster, and the storage utilization rate of the server or the virtual machine in the cluster.
8. The prediction method of claim 1, wherein the data cleaning and denoising comprises cleaning all collected traffic data and IT capacity data, supplementing vacancy values, removing noise points, and performing accurate calculation on the data.
9. A traffic based IT capacity prediction system, the system comprising:
a service volume data acquisition module: the module collects historical traffic data of different services and historical data of capacities of an IT application system, a computer server, network equipment, storage equipment, a cluster and the like related to the services;
a service scene screening module: according to the predicted service scene, the service volume data and the time points required by meeting scene conditions can be screened out; wherein the content of the first and second substances,
the business scene can be the scenes of a bank counter, an internet bank or a mobile phone client and the like, and can also be the number of passengers in the transportation industry and the number of patients in outpatient service or emergency treatment of a hospital;
the time points can be the time points such as the release date of bank financing products, the date of the explosion of the gold or stock market, the date of the promotion of the Taobao double eleven network and the like, the traffic and transportation peaks such as national day, spring festival and the like, or the incidence peaks of flu in autumn and winter;
an IT capacity data acquisition module: the module collects server resource data consumed by a background IT system, storage resource data, network resource data, cluster capacity resources and other historical data;
an IT capacity scene screening module: the module screens data in a time period range of keeping an application version, a system software version and IT resource configuration stable;
a data cleaning and noise reduction module: the module carries out cleaning for supplementing vacancy values and removing noise points on all collected traffic data and IT capacity data;
a traffic prediction module: the module calculates the predicted future traffic data;
an IT capacity resource consumption modeling module: the module is used for establishing an IT capacity resource consumption model which describes the relationship between the business volume and the IT capacity resource consumption and is used for calculating the IT capacity use condition of hardware resources consumed by different business volumes running on the system;
an IT capacity resource consumption prediction module: the module calculates the predicted IT capacity consumption data according to the predicted future traffic data and the IT capacity resource consumption model;
wherein the IT capacity includes: server capacity, storage capacity, network capacity, cluster capacity, etc.
10. The forecasting system of claim 9, wherein the corresponding hardware resource configuration is performed according to the IT capacity forecast consumption data calculated by the IT capacity resource consumption module.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011092741.9A CN112269811A (en) | 2020-10-13 | 2020-10-13 | IT capacity prediction method and system based on traffic |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011092741.9A CN112269811A (en) | 2020-10-13 | 2020-10-13 | IT capacity prediction method and system based on traffic |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112269811A true CN112269811A (en) | 2021-01-26 |
Family
ID=74338027
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011092741.9A Withdrawn CN112269811A (en) | 2020-10-13 | 2020-10-13 | IT capacity prediction method and system based on traffic |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112269811A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113032239A (en) * | 2021-05-28 | 2021-06-25 | 北京宝兰德软件股份有限公司 | Risk prompting method and device, electronic equipment and storage medium |
| CN113344282A (en) * | 2021-06-23 | 2021-09-03 | 中国光大银行股份有限公司 | Method, system and computer readable medium for capacity data processing and allocation |
| CN114007256A (en) * | 2021-12-15 | 2022-02-01 | 中国电信股份有限公司 | Terminal device, energy-saving feedback method, non-transitory storage medium, and program product |
| CN114257527A (en) * | 2021-11-01 | 2022-03-29 | 北京思特奇信息技术股份有限公司 | Network bearing capacity estimation method |
| CN114518988A (en) * | 2022-02-10 | 2022-05-20 | 中国光大银行股份有限公司 | Resource capacity system, method of controlling the same, and computer-readable storage medium |
| WO2023024679A1 (en) * | 2021-08-27 | 2023-03-02 | 深圳前海微众银行股份有限公司 | Method and apparatus for predicting server capacity |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20120303413A1 (en) * | 2011-05-27 | 2012-11-29 | Vpisystems Inc. | Methods and systems for network traffic forecast and analysis |
| CN105120487A (en) * | 2015-09-02 | 2015-12-02 | 中国联合网络通信集团有限公司 | Forecasting method and device for business data |
| WO2016033973A1 (en) * | 2014-09-05 | 2016-03-10 | 中兴通讯股份有限公司 | Method and system for predicting resource occupancy |
| CN107784440A (en) * | 2017-10-23 | 2018-03-09 | 国网辽宁省电力有限公司 | A kind of power information system resource allocation system and method |
| CN108923996A (en) * | 2018-05-11 | 2018-11-30 | 中国银联股份有限公司 | A kind of capacity analysis method and device |
| CN110618867A (en) * | 2018-06-19 | 2019-12-27 | 北京京东尚科信息技术有限公司 | Method and device for predicting resource usage amount |
| CN111045907A (en) * | 2019-12-12 | 2020-04-21 | 苏州博纳讯动软件有限公司 | System capacity prediction method based on traffic |
-
2020
- 2020-10-13 CN CN202011092741.9A patent/CN112269811A/en not_active Withdrawn
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20120303413A1 (en) * | 2011-05-27 | 2012-11-29 | Vpisystems Inc. | Methods and systems for network traffic forecast and analysis |
| WO2016033973A1 (en) * | 2014-09-05 | 2016-03-10 | 中兴通讯股份有限公司 | Method and system for predicting resource occupancy |
| CN105120487A (en) * | 2015-09-02 | 2015-12-02 | 中国联合网络通信集团有限公司 | Forecasting method and device for business data |
| CN107784440A (en) * | 2017-10-23 | 2018-03-09 | 国网辽宁省电力有限公司 | A kind of power information system resource allocation system and method |
| CN108923996A (en) * | 2018-05-11 | 2018-11-30 | 中国银联股份有限公司 | A kind of capacity analysis method and device |
| CN110618867A (en) * | 2018-06-19 | 2019-12-27 | 北京京东尚科信息技术有限公司 | Method and device for predicting resource usage amount |
| CN111045907A (en) * | 2019-12-12 | 2020-04-21 | 苏州博纳讯动软件有限公司 | System capacity prediction method based on traffic |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113032239A (en) * | 2021-05-28 | 2021-06-25 | 北京宝兰德软件股份有限公司 | Risk prompting method and device, electronic equipment and storage medium |
| CN113344282A (en) * | 2021-06-23 | 2021-09-03 | 中国光大银行股份有限公司 | Method, system and computer readable medium for capacity data processing and allocation |
| WO2023024679A1 (en) * | 2021-08-27 | 2023-03-02 | 深圳前海微众银行股份有限公司 | Method and apparatus for predicting server capacity |
| CN114257527A (en) * | 2021-11-01 | 2022-03-29 | 北京思特奇信息技术股份有限公司 | Network bearing capacity estimation method |
| CN114257527B (en) * | 2021-11-01 | 2024-02-02 | 北京思特奇信息技术股份有限公司 | Network bearing capacity estimation method |
| CN114007256A (en) * | 2021-12-15 | 2022-02-01 | 中国电信股份有限公司 | Terminal device, energy-saving feedback method, non-transitory storage medium, and program product |
| CN114007256B (en) * | 2021-12-15 | 2024-04-09 | 中国电信股份有限公司 | Terminal device, energy-saving feedback method, non-transitory storage medium, and program product |
| CN114518988A (en) * | 2022-02-10 | 2022-05-20 | 中国光大银行股份有限公司 | Resource capacity system, method of controlling the same, and computer-readable storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112269811A (en) | IT capacity prediction method and system based on traffic | |
| CN104350471B (en) | Method and system for detecting anomalies in real-time in processing environment | |
| US8423634B2 (en) | System and method for determining cost of website performance | |
| US20080195404A1 (en) | Compliant-based service level objectives | |
| US20110188701A1 (en) | Optimizing video stream processing | |
| US9176789B2 (en) | Capacity control | |
| US10902365B2 (en) | Identifying recurring series from transactional data | |
| WO2022155740A1 (en) | Systems and methods for outlier detection of transactions | |
| JP6251383B2 (en) | Calculating the probability of a defaulting company | |
| JP2023518771A (en) | Data monitoring based on machine learning | |
| US8219368B1 (en) | Capacity modeling system | |
| WO2016106100A1 (en) | Retention risk mitigation system | |
| CN113835947B (en) | Method and system for determining abnormality cause based on abnormality recognition result | |
| CA3182205A1 (en) | Financial risk assessment | |
| KR20110114181A (en) | Loan underwriting method for improving forecasting accuracy | |
| CN110910241A (en) | Cash flow evaluation method, apparatus, server device and storage medium | |
| CN113703974A (en) | Method and device for predicting server capacity | |
| JP4610732B2 (en) | Loan management system and method | |
| CN111738868A (en) | Anti-money laundering anti-terrorist financing risk monitoring method, device, computer equipment and storage medium | |
| JPWO2018061136A1 (en) | Demand forecasting method, demand forecasting system and program thereof | |
| CN115004182A (en) | Data quantization method based on determined value and estimated value | |
| US20120029960A1 (en) | Method and system for predicting resource allocation in cohort time | |
| JP6121017B2 (en) | Financial institution management support system and program | |
| CN112418534B (en) | Method and device for predicting quantity of collected parts, electronic equipment and computer readable storage medium | |
| US11494778B2 (en) | Enhanced data security and presentation system and method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20210126 |
|
| WW01 | Invention patent application withdrawn after publication |