CN109766234A - Disk storage capacity prediction technique based on time series models - Google Patents
Disk storage capacity prediction technique based on time series models Download PDFInfo
- Publication number
- CN109766234A CN109766234A CN201811511746.3A CN201811511746A CN109766234A CN 109766234 A CN109766234 A CN 109766234A CN 201811511746 A CN201811511746 A CN 201811511746A CN 109766234 A CN109766234 A CN 109766234A
- Authority
- CN
- China
- Prior art keywords
- data
- disk
- time series
- model
- prediction
- 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.)
- Pending
Links
Landscapes
- Debugging And Monitoring (AREA)
Abstract
The disk storage capacity prediction technique based on time series models that the invention discloses a kind of monitors the operation of storage system comprising steps of (1) establishes the database of storage disk capacity service condition, data needed for providing prediction;(2) reading database obtains historical data, carries out data processing, constructs the prediction model based on time series models, is predicted;(3) it is compared according to capacity prediction result with residual capacity, assessment judges whether to meet predetermined threshold, and judges whether to issue early warning, and operation maintenance personnel is reminded to safeguard disk system.Disk size intelligent predicting technology of the invention avoids the limitation of single method prediction result according to the different time series models of historical data feature selecting, realizes the Storage Estimation demand to different disk system, has very strong adaptability and popularity.The present invention can more accurately predict disk size in actual test, reduce the manpower and financial resources cost of system O&M.
Description
Technical field
The present invention relates to disk monitoring and disk size Predicting Technique more particularly to a kind of magnetic based on time series models
Disk capacity prediction technique.
Background technique
With the rapid development of internet big data, large-scale data center increasingly becomes the need of state's net development of company
It wants.This is just that disk management in data center causes huge challenge, the disk pipe of a data center for possessing PB grades of storages
It is very big to manage difficulty, still more current development trend is that bigger data center is gradually generating.For example, a tool
The monitoring for the data center for having 1000PB to store, can no longer meet with traditional disk monitor mode, even if traditional monitoring
Mode can monitor the state of disk, also not can guarantee the abundant efficient utilization to disk size, often result in certain disks
Long-time is idle and other disk resources expend totally, also draws high further the cost of disk, that is, causes the wave of resource
Take, be really achieved storage can not and most sufficiently, reasonably use.In information system, storage system is basic as one
Component part, the items superiority and inferiority of system parameter such as capacity, speed, safety and the comfort level of system administration, for being
Can system other parts well run conclusive influence.The important prerequisite for constructing a handy storage system, is pair
In the correct assurance of the storage demand of the user of current and future.
It is less in number of users in the past, storage size is smaller and memory system architecture it is fairly simple under conditions of, people
The demand that work point analyses user is still can reaching for a task, and has been arrived nowadays, in face of such bulky complex and is incited somebody to action
Become the storage system and user group of more bulky complex, if it has to by increasingly automated and intelligent enough technology
To be analyzed storage demand and be predicted.
In the past, the disk size prediction technique based on Historical Monitoring data can sequence according to disk Historical Monitoring data
The characteristics of change, is monitored and is serialized at the storage of disk size data and disk Historical Monitoring data analysis using disk poll
It manages and carries out disk size prediction according to data analysis result, realize that disk monitoring modular carries out the acquisition of high-performance polling data,
Disk Analysis on monitoring data module is according to the time period analyzed and processed disk monitoring data sequent data, and disk size predicts mould
Root tuber is predicted according to the disk size after data analysis result selectable completion following a period of time, to solve disk size
The problems such as early warning when will exhaust, server outage and service application are interrupted caused by avoiding because of disk size failure, into
And the utilization rate of disk is maximized, it reduces because of the inadequate bring risk of disk size and the wasting of resources.
When using disk size prediction technique based on Historical Monitoring data, since disk size depends not only on history
State, it is often more important that current system operating status can all make it simultaneously because not carrying out cleaning reconstruct to historical data
Prediction result is likely to occur relatively large deviation.
Summary of the invention
Goal of the invention: in view of the above problems, the present invention proposes that a kind of disk storage capacity based on time series models is pre-
Survey method is predicted the storage demand variation that system in future may face, to realize the supervision to entire storage environment
And early warning.
Technical solution: to achieve the purpose of the present invention, the technical scheme adopted by the invention is that: one kind being based on time series
The disk storage capacity prediction technique of model, comprising steps of
(1) database for establishing storage disk capacity service condition monitors the operation of storage system, provides needed for prediction
Data;
(2) reading database obtains historical data, carries out data processing, constructs the prediction mould based on time series models
Type is predicted;
(3) it is compared according to capacity prediction result with residual capacity, assessment judges whether to meet predetermined threshold, and judges
Early warning whether is issued, operation maintenance personnel is reminded to safeguard disk system.
Further, the step 2 includes:
(2.1) using the time of the identification number of attribute and acquisition index as condition, performance data is extracted, obtains the past
A period of time disk service condition data;
(2.2) data analysis and processing are carried out;
(2.3) using treated, data are modeled, and carry out model testing and modification;
(2.4) data prediction is carried out using the model by examining.
(2.5) 3 statistics figureofmerits: mean absolute error, root-mean-square error and average absolute percentage error, weighing apparatus are used
Measure model prediction accuracy.
Further, the step (2.2) includes:
(2.2.1) carries out periodicity analysis, the stationarity of heuristic data;
(2.2.2) data cleansing, rejects the repeated data of disk size, and using the disk size of Servers-all as
One fixed value;
(2.2.3) carries out attribute reconstruct, and tri- attribute values of NAME, TARGET_ID, ENTITY are merged, new category is constructed
Property.
Further, the step (2.3) includes:
(2.3.1) determines model used by making auto-correlation and partial correlation figure judgement data stationarity to data;
(2.3.2) if auto-correlation is hangover, AR algorithm is then used in partial correlation truncation;If auto-correlation truncation, partial correlation
MA algorithm is then used in hangover;If auto-correlation and partial correlation are all hangovers, ARMA algorithm is used;
(2.3.3) carries out the parameter Estimation of model using maximum-likelihood method, estimates the value of parameters;
(2.3.4) carries out model using BIC information criterion to determine rank for each different models, determines p, q parameter, from
And select optimal models.
The utility model has the advantages that disk size intelligent predicting technology of the invention, according to the historical data feature selecting different time
Series model avoids the limitation of single method prediction result, realizes the Storage Estimation demand to different disk system, tool
There are very strong adaptability and popularity.
The intelligentized Predicting Technique of present invention height, can more accurately predict disk size in actual test, can
It realizes making effective use of for disk system, reduces the manpower and financial resources cost of system O&M.
Specific embodiment
Below with reference to embodiment, further description of the technical solution of the present invention.
Due to disk storage capacity demand, the service history of the past period is depended not only on, is also depended on current
System running state, thus only the prediction technique based on historical data (such as front based on Historical Monitoring data
Disk size prediction technique) it is not enough to cope with disk size requirement forecasting problem, and taken on to capacity prediction methods
The prediction model with feedback modifiers function of effect.
For this two o'clock, it is based on history data in magnetic disk, the methods of cleaning reconstruct processing is carried out to data, using time series
Analytic approach constructs reasonable prediction model, the disk storage capacity prediction technique based on time series models is formed, to predict
The size of application system server disk use space, provides the early warning of customization for administrator.
Disk storage capacity prediction technique of the present invention based on time series models, may face system in future
Storage demand variation predicted, to realize supervision and early warning to entire storage environment, comprising steps of
(1) storage disk capacity is established using database, monitors the operation of storage system, to provide number needed for prediction
According to;
(2) the prediction module reading database based on time series models obtains historical data, carries out data processing, mould
Type construction, prediction;And judge whether to meet predetermined threshold by assessment;
Including four key steps:
(2.1) data acquisition;
In order to extract data in magnetic disk, with the time of identification number (TARGET_ID) and acquisition index of attribute
(COLLECTTIME) it is condition, performance data is extracted, this step mainly passes through system command and directly acquires over
A period of time disk service condition numerical value.
(2.2) data processing;
1, periodicity analysis, the stationarity of heuristic data are carried out;
The present invention is modeled using time Sequence Analysis Method, for the needs of modeling, needs the stationarity of heuristic data.It is logical
The stationarity of data can tentatively be found by crossing timing diagram, for the use size of server disk, as unit of day, to data
Visualized operation is carried out, the service condition of disk does not have periodicity under normal circumstances, they show Retarder theory increasing
It is long, Trendline is presented, preliminary to judge, data are non-stable.
2, the repeated data of disk size is rejected in data cleansing, and using the disk size of Servers-all as one
Fixed value facilitates model pre-warning;
In practical business, monitoring system timing can be collected the information of disk daily.But under normal circumstances, magnetic
The capacity attribute of disk is a definite value, therefore can have the repeated data of disk size in disk original data.In data cleansing
In the process, the repeated data of disk size is rejected, and using the disk size of Servers-all as a fixed value, it is convenient
Model pre-warning.
3, attribute reconstruct is carried out, tri- attribute values of NAME, TARGET_ID, ENTITY are merged, new attribute is constructed;
In data storage, disk size is as unit of KB.Because the disc information of every server can be by table
Tri- attributes of NAME, TARGET_ID, ENTITY distinguish, and the above three attribute of every server is constant, institute
Can merge this three attribute values, new attribute is constructed.
(2.3) Construction of A Model;
By treated, data are divided into two parts, and a part is modeling sample data, and a part is model verify data.Choosing
Last 5 data for evidence of fetching is verify data, other data are modeling sample data.
1, model selects;
Requirement due to ARIMA/ARMA model to time series is leveling style, it is therefore desirable to carry out stationary test.This
Invention determines model used by making auto-correlation and partial correlation figure judgement data stationarity to data.If auto-correlation is to drag
AR algorithm is then used in tail, partial correlation truncation;If auto-correlation truncation, MA algorithm is then used in partial correlation hangover;If auto-correlation and partially
Related is all hangover, then uses ARMA algorithm.The parameter Estimation that model is carried out using maximum-likelihood method, estimates the value of parameters.
For each different models, model is carried out using BIC information criterion to determine rank, determines p, q parameter, to select optimal models.
There is no Stochastic Trends or deterministic trend to determine in original data sequence, needs to carry out stationarity inspection to data
The phenomenon that testing, otherwise will generating " shadowing property ".The present invention carries out stationary test, such as original sequence using ADF method
Column are attributed to steadily after 1 order difference, and d value is determined as 1 at this time.
AR model is known as autoregression model (Auto Regressive model);MA model is known as moving average model(MA model)
(Moving Average model);ARMA is known as ARMA model (Auto Regressive and Moving
Average model);ARIMA model is known as difference ARMA model.
AR model
If any number of some time series can be expressed as following regression equation, which is obeyed
The autoregressive process of p rank can be expressed as AR (p):
Wherein, xt,xt-1,xt-2,……,xt-pFor different time points record index value,For
Autoregressive coefficient, utWhite noise as the time series.
The autoregressive process AR (1) of referred to as 1 rank,Referred to as 2 ranks from
Regression process AR (2).
It can be found that AR model utilizes the correlativity (auto-correlation) of numerical value early period and later period numerical value, establishing includes early period
The regression equation of numerical value and later period numerical value achievees the purpose that prediction, therefore becomes autoregressive process.Here white noise can be with
It is understood as the random fluctuation of time series numerical value, the summation of these random fluctuations can be equal to 0.
MA model
If any number of some time series can be expressed as following regression equation, which is obeyed
The moving average process of q rank can be expressed as MA (q):
Wherein, ut,ut-1,ut-2,……ut-qIndicate the white noise item of different time points, θ1,θ2,θ3,……,θqFor movement
Regression equation coefficient, xtIndicate the corresponding index value of time point t.
It can be found that the index value at some time point is equal to the weighted sum of white noise sequence, if in regression equation, it is white
Noise only has two, then the moving average process is 2 rank moving average process MA (2).Compare autoregressive process and movement is flat
Equal process solves white noise in autoregression variance it is found that moving average process can be used as the supplement of autoregressive process in fact
The combination of Solve problems, the two just becomes autoregressive moving-average (ARMA) process.
Arma modeling
ARMA model consists of two parts: autoregression part and rolling average part, therefore includes two
Order can be expressed as ARMA (p, q), and p is Autoregressive, and q is moving average order, and regression equation indicates are as follows:
From regression equation it is found that ARMA model combines the advantage of two models of AR and MA, in ARMA mould
In type, autoregressive process is responsible for quantifying the relationship between current data and Primary Stage Data, and moving average process is responsible for solving random
The Solve problems of item are changed, therefore, the model is more effectively and common.
ARIMA model
ARIMA model can be used in the analysis of nonhomogeneous nonstationary time series, and here homogeneous refers to originally unstable
Time series after d difference become stationary time series.Although many time serieses itself are unstable, pass through
After difference (index value of adjacent time point subtracts each other), the new time series of formation reforms into stationary time series.Cause
This, difference ARMA model is write as ARIMA (p, d, q).P represents Autoregressive;D represents difference number;Q generation
Table moving average order.
2, model testing;
After model determines, examine whether its residual sequence is white noise.If not white noise, illustrate in residual error there is also
Useful information needs to modify model or further extracts.
Whether it has been extracted to verify information useful in sequence and has finished, has needed to carry out white noise verification to sequence.Such as
Infructescence column verify as white noise sequence, just illustrate that information useful in sequence has been extracted and finish, remaining is random entirely
Disturbance, can not be predicted and be used.The present invention carries out white noise verification using the method for LB statistic.
(2.4) model prediction;
1, it is predicted using the model by examining, obtains following 5 days predicted values, and compared with actual value, it is preceding
Last 5 data are not used in modeling in we in face, we are given a forecast verifying with these data.
2, in order to evaluate the quality of Time series forecasting model effect, this experiment using 3 measurement model prediction accuracies statistics
Figureofmerit: mean absolute error, root-mean-square error and average absolute percentage error.Never ipsilateral reflects this 3 indexs
The precision of prediction of algorithm.
(3) it is compared according to the result that capacity prediction provides with residual capacity, makes and whether issue early warning, remind O&M people
Member carries out maintenance or dilatation to disk system.
The invention adopts the above technical scheme, can effectively solve the problem that the history-dependent problem of the information in requirement forecasting,
Disk size can be more accurately predicted in actual test.
Above-mentioned disk size intelligent predicting technology is avoided according to the different time series models of historical data feature selecting
The limitation of single method prediction result realizes the Storage Estimation demand to different disk system, has very strong adaptability
And popularity.
The intelligentized Predicting Technique of height provided by the invention, it can be achieved that disk system makes effective use of, and for because
Service delay machine caused by disk is insufficient makes early warning, greatly reduces manpower consumption, reduces the manpower and financial resources of system O&M
Cost.
Claims (5)
1. a kind of disk storage capacity prediction technique based on time series models, which is characterized in that comprising steps of
(1) database for establishing storage disk capacity service condition monitors the operation of storage system, number needed for providing prediction
According to;
(2) reading database obtains historical data, carries out data processing, constructs the prediction model based on time series models, into
Row prediction;
(3) it is compared according to capacity prediction result with residual capacity, assessment judges whether to meet predetermined threshold, and judges whether to send out
Early warning out reminds operation maintenance personnel to safeguard disk system.
2. the disk storage capacity prediction technique according to claim 1 based on time series models, which is characterized in that institute
Stating step 2 includes:
(2.1) using the time of the identification number of attribute and acquisition index as condition, performance data is extracted, obtains one section in the past
Time disk service condition data;
(2.2) data analysis and processing are carried out;
(2.3) using treated, data are modeled, and carry out model testing and modification;
(2.4) data prediction is carried out using the model by examining.
3. the disk storage capacity prediction technique according to claim 2 based on time series models, which is characterized in that also
Including step (2.5), using 3 statistics figureofmerits: mean absolute error, root-mean-square error and average absolute percentage error, weighing apparatus
Measure model prediction accuracy.
4. the disk storage capacity prediction technique according to claim 2 based on time series models, which is characterized in that institute
Stating step (2.2) includes:
(2.2.1) carries out periodicity analysis, the stationarity of heuristic data;
The repeated data of disk size is rejected in (2.2.2) data cleansing, and using the disk size of Servers-all as one
Fixed value;
(2.2.3) carries out attribute reconstruct, and tri- attribute values of NAME, TARGET_ID, ENTITY are merged, new attribute is constructed.
5. the disk storage capacity prediction technique according to claim 2 based on time series models, which is characterized in that institute
Stating step (2.3) includes:
(2.3.1) determines model used by making auto-correlation and partial correlation figure judgement data stationarity to data;
(2.3.2) if auto-correlation is hangover, AR algorithm is then used in partial correlation truncation;If auto-correlation truncation, partial correlation hangover,
Then use MA algorithm;If auto-correlation and partial correlation are all hangovers, ARMA algorithm is used;
(2.3.3) carries out the parameter Estimation of model using maximum-likelihood method, estimates the value of parameters;
(2.3.4) carries out model using BIC information criterion to determine rank, determines p, q parameter, to select for each different models
Select optimal models.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811511746.3A CN109766234A (en) | 2018-12-11 | 2018-12-11 | Disk storage capacity prediction technique based on time series models |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811511746.3A CN109766234A (en) | 2018-12-11 | 2018-12-11 | Disk storage capacity prediction technique based on time series models |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109766234A true CN109766234A (en) | 2019-05-17 |
Family
ID=66451159
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811511746.3A Pending CN109766234A (en) | 2018-12-11 | 2018-12-11 | Disk storage capacity prediction technique based on time series models |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109766234A (en) |
Cited By (18)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110321240A (en) * | 2019-06-28 | 2019-10-11 | 阿里巴巴集团控股有限公司 | A kind of business impact assessment method and apparatus based on time series forecasting |
CN110377228A (en) * | 2019-06-19 | 2019-10-25 | 深圳壹账通智能科技有限公司 | Automatic expansion method, device, O&M terminal and the storage medium of block chain node |
CN110489062A (en) * | 2019-08-27 | 2019-11-22 | 浪潮云信息技术有限公司 | A kind of disk expansion method and system based on OpenStack environment |
CN110569102A (en) * | 2019-07-22 | 2019-12-13 | 华为技术有限公司 | Method and device for deploying container instances |
CN110851333A (en) * | 2019-11-14 | 2020-02-28 | 北京金山云网络技术有限公司 | Monitoring method and device of root partition and monitoring server |
CN110865928A (en) * | 2019-11-26 | 2020-03-06 | 上海新炬网络技术有限公司 | Method for realizing capacity prediction based on ARIMA prediction model and gray prediction model |
CN111241066A (en) * | 2020-01-10 | 2020-06-05 | 平安科技(深圳)有限公司 | Automatic operation and maintenance method and device for platform database and computer readable storage medium |
CN111427753A (en) * | 2020-03-23 | 2020-07-17 | 上海新炬网络信息技术股份有限公司 | ARIMA model-based capacity prediction device and control method thereof |
CN111666196A (en) * | 2020-05-28 | 2020-09-15 | 北京思特奇信息技术股份有限公司 | Method for evaluating service cluster capacity based on ARMA model |
CN112115416A (en) * | 2020-08-06 | 2020-12-22 | 深圳市水务科技有限公司 | Predictive maintenance method, apparatus, and storage medium |
CN112148557A (en) * | 2020-09-15 | 2020-12-29 | 北京基调网络股份有限公司 | Method for predicting performance index in real time, computer equipment and storage medium |
CN112748847A (en) * | 2019-10-29 | 2021-05-04 | 伊姆西Ip控股有限责任公司 | Method, apparatus and program product for managing storage space in a storage system |
CN112988071A (en) * | 2021-03-15 | 2021-06-18 | 中国建设银行股份有限公司 | Distributed storage capacity expansion method and device, storage medium and equipment |
CN113127803A (en) * | 2019-12-30 | 2021-07-16 | ***通信集团四川有限公司 | Method and device for establishing service cluster capacity estimation model and electronic equipment |
CN113157204A (en) * | 2021-01-29 | 2021-07-23 | 杭州优云软件有限公司 | Disk capacity prediction method for identifying manual cleaning behavior based on second-order difference method |
CN113238714A (en) * | 2021-05-28 | 2021-08-10 | 广东好太太智能家居有限公司 | Disk capacity prediction method and system based on historical monitoring data and storage medium |
CN114490650A (en) * | 2022-01-17 | 2022-05-13 | 成都飞机工业(集团)有限责任公司 | System, device, equipment and storage medium for predicting database space |
CN115145494A (en) * | 2022-08-11 | 2022-10-04 | 江苏臻云技术有限公司 | Disk capacity prediction system and method based on big data time series analysis |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6928398B1 (en) * | 2000-11-09 | 2005-08-09 | Spss, Inc. | System and method for building a time series model |
CN105094698A (en) * | 2015-07-08 | 2015-11-25 | 浪潮(北京)电子信息产业有限公司 | Method for predicting disc capacity based on historical monitoring data |
WO2017031837A1 (en) * | 2015-08-25 | 2017-03-02 | 北京百度网讯科技有限公司 | Disk capacity prediction method, device and apparatus |
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 |
CN108052528A (en) * | 2017-11-09 | 2018-05-18 | 华中科技大学 | A kind of storage device sequential classification method for early warning |
-
2018
- 2018-12-11 CN CN201811511746.3A patent/CN109766234A/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6928398B1 (en) * | 2000-11-09 | 2005-08-09 | Spss, Inc. | System and method for building a time series model |
CN105094698A (en) * | 2015-07-08 | 2015-11-25 | 浪潮(北京)电子信息产业有限公司 | Method for predicting disc capacity based on historical monitoring data |
WO2017031837A1 (en) * | 2015-08-25 | 2017-03-02 | 北京百度网讯科技有限公司 | Disk capacity prediction method, device and apparatus |
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 |
CN108052528A (en) * | 2017-11-09 | 2018-05-18 | 华中科技大学 | A kind of storage device sequential classification method for early warning |
Non-Patent Citations (2)
Title |
---|
李刚: "基于时间序列的存储负载预警研究", 《智能计算机与应用》 * |
樊重俊: "《大数据分析与应用》", 31 January 2016 * |
Cited By (26)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110377228A (en) * | 2019-06-19 | 2019-10-25 | 深圳壹账通智能科技有限公司 | Automatic expansion method, device, O&M terminal and the storage medium of block chain node |
CN110321240A (en) * | 2019-06-28 | 2019-10-11 | 阿里巴巴集团控股有限公司 | A kind of business impact assessment method and apparatus based on time series forecasting |
CN110569102A (en) * | 2019-07-22 | 2019-12-13 | 华为技术有限公司 | Method and device for deploying container instances |
CN110569102B (en) * | 2019-07-22 | 2022-12-13 | 华为技术有限公司 | Method and device for deploying container instances |
CN110489062A (en) * | 2019-08-27 | 2019-11-22 | 浪潮云信息技术有限公司 | A kind of disk expansion method and system based on OpenStack environment |
CN112748847A (en) * | 2019-10-29 | 2021-05-04 | 伊姆西Ip控股有限责任公司 | Method, apparatus and program product for managing storage space in a storage system |
CN112748847B (en) * | 2019-10-29 | 2024-04-19 | 伊姆西Ip控股有限责任公司 | Method, apparatus and program product for managing storage space in a storage system |
CN110851333A (en) * | 2019-11-14 | 2020-02-28 | 北京金山云网络技术有限公司 | Monitoring method and device of root partition and monitoring server |
CN110851333B (en) * | 2019-11-14 | 2023-09-01 | 北京金山云网络技术有限公司 | Root partition monitoring method and device and monitoring server |
CN110865928B (en) * | 2019-11-26 | 2024-04-26 | 上海新炬网络技术有限公司 | Method for realizing capacity prediction based on ARIMA prediction model and gray prediction model |
CN110865928A (en) * | 2019-11-26 | 2020-03-06 | 上海新炬网络技术有限公司 | Method for realizing capacity prediction based on ARIMA prediction model and gray prediction model |
CN113127803A (en) * | 2019-12-30 | 2021-07-16 | ***通信集团四川有限公司 | Method and device for establishing service cluster capacity estimation model and electronic equipment |
CN111241066A (en) * | 2020-01-10 | 2020-06-05 | 平安科技(深圳)有限公司 | Automatic operation and maintenance method and device for platform database and computer readable storage medium |
CN111427753A (en) * | 2020-03-23 | 2020-07-17 | 上海新炬网络信息技术股份有限公司 | ARIMA model-based capacity prediction device and control method thereof |
CN111427753B (en) * | 2020-03-23 | 2024-04-23 | 上海新炬网络信息技术股份有限公司 | Capacity prediction device based on ARIMA model and control method thereof |
CN111666196A (en) * | 2020-05-28 | 2020-09-15 | 北京思特奇信息技术股份有限公司 | Method for evaluating service cluster capacity based on ARMA model |
CN112115416A (en) * | 2020-08-06 | 2020-12-22 | 深圳市水务科技有限公司 | Predictive maintenance method, apparatus, and storage medium |
CN112148557A (en) * | 2020-09-15 | 2020-12-29 | 北京基调网络股份有限公司 | Method for predicting performance index in real time, computer equipment and storage medium |
CN112148557B (en) * | 2020-09-15 | 2021-10-08 | 北京基调网络股份有限公司 | Method for predicting performance index in real time, computer equipment and storage medium |
CN113157204A (en) * | 2021-01-29 | 2021-07-23 | 杭州优云软件有限公司 | Disk capacity prediction method for identifying manual cleaning behavior based on second-order difference method |
CN112988071A (en) * | 2021-03-15 | 2021-06-18 | 中国建设银行股份有限公司 | Distributed storage capacity expansion method and device, storage medium and equipment |
CN112988071B (en) * | 2021-03-15 | 2024-05-24 | 中国建设银行股份有限公司 | Distributed storage capacity expansion method, device, storage medium and equipment |
CN113238714A (en) * | 2021-05-28 | 2021-08-10 | 广东好太太智能家居有限公司 | Disk capacity prediction method and system based on historical monitoring data and storage medium |
CN114490650A (en) * | 2022-01-17 | 2022-05-13 | 成都飞机工业(集团)有限责任公司 | System, device, equipment and storage medium for predicting database space |
CN115145494B (en) * | 2022-08-11 | 2023-09-15 | 江苏臻云技术有限公司 | Disk capacity prediction system and method based on big data time sequence analysis |
CN115145494A (en) * | 2022-08-11 | 2022-10-04 | 江苏臻云技术有限公司 | Disk capacity prediction system and method based on big data time series analysis |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109766234A (en) | Disk storage capacity prediction technique based on time series models | |
CN108052528B (en) | A kind of storage equipment timing classification method for early warning | |
Arlitt et al. | Iotabench: an internet of things analytics benchmark | |
CN107967485B (en) | Fault analysis method and device for electricity metering equipment | |
Wang et al. | Estimating the completion time of crowdsourced tasks using survival analysis models | |
Zhang et al. | Resource requests prediction in the cloud computing environment with a deep belief network | |
CN104516808B (en) | Data prediction device and method | |
CN101408769B (en) | On-line energy forecasting system and method based on product ARIMA model | |
US20150302432A1 (en) | Classifying, Clustering, and Grouping Demand Series | |
CN104572734A (en) | Question recommendation method, device and system | |
CN112183990B (en) | Self-adaptive auditing monitoring management platform and method based on big data machine learning | |
CN109583491A (en) | A kind of shared bicycle intelligent dispatching method | |
CN109767032A (en) | A kind of business finance operation digital management optimization system based on data analysis | |
CN110457867A (en) | A kind of time series data based on machine learning is filled up and restoring method | |
CN114493049A (en) | Production line optimization method and device based on digital twin, electronic equipment and medium | |
CN111737099B (en) | Data center anomaly detection method and device based on Gaussian distribution | |
JP7304698B2 (en) | Water demand forecasting method and system | |
CN117853158A (en) | Enterprise operation data prediction method and device based on dynamic quantity benefit analysis | |
CN107480703A (en) | Transaction fault detection method and device | |
JP7062505B2 (en) | Equipment management support system | |
CN107194529B (en) | Power distribution network reliability economic benefit analysis method and device based on mining technology | |
CN113763181A (en) | Risk pressure test system | |
CN111488284A (en) | Simulation operation active detection method for OpenStack cloud platform | |
CN110148022A (en) | A kind of method of single probability under real-time forecast updating user | |
CN114968744B (en) | Implementation method and system based on financial industry capacity management prediction analysis AI algorithm |
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 | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190517 |