CN104010029A - Distributed computing environment performance predicting method based on transverse and longitudinal information integration - Google Patents
Distributed computing environment performance predicting method based on transverse and longitudinal information integration Download PDFInfo
- Publication number
- CN104010029A CN104010029A CN201410198278.4A CN201410198278A CN104010029A CN 104010029 A CN104010029 A CN 104010029A CN 201410198278 A CN201410198278 A CN 201410198278A CN 104010029 A CN104010029 A CN 104010029A
- Authority
- CN
- China
- Prior art keywords
- prediction
- time
- data
- horizontal
- longitudinal
- 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.)
- Granted
Links
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
A distributed computing environment performance predicting method based on transverse and longitudinal information integration includes the steps that a long time sequence datum is divided into a plurality of sub-time sequences equal in length, and the length of all the sub-time sequences is T; all the sub-time sequences generated previously are predicted by adopting a longitudinal prediction algorithm and an exponential type smooth curve algorithm, relevance between periods is calculated, and longitudinal prediction data are obtained and integrated; the longitudinal prediction data obtained in the previous step are predicted through the relation between two time points in one time period by adopting a transverse prediction algorithm, so that relations of the data in the periods are found to construct a relation model, all the found transverse relations existing in the time periods are stored, prediction in the periods is carried out, a prediction result is stacked in a weighted mode, and the transverse relations are adjusted to obtain a final prediction result. Accuracy and reliability of resource scheduling through a server are improved.
Description
Technical field
The information that the present invention relates to is integrated, time series forecasting field, particularly a kind of based on the Forecasting Methodology of horizontal longitudinally algorithm to server performance under distributed environment.
Background technology
In the load balancing of server, need the relevant information of various resources in acquisition system to determine whether resource can be used, then dispatching algorithm is according to determining the priority of task and distribute to their available resources the running time of the availability of resource, task etc.But along with the operation of task, the state of various resources, as cpu load, free memory, hard disk remaining space etc. can change at any time, therefore needs the prediction of server performance load balancing and the dispatching algorithm of direct server.
To server performance, prediction can be taked time series algorithm, namely makes prediction based on historical data.The easy steps of the method: 1) by time sequence period ground acquisition server performance data; 2), based on these historical datas, set up one about the relational model between server performance and time variable; 3) adopt this model to calculate corresponding server performance value of fixed time, and predicted value using this value as server performance.Use such model, can predict server performance, thereby help scheduler program Resources allocation, management role better, improve the operating efficiency of whole distributed system.
Adopt seasonal effect in time series method to predict to server performance, can predict by time series autoregression model AR model, moving average model MA model etc., these models mainly can be predicted accurately to the data of some stable states, if but data are not in stable situation, predict the outcome just not ideal enough, and these models can not well carry out long-term forecast.
Research in the past shows, server performance situation exists periodically.Server performance can be counted as multiple periodic synergistic effects little, that have different cycles, how to utilize the periodicity of data to instruct prediction, is a problem that letter is to be solved.
Summary of the invention
The present invention is directed to prior art above shortcomings, the method for server performance being carried out to long-term forecast is provided.The present invention is achieved through the following technical solutions:
Based on the laterally longitudinal integrated distributed computing environment (DCE) performance prediction method of information, comprise the following steps:
Step 1: a long time series data is cut into the Time Sub-series of multiple equal lengths, the length of each Time Sub-series is T;
Step 2: adopt longitudinal prediction algorithm to use Exponential Smoothing curved line arithmetic to predict the Time Sub-series of all generations of step 1 is predicted, calculate the relevance between cycle and cycle, obtain longitudinal prediction data and the longitudinal prediction data obtaining is integrated;
Step 3: adopt lateral prediction algorithm to predict by longitudinal prediction data that the relation between two time points obtains step 2 within the time cycle, specifically find the contact of the data in the cycle, build relational model, the horizontal relationship existing in all time cycles of finding is stored, then carry out prediction in the cycle, weighted superposition predicts the outcome, and adjusts horizontal relationship by feedback system, is finally predicted the outcome.
According to described in preferred embodiment of the present invention based on the horizontal longitudinally integrated distributed computing environment (DCE) performance prediction method of information, step 1 is specially a historical data and is converted into many group Time Sub-series according to period of time T, wherein:
Historical data is: { x
1, x
2, x
3..., x
n, length is n;
The length of each Time Sub-series is T, is cut into n/T Time Sub-series, i subsequence S
ibe expressed as { x
i1, x
i2, x
i3..., x
iT, historical data is cut into <S
1, S
2..., S
i..., S
l>, wherein L=n/T:
Wherein, x
1, x
2, x
3..., x
nhistorical data, S
itime Sub-series,
X
i1, x
i2, x
i3..., x
iTthe data of Time Sub-series after cutting.
According to described in preferred embodiment of the present invention based on the horizontal longitudinally integrated distributed computing environment (DCE) performance prediction method of information, step 2 comprises the following steps:
Step 21: use Exponential Smoothing curved line arithmetic to carry out long-term forecast to the time series of all generations of the first step, the length of prediction is T step, predicts the outcome as follows:
Wherein,
It is longitudinal prediction data;
Step 22: consolidated forecast data, step 21 prediction data point is combined to form to prediction data according to original sampled point sequence, the direction of each predictor predicts is longitudinal, the last splicing of data is horizontal, as
For horizontal prediction data, length is T.
According to described in preferred embodiment of the present invention based on the horizontal longitudinally integrated distributed computing environment (DCE) performance prediction method of information, step 3 comprises the following steps:
Step 31: calculate the horizontal relationship between data in the same cycle
Horizontal relationship refers to the relation between two time points within the time cycle, specific as follows:
Suppose that two data points are x, y, definition y=bx+a is x, the horizontal association between y, wherein, and a, b is parameter;
Suppose that two data points are x, y, definition y=bx+a is x, the horizontal association between y, wherein, and a, b is parameter;
A, b parameter directly obtains from linear regression, and linear regression method least square method formula is as follows:
Q＝(y
1-bx
1-a)
2+(y
2-bx
2-a)
2+…+(y
n-bx
n-a)
2
Q is minimized, and first derivation, show that design factor formula is as follows:
Wherein, Q is the error of predicted value and calculated value;
Whether there is horizontal relationship for two data points, by cycle training linear regression parameters and checking out at the same time, if its error is less than certain threshold values, think the horizontal relation that exists; Otherwise, if error exceedes threshold values, think that these two time points do not exist horizontal association before;
Step 32: find the horizontal relationship existing in all time cycles, and it is stored;
Step 33: weighted superposition predicts the outcome
When the predicted value of longitudinal algorithm and the predicted value of horizontal algorithm being superposeed with weighting scheme after the horizontal relationship of having searched for all existence, obtain predicting that weighted results is as follows:
Wherein:
it is the prediction weighted results of the capable j row of i;
W
1with w
2it is respectively weight horizontal and longitudinal algorithm;
it is the result of horizontal algorithm predicts;
it is longitudinal algorithm predicts result;
Step 34: data are adjusted by horizontal relationship
Horizontal relationship, not only for prediction, also, for to the last adjustment to data, supposes that it is y=ax+b that A, two time points of B exist horizontal relationship, wherein, A is corresponding to x, and B is corresponding to y, a, and b is known parameters, and the prediction data that obtains A, B through prediction is above xp, yp,
Set up cost function:
Q＝(y-y
p)
2+(x-x
p)
2＝(ax+b-y
p)
2+(x-x
p)
2
It is exactly that two data points meet horizontal relationship that Cost function minimizes, and carries out first derivation and is made as 0:
2a(aX+b-y
p)+2(X-x
p)＝0
Obtaining new prediction data point A, B is respectively X, aX+b.
According to described in preferred embodiment of the present invention based on the horizontal longitudinally integrated distributed computing environment (DCE) performance prediction method of information, also comprise:
Step 4: the laterally longitudinal prediction algorithm of dynamic integrity
Corresponding different machines different time sections, cycle time relevance be not fix situation adopt dynamically laterally longitudinally prediction algorithm, be specially:
Length Ratio is more approaching in the future to suppose prediction, suffered time cycle impact is the same, in the time of each prediction, historical data is removed to last identical prediction length, then call respectively laterally longitudinally prediction algorithm with a series of candidate time cycle and predict and calculate its error, be used for predicting using that time cycle of error minimum as the parameter of laterally longitudinal prediction algorithm, according to the different time cycles, laterally longitudinal prediction algorithm is become to multiple fallout predictors, according to predicting the outcome of last time, select the fallout predictor corresponding to time cycle of current optimum.
According to described in preferred embodiment of the present invention based on the horizontal longitudinally integrated distributed computing environment (DCE) performance prediction method of information, the candidate in step 4 is respectively 6 hours, 12 hours, 1 day and 1 week the time cycle.
The present invention is directed to the feature of server performance, by laterally longitudinal prediction algorithm, performance situation is carried out to long-term forecast.By the load data of acquisition server, a fine-grained analysis of performance data of server, be cut into multiple time series datas, there is a data cycle in each time series data, carry out long-term forecast in conjunction with laterally carrying out data with longitudinal algorithm, last integrated predicting the outcome of gathering, improves accuracy and the reliability of server to scheduling of resource.
Brief description of the drawings
Fig. 1 is horizontal vertical structure principle schematic of the present invention;
Fig. 2 is laterally longitudinal algorithm flow schematic diagram of the present invention;
Fig. 3 is weighting algorithm schematic diagram of the present invention.
Embodiment
Below with reference to accompanying drawing of the present invention; technical scheme in the embodiment of the present invention is carried out to clear, complete description and discussion; obviously; as described herein is only a part of example of the present invention; it is not whole examples; based on the embodiment in the present invention, the every other embodiment that those of ordinary skill in the art obtain under the prerequisite of not making creative work, belongs to protection scope of the present invention.
For the ease of the understanding to the embodiment of the present invention, be further explained as an example of specific embodiment example below in conjunction with accompanying drawing, and each embodiment does not form the restriction to the embodiment of the present invention.
A kind of based on laterally longitudinal integrated distributed computing environment (DCE) performance prediction method of information, a fine-grained analysis of performance data of server, be cut into multiple time series datas, there is a data cycle in each time series data, in conjunction with laterally with longitudinal algorithm, data being carried out to long-term forecast, last integrated predicting the outcome of gathering.See Fig. 1,2 schematic diagrames, wherein:
Lateral prediction algorithm: carry out Analysis server performance data from horizontal angle, calculate the relation between the data of one-period, whole data belong to periodically, and adjacent data the former can have impact to the latter, known point and the future position cycle at one time, can the value of making prediction.
Longitudinally prediction algorithm: analyze the contact in multiple cycles from regulation of longitudinal angle, calculate the relevance between cycle and cycle, between the cycle, data are interactive, thereby predict next cycle data according to cycle situation.
Concrete steps are:
S1: a long time series data is cut into the Time Sub-series of multiple equal lengths, the length of each Time Sub-series is T.
Be specially a historical data and be converted into many group time serieses according to the time cycle (T).
Historical data is: { x
1, x
2, x
3..., x
n, length is n;
The length of each Time Sub-series is T, can be cut into n/T Time Sub-series, i subsequence S
ican be expressed as { x
i1, x
i2, x
i3..., x
iT, historical data can be cut into <S
1, S
2..., S
i..., S
l>, wherein L=n/T:
Wherein x
1, x
2, x
3..., x
nhistorical data, S
isubsequence data,
X
i1, x
i2, x
i3..., x
iTdata after cutting.
Give an example: be t=1h if the former loaded sampling period is 1h, T=24h generates 24 time series cycles after this step, and the sampling period is 24h, every day 1 point, every days 2 point.。。24 of every days totally 24 time serieses.
Function Mapping is:
Input: { x
1, x
2, x
3.., x
n}
Output: { { x
i, x
i+1, x
i+2..., x
i+T| i=1,2 ... T}
More direct form (1) carrys out tracking data.
Table 1 builds many group time series schematic diagrames
About the selection of time cycle, general select 1 day or 1 week more meaningful, because no matter machine or the mankind's activity 1 day or the periodic associated property of 1 time-of-week are more intense, loadtype is specific one type relevant to network, has also further confirmed that it is rational that the time cycle is selected 1 day and 1 week.Choose the other times cycle perhaps feasible, but be and particular machine relation, applicability is little.
S2, adopt longitudinal prediction algorithm to use Exponential Smoothing curved line arithmetic to predict the Time Sub-series of all generations of step 1 is predicted, calculate the relevance between cycle and cycle, obtain longitudinal prediction data and the longitudinal prediction data obtaining is integrated.
Longitudinal prediction algorithm can be selected existing any one prediction algorithm.Because Scheme has caused in the time longitudinally predicting, the time series models of each decomposition only need be predicted little point, such as the legacy data sampling period is 1h (t=1h), be decided to be 1 week cycle time, if each time series is predicted a value, prediction length can reach 168 steps, and the algorithm that longitudinally prediction is selected as can be seen here should be considered the algorithm that short-term forecast accuracy is higher as far as possible, so just can reach reasonable effect.This algorithm has been selected generally acknowledged short-term forecast algorithm comparatively accurately---Exponential Smoothing curved line arithmetic (exponential smoothing).This algorithm has reasonable effect for the short-term forecast of any irregular data.
Particularly, step S2 comprises the following steps:
S21: use Exponential Smoothing curved line arithmetic to carry out long-term forecast to the time series of all generations of the first step, the length of prediction is T step, predicts the outcome as follows:
Wherein,
It is longitudinal prediction data;
Because data are to exist periodically, between the data of same position, there is correlation, be correlated with as first data in each cycle, therefore adopt longitudinal prediction algorithm to carry out long-term forecast, the length of prediction can be T step.
S22: consolidated forecast data, step S21 prediction data point is combined to form to prediction data according to original sampled point sequence, the direction of each predictor predicts is longitudinal, the last splicing of data is horizontal, as
For horizontal prediction data, length is T.
S3: adopt lateral prediction algorithm to predict by longitudinal prediction data that the relation between two time points obtains step 2 within the time cycle, specifically find the contact of the data in the cycle, build relational model, the horizontal relationship existing in all time cycles of finding is stored, then carry out prediction in the cycle, weighted superposition predicts the outcome, and adjusts horizontal relationship by feedback system, is finally predicted the outcome.
S31: calculate the horizontal relationship between data in the same cycle
Horizontal relationship refers to the relation between two time points (generally referring to the association of contiguous time point) within the time cycle.If the association of two time points is linear relationships, because data itself are discrete, the line of two adjacent data is exactly a broken line, and linear model is simpler than other models.So be defined as a kind of relation of linearity with regard to horizontal association at this model, specific as follows:
Suppose that two data points are x, y, definition y=bx+a is x, the horizontal association between y, wherein, and a, b is parameter;
A, b parameter directly obtains from linear regression, and linear regression method least square method formula is as follows:
Q＝(y
1-bx
1-a)
2+(y
2-bx
2-a)
2+…+(y
n-bx
n-a)
2
Q is minimized, and first derivation, show that design factor formula is as follows:
Wherein, Q is the error of predicted value and calculated value;
Whether there is horizontal relationship for two data points, by cycle training linear regression parameters and checking out at the same time, if its error is less than certain threshold values, think the horizontal relation that exists; Otherwise, if error exceedes threshold values, think that these two time points do not exist horizontal association before.
S32: find the horizontal relationship existing in all time cycles, and it is stored.
S33: weighted superposition predicts the outcome
In the time having searched for the horizontal relationship of all existence, if length of history data is not the integral multiple of time cycle just in time, unnecessary time point can be used for prediction.Put and between future position, do not have horizontal relationship when over head time, just can predict according to horizontal line association.Because longitudinal algorithm can be predicted a value, so laterally the value of algorithm predicts can superpose with weighting scheme, specific algorithm is shown in Fig. 3.
Wherein
it is the prediction weighted results of the capable j row of i;
W
1with w
2it is respectively weight horizontal and longitudinal algorithm;
it is the result of horizontal algorithm predicts;
it is longitudinal algorithm predicts result.
It should be noted that length of history data is at that time just in time that time cycle or unnecessary time point are inscrutable with there is not horizontal relationship the method for future position.The method can only be made prediction to part future position.Give an example, while adding historical data from July 1 0:00 to 12:00 on July 3, the setting-up time cycle is the words of 1 day, if the 0:00 on July 3 finds through historical data to 12:00 finally have more one section time, 12:00 exists certain associated with 13:00, can pass through 12:00 load estimation 13:00 load.If but previous step search horizontal relationship do not find 13:00 with before 0:00 to horizontal associated of 12:00,13:00 is to make prediction in this step so, specific algorithm is shown in Fig. 3.
S34: data are adjusted by horizontal relationship
Horizontal relationship is not only for prediction, also can be further to the last adjustment to data, such as knowing that historical record shows that the 8:00 (x) of every day and 8:01 (y) exist the relation of y=x+1, although data point x has been got well in prediction above
p, yp, but can adjust and make it meet this relation.The concrete derivation of equation is as follows:
(A is corresponding to x, and B is y=ax+b (a, b is known) corresponding to y) there is horizontal relationship, and xp yp while obtaining A B prediction data through prediction above to suppose two time points of A B.
Set up cost function:
Q＝(y-y
p)
2+(x-x
p)
2＝(ax+b-y
p)
2+(x-x
p)
2
It is exactly that two data points meet this relation that Cost function minimizes, and carries out first derivation and is made as 0:
2a(aX+b-y
p)+2(X-x
P)＝0
So new prediction data point AB is respectively X, aX+b
This prediction not only can for the data point of two needs prediction it, also can be with between known data point and unknown number strong point, because all horizontal relationships of finding in the first step of lateral prediction can be used.
In addition, consider different machines different time sections, cycle time, relevance was not fixed, many times the impact that is being subject to multiple time cycles in other words, be not a well selection so fix a time cycle, therefore, the present invention also comprises dynamically laterally longitudinal prediction algorithm, can reasonablely avoid this point, dynamically laterally longitudinally prediction algorithm supposes that the suffered time cycle impact of adjacent prediction length is the same.
Particularly, the present invention is further comprising the steps of:
S4: the laterally longitudinal prediction algorithm of dynamic integrity
Corresponding different machines different time sections, cycle time relevance be not fix situation adopt dynamically laterally longitudinally prediction algorithm, be specially:
Length Ratio is more approaching in the future to suppose prediction, suffered time cycle impact is the same, in the time of each prediction, historical data is removed to last identical prediction length, then call respectively laterally longitudinally prediction algorithm with a series of candidate time cycle and predict and calculate its error, be used for predicting using that time cycle of error minimum as the parameter of laterally longitudinal prediction algorithm, according to the different time cycles, laterally longitudinal prediction algorithm is become to multiple fallout predictors, according to predicting the outcome of last time, select the fallout predictor corresponding to time cycle of current optimum.
Particularly, candidate is respectively 6 hours the time cycle, and 12 hours, 1 day and 1 week.
The present invention is directed to the feature of server performance, by laterally longitudinal prediction algorithm, performance situation is carried out to long-term forecast.Be difficult to prediction for time series data and search out relativity problem, the present invention finds data dependence by data being cut into multiple subsequence data.By the correlation of this data, use respectively longitudinal algorithm and horizontal algorithm to predict data, and the result of prediction is adjusted, finally further improve accuracy and the reliability of prediction by the periodicity of dynamic adjusting data.
Disclosed is above only several specific embodiment of the present invention, but the present invention is not limited thereto, and the changes that any person skilled in the art can think of all should drop in protection scope of the present invention.
Claims (6)
1. based on the laterally longitudinal integrated distributed computing environment (DCE) performance prediction method of information, it is characterized in that, comprise the following steps:
Step 1: a long time series data is cut into the Time Sub-series of multiple equal lengths, the length of each Time Sub-series is T;
Step 2: adopt longitudinal prediction algorithm to use Exponential Smoothing curved line arithmetic to predict the Time Sub-series of all generations of step 1 is predicted, calculate the relevance between cycle and cycle, obtain longitudinal prediction data and the longitudinal prediction data obtaining is integrated;
Step 3: adopt lateral prediction algorithm to predict by longitudinal prediction data that the relation between two time points obtains step 2 within the time cycle, specifically find the contact of the data in the cycle, build relational model, the horizontal relationship existing in all time cycles of finding is stored, then carry out prediction in the cycle, weighted superposition predicts the outcome, and adjusts horizontal relationship by feedback system, is finally predicted the outcome.
2. according to claim 1ly it is characterized in that based on the horizontal longitudinally integrated distributed computing environment (DCE) performance prediction method of information, step 1 is specially a historical data and is converted into many group Time Sub-series according to period of time T, wherein:
Historical data is: { x
1, x
2, x
3..., x
n, length is n;
The length of each Time Sub-series is T, is cut into n/T Time Sub-series, i subsequence S
ibe expressed as { x
i1, x
i2, x
i3..., x
iT, historical data is cut into <S
1, S
2..., S
i..., S
l>, wherein L=n/T:
Wherein, x
1, x
2, x
3..., x
nhistorical data, S
itime Sub-series,
X
i1, x
i2, x
i3..., x
iTthe data of Time Sub-series after cutting.
3. according to claim 2ly it is characterized in that based on the horizontal longitudinally integrated distributed computing environment (DCE) performance prediction method of information, step 2 comprises the following steps:
Step 21: use Exponential Smoothing curved line arithmetic to carry out long-term forecast to the time series of all generations of the first step, the length of prediction is T step, predicts the outcome as follows:
Wherein,
it is longitudinal prediction data;
Step 22: consolidated forecast data, step 21 prediction data point is combined to form to prediction data according to original sampled point sequence, the direction of each predictor predicts is longitudinal, the last splicing of data is horizontal, as
for horizontal prediction data, length is T.
4. according to claim 3ly it is characterized in that based on the horizontal longitudinally integrated distributed computing environment (DCE) performance prediction method of information, step 3 comprises the following steps:
Step 31: calculate the horizontal relationship between data in the same cycle
Horizontal relationship refers to the relation between two time points within the time cycle, specific as follows:
Suppose that two data points are x, y, definition y=bx+a is x, the horizontal association between y, wherein, and a, b is parameter;
A, b parameter directly obtains from linear regression, and linear regression method least square method formula is as follows:
Q is minimized, and first derivation, show that design factor formula is as follows:
Wherein, Q is the error of predicted value and calculated value;
Whether there is horizontal relationship for two data points, by cycle training linear regression parameters and checking out at the same time, if its error is less than certain threshold values, think the horizontal relation that exists; Otherwise, if error exceedes threshold values, think that these two time points do not exist horizontal association before;
Step 32: find the horizontal relationship existing in all time cycles, and it is stored;
Step 33: weighted superposition predicts the outcome
When the predicted value of longitudinal algorithm and the predicted value of horizontal algorithm being superposeed with weighting scheme after the horizontal relationship of having searched for all existence, obtain predicting that weighted results is as follows:
Wherein:
it is the prediction weighted results of the capable j row of i;
with
it is respectively weight horizontal and longitudinal algorithm;
it is the result of horizontal algorithm predicts;
it is longitudinal algorithm predicts result;
Step 34: data are adjusted by horizontal relationship
Horizontal relationship, not only for prediction, also, for to the last adjustment to data, supposes that it is y=ax+b that A, two time points of B exist horizontal relationship, wherein, A is corresponding to x, and B is corresponding to y, a, and b is known parameters, and the prediction data that obtains A, B through prediction is above xp, yp,
Set up cost function:
It is exactly that two data points meet horizontal relationship that Cost function minimizes, and carries out first derivation and is made as 0:
Obtaining new prediction data point A, B is respectively X, aX+b.
5. according to claim 1ly it is characterized in that based on the horizontal longitudinally integrated distributed computing environment (DCE) performance prediction method of information, also comprise:
Step 4: the laterally longitudinal prediction algorithm of dynamic integrity
Corresponding different machines different time sections, cycle time relevance be not fix situation adopt dynamically laterally longitudinally prediction algorithm, be specially:
Length Ratio is more approaching in the future to suppose prediction, suffered time cycle impact is the same, in the time of each prediction, historical data is removed to last identical prediction length, then call respectively laterally longitudinally prediction algorithm with a series of candidate time cycle and predict and calculate its error, be used for predicting using that time cycle of error minimum as the parameter of laterally longitudinal prediction algorithm, according to the different time cycles, laterally longitudinal prediction algorithm is become to multiple fallout predictors, according to predicting the outcome of last time, select the fallout predictor corresponding to time cycle of current optimum.
6. according to claim 5ly it is characterized in that based on the horizontal longitudinally integrated distributed computing environment (DCE) performance prediction method of information, the candidate in step 4 is respectively 6 hours, 12 hours, 1 day and 1 week the time cycle.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410198278.4A CN104010029B (en) | 2014-05-12 | 2014-05-12 | DCE performance prediction method based on laterally longitudinal information integration |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410198278.4A CN104010029B (en) | 2014-05-12 | 2014-05-12 | DCE performance prediction method based on laterally longitudinal information integration |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104010029A true CN104010029A (en) | 2014-08-27 |
| CN104010029B CN104010029B (en) | 2017-09-08 |
Family
ID=51370520
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410198278.4A Active CN104010029B (en) | 2014-05-12 | 2014-05-12 | DCE performance prediction method based on laterally longitudinal information integration |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104010029B (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106201718A (en) * | 2016-07-05 | 2016-12-07 | 北京邮电大学 | A kind of cloud computing resources dynamic retractility method based on load estimation |
| CN108200473A (en) * | 2018-02-01 | 2018-06-22 | 深圳创维－Rgb电子有限公司 | Cpu power control method, intelligent TV set and storage medium |
| CN108241864A (en) * | 2016-12-26 | 2018-07-03 | 摩根士丹利服务集团有限公司 | Server performance Forecasting Methodology based on multivariable grouping |
| CN108256893A (en) * | 2016-12-29 | 2018-07-06 | 北京国双科技有限公司 | The analysis method and device of advertisement delivery effect |
| CN110928634A (en) * | 2018-09-19 | 2020-03-27 | 阿里巴巴集团控股有限公司 | Data processing method, device and equipment |
| CN116186017A (en) * | 2023-04-25 | 2023-05-30 | 蓝色火焰科技成都有限公司 | Big data collaborative supervision method and platform |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20060031046A1 (en) * | 2004-08-09 | 2006-02-09 | Asper Robert W | Finite element analysis tire footprint smoothing algorithm using multiple load cases |
| US20100076629A1 (en) * | 2008-09-05 | 2010-03-25 | Thales | System and Method for Merging Meteorological Data Predicted and Measured on an Aircraft |
| CN103544362A (en) * | 2013-11-04 | 2014-01-29 | 国网上海市电力公司 | Harmonic medium and long term prediction method based on two-dimensional curve prediction |
-
2014
- 2014-05-12 CN CN201410198278.4A patent/CN104010029B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20060031046A1 (en) * | 2004-08-09 | 2006-02-09 | Asper Robert W | Finite element analysis tire footprint smoothing algorithm using multiple load cases |
| US20100076629A1 (en) * | 2008-09-05 | 2010-03-25 | Thales | System and Method for Merging Meteorological Data Predicted and Measured on an Aircraft |
| CN103544362A (en) * | 2013-11-04 | 2014-01-29 | 国网上海市电力公司 | Harmonic medium and long term prediction method based on two-dimensional curve prediction |
Non-Patent Citations (4)
| Title |
|---|
| DINGYU YANG 等: "《A pattern fusion model for multi-step-ahead CPU load prediction》", 《THE JOURNAL OF SYSTEMS AND SOFTWARE》 * |
| 付继文: "《分布式环境中的性能预测方法》", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
| 戚成功: "《时间序列预测中提高预测准确度的几种方法》", 《预测》 * |
| 赵海青 等: "《电力负荷与电量组合式预测模型及其应用》", 《水电能源科学》 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106201718A (en) * | 2016-07-05 | 2016-12-07 | 北京邮电大学 | A kind of cloud computing resources dynamic retractility method based on load estimation |
| CN108241864A (en) * | 2016-12-26 | 2018-07-03 | 摩根士丹利服务集团有限公司 | Server performance Forecasting Methodology based on multivariable grouping |
| CN108256893A (en) * | 2016-12-29 | 2018-07-06 | 北京国双科技有限公司 | The analysis method and device of advertisement delivery effect |
| CN108200473A (en) * | 2018-02-01 | 2018-06-22 | 深圳创维－Rgb电子有限公司 | Cpu power control method, intelligent TV set and storage medium |
| CN110928634A (en) * | 2018-09-19 | 2020-03-27 | 阿里巴巴集团控股有限公司 | Data processing method, device and equipment |
| CN110928634B (en) * | 2018-09-19 | 2023-04-07 | 阿里巴巴集团控股有限公司 | Data processing method, device and equipment |
| CN116186017A (en) * | 2023-04-25 | 2023-05-30 | 蓝色火焰科技成都有限公司 | Big data collaborative supervision method and platform |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104010029B (en) | 2017-09-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104010029A (en) | Distributed computing environment performance predicting method based on transverse and longitudinal information integration | |
| CN102034350B (en) | Short-time prediction method and system of traffic flow data | |
| CN106549772B (en) | Resource prediction method, system and capacity management device | |
| US20160320825A1 (en) | System and Method for Optimizing Energy Consumption by Processors | |
| Zacheilas et al. | Elastic complex event processing exploiting prediction | |
| CN105760213B (en) | The early warning system and method for resources of virtual machine utilization rate under cloud environment | |
| JP2015076091A (en) | Time series forecasting ensemble | |
| CN103176974A (en) | Method and device used for optimizing access path in data base | |
| Zhao et al. | A data placement strategy for data-intensive scientific workflows in cloud | |
| CN106952208A (en) | Crime automatic prediction method and system | |
| CN108573323A (en) | A kind of energy Internet user electricity demand forecasting method and system | |
| CN109523066A (en) | A kind of newly-increased mobile site site selecting method of the PM2.5 based on Kriging regression | |
| Li et al. | Efficient resource scaling based on load fluctuation in edge-cloud computing environment | |
| CN109492826A (en) | A kind of information system operating status Risk Forecast Method based on machine learning | |
| CN104217091A (en) | Website page view prediction method based on historical tendency weights | |
| CN106980874B (en) | A kind of multi-time Scales dimension data fusion method towards distribution big data | |
| CN105740431A (en) | Time series data prediction method based on multi-layer period exponent hierarchy | |
| CN103679289B (en) | Methods of electric load forecasting based on multiple regression extrapolation | |
| CN113837383B (en) | Model training method and device, electronic equipment and storage medium | |
| CN108205713A (en) | A kind of region wind power prediction error distribution determination method and device | |
| CN116244069A (en) | Capacity expansion and contraction method and device, electronic equipment and readable storage medium | |
| CN103957537B (en) | Frequency spectrum sensing method and device based on channel cluster | |
| CN106779147B (en) | Power load prediction method based on self-adaptive hierarchical time sequence clustering | |
| US20220253334A1 (en) | Scheduling control apparatus, scheduling control method and program | |
| CN108241864A (en) | Server performance Forecasting Methodology based on multivariable grouping |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20200116 Address after: Room 1709, Building No. 8, Binjiang West Road, Jiangyin City, Wuxi City, Jiangsu Province Patentee after: Jiangyin Daily Information Technology Co., Ltd. Address before: 200240 Dongchuan Road, Shanghai, No. 800, No. Patentee before: Shanghai Jiaotong University |
|
| TR01 | Transfer of patent right |