CN102129397A - Method and system for predicating self-adaptive disk array failure - Google Patents
Method and system for predicating self-adaptive disk array failure Download PDFInfo
- Publication number
- CN102129397A CN102129397A CN 201010611191 CN201010611191A CN102129397A CN 102129397 A CN102129397 A CN 102129397A CN 201010611191 CN201010611191 CN 201010611191 CN 201010611191 A CN201010611191 A CN 201010611191A CN 102129397 A CN102129397 A CN 102129397A
- Authority
- CN
- China
- Prior art keywords
- parameter
- disk array
- prediction
- module
- granularity
- 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
Images
Landscapes
- Debugging And Monitoring (AREA)
Abstract
The invention discloses a method and a system for predicating self-adaptive disk array failure, which is characterized in that: based on a back-propagation neutral network, historical data of a parameter capable of characterizing the health state of disk array is collected by means of self-learning mechanism of the neutral network, so the rule characterizing the variation of the parameter along with time can be summarized by an approximation method, further predicating the value of the parameter. In the method, three parameter collecting policies from low level to high level, i.e. the first, second and third parameter collecting policies, are set in total, the number of the parameters increases gradually as the time granularity decreases gradually, and a universal calculation model for calculating the health state of disk array according to predicated parameter values and parameter failure thresholds are provided, therefore, the failure condition of disk array can be effectively predicated. The self-adaptive disk array failure predicating system comprises an information acquisition module, a self-adaptive configuration module, a data storage module, a self-adaptive index predication module, a parameter failure threshold module, a user management module and a user interface module.
Description
Technical field
The present invention relates to technical field of computer network management, be specifically related to a kind of implementation method and device of disk array failure prediction system.
Background technology
Along with constantly coming into operation of mainframe computer infosystem, the mass data storage system has obtained application more and more widely, wherein the most important thing is disc array system (RAID, RedundantArray ofIndependent Disks).RAID is put forward in its paper by the D.A.Patterson of University of California-Berkeley professor, preserve means as a kind of data, its effect provides in the private server and to insert a plurality of disks when (specially referring to hard disk), forms a vast capacity, response speed is fast, reliability is high storage subsystem in the disk array mode.
By deblocking technology and redundancy scheme, disc array system has possessed certain data reliability guarantee, but this security mechanism is limited: specific to the disk array of a certain operation system, often the same model that the service property (quality) consistance is close and batch disk, after long-term the use, when a disk breaks down, the possibility that other disk breaks down is also just very big, if the polylith disk recurs fault in the short time, then rely on the reliability mechanisms of disc array system self can't restore data.
Owing to often storing very important business datum on the disk array,, will cause loss difficult to the appraisal in case expendable bust occurs.The online monitoring system of disk array operation at present often can only detect and alarm the disk array fault that has taken place, and further, people wish and can predict the contingent fault of disc array system.The Chinese patent of application number 200510088506.3 " failure prediction method of disk set and use the disk set of this method " and the Chinese patent " device, the system and method for prediction memory device fault " of application number 200710004243.2 are set forth this respectively.In traditional disk failure Forecasting Methodology, it mainly is failure monitoring index by real-time collection disk system, and by comparing with the prediction fault indices threshold value of presetting, when the index of real-time collection surpasses prediction fault indices threshold value, by carrying out juggling after certain analytical calculation.Mainly there are the following problems for traditional disk array failure prediction method:
(1) the parameter acquisition mode lacks adaptivity.Often just fixing several failure prediction parameters are gathered and calculated, so can only carry out information acquisition and failure prediction at specific disc array system.
For the short system of the many collection period of acquisition parameter, often influence system performance.And for the few system of acquisition parameter, the accuracy of failure prediction is affected again.
Shortage is to the prediction in advance of fault warning time.When detecting fault parameter, send alarm often above threshold value, and for when fault parameter will surpass that threshold value does not have any prediction or precision of prediction is poor, and it is very short time of physical fault to occur from disk probably when fault parameter surpasses threshold value, is unfavorable for the preparation in advance of standby redundancy.
Summary of the invention
The present invention is directed to the problems referred to above; a kind of self-adaptation disk array failure prediction system is proposed; can infer the residing health status of disk array according to the history data of disk array; and then according to the quantity and the sense cycle of the adaptive adjusting malfunction determination parameter of the health status of disk array; reach the self-adaptation granularity of disk array failure prediction is regulated, thus more efficient reliable protection data in magnetic disk more.
To achieve these goals, self-adaptation disk array failure prediction method described in the invention comprises following key step: gather the health and the performance index parameter of the disk array in a period of time and set up the historical data base of each index parameter according to thick parameter granularity, thick time granularity strategy; The historical data base of each index parameter to each index parameter that newly collects, is trained a parameter prediction network according to its historical data by Artificial Neural Network after setting up; And utilize the parameter prediction network that trains that the health and the performance index parameter of disk array are carried out initial forecast; Result and certain parameter granularity, the pairing reference threshold of certain hour granularity of initial forecast are compared, calculate and analyze according to given computation model, if it is good to judge interior health status of following a period of time of disk array, then can continue to implement prediction according to this parameter granularity, time granularity strategy; The disk array health status will begin to worsen in following a period of time if judge, and then increase the time granularity of the granularity and the shortening prediction fault of detected parameters, to obtain the more accurate failure prediction information of disc array system; Again to each index parameter that newly collects after increasing detected parameters and shortening the time granularity of predicting fault, predict network according to its historical data by Artificial Neural Network training parameter again, and utilize the new parameter prediction network that trains that the health and the performance index parameter of disk array are predicted, predicted the outcome;
Predicting the outcome once more compared with second reference threshold of the time granularity of the detected parameters granularity of corresponding increase and shortening, calculate and analyze according to given computation model, if it is good to judge interior health status of following a period of time of disk array, then can keep current detected parameters and time granularity to implement prediction; Otherwise, if when judging disc array system and being about to break down, then produce warning message, notify related personnel or external system to handle.
Aforementioned increase detected parameters granularity and the second time prediction of shortening time granularity can have repeatedly, if in prediction repeatedly once judge arbitrarily disk array in following a period of time health status good, then can keep current detected parameters and time granularity to implement prediction; The disk array health status will further worsen in following a period of time if judge, then further increase detected parameters to thin parameter granularity and shortening and predict that the time granularity of fault is to thin time granularity, to obtain the more accurate health and fitness information of disc array system; When judging disc array system and be about to break down, then produce warning message, notify related personnel or external system to handle.
Disc array system health of the present invention and performance parameter comprise data in magnetic disk reading speed, data in magnetic disk writing speed, data in magnetic disk read or write speed, mistake read rate, startup/stop number of times at least, redistribute sector number, rotation number of retries, disk calibration number of retries, ULTRA DMA parity error rate, multizone error rate etc.
The described detected parameters granularity of method of the present invention and detection time granularity, can carry out human configuration by system by the system manager and obtain.Optional 3~5 the main detected parameters of initial generally speaking parameter granularity, middle parameter granularity increases to 6~9 detected parameters, and final parameter granularity increases to 10~15 detected parameters; Initial time granularity is chosen as 1 month, and middle time granularity elected for 1 week as, and final time granularity is chosen as 1 day.Actual policy configurations should be grasped flexibly according to the actual conditions of system.Certainly, persons skilled in the art should be appreciated that these exemplary value are not to be intended to limit protection scope of the present invention.
Self-adaptation disk array failure prediction system of the present invention links to each other respectively with warning system with the external system disc array system.Failure prediction system is regularly gathered the health and the performance parameter of disk array, by the parameter prediction and the analysis-by-synthesis of internal system, sends alarm signal to outside warning system when doping the disk array fault, and system is further processed by external alarm.The disk array failure prediction system self includes the following function module:
Information acquisition module: regularly gather the health and the performance parameter information of disk array,, be loaded into unified information object data structure, make things convenient for subsequent treatment by the regular conversion of data.
The adaptive configuration module: according at present performed acquisition strategies and to the predicting the outcome of disk array health status, whether adaptive decision changes performed acquisition strategies.Specifically, promptly whether rise to the second cover strategy, perhaps whether rise to the 3rd cover strategy by the second cover strategy by the first cover strategy.
Data memory module: be responsible for various data storage in the system are advanced in the corresponding database, main data comprise the disk array health and the performance parameter of system's timing acquiring, bear the various network parameters of the artificial neural network of parameter prediction work, and the prediction fault reference threshold value of various health and performance parameter etc.
Self-adaptation index prediction module:, utilize a plurality of failure prediction neural networks of the adaptive structure of historical data of nearest a period of time at the different parameters index according to the determined acquisition strategies of system.At first set up neural network, then by adding up the historical healthy and performance index parameter of system in a period of time, input neural network is with the weights of neural network training, reach the training termination condition after, just obtained the neural network that trains at a certain index.For the neural network that trains, after the parameter for the previous period that input is gathered in real time, output obtains the parameter prediction value of next chronomere.
With reference to the fault threshold module: the threshold value of the healthy and performance parameter of reading disk array system, the system manager can reset the threshold value of each parameter by craft.
The analysis decision module: the parameter prediction value to the output of self-adaptation index prediction module is analyzed, model and correlation parameter fault threshold information according to the monitoring disk array, computation model based on system carries out comprehensive analytical calculation, draw the health status of the disc array system of monitoring, and according to next step operation of monitoring strategies decision-making that system adopted.
User management module: be responsible for authentication, have only through user's validated user of authorizing and just can consult the disk array object health status information of being monitored to the users' at different levels of system increase, deletion, mandate and user.User's role comprises system manager and domestic consumer's two-stage at least, and the information that the user need submit to includes but not limited to name, department, position, E-mail address, mobile phone, office telephone etc.
Subscriber interface module: be that user and system carry out mutual interface, the system manager carries out system configuration and system management by the system demonstration module.The system demonstration module is by the health status information of the overall situation and the development trend of disk array health status etc. in current and following a period of time of form display systems such as various forms, broken line graph, topological diagram.
Adopt technique scheme, beneficial technical effects of the present invention is: by setting up the failure prediction system at disc array system, can infer the parameter index that it is following according to the health of disk array and the historical data of performance parameter, thereby the contingent fault of disc array system is predicted in advance.Owing to adopted a series of adaptive technology, can either guarantee the performance of system, can guarantee the failure prediction precision of system preferably again, thus the data of effectively protecting on the disc array system to be stored.
Description of drawings
The invention will be further described below in conjunction with drawings and Examples, in the accompanying drawing:
Fig. 1 is the structured flowchart of self-adaptation disk array failure prediction method of the present invention and system embodiment;
Fig. 2 is the network topological diagram that is used for the artificial neural network of parameter prediction in self-adaptation disk array failure prediction method of the present invention and the system;
Fig. 3 is the process flow diagram of failure prediction method in self-adaptation disk array failure prediction method of the present invention and the system.
Embodiment
As shown in Figure 1, be the structured flowchart of self-adaptation disk array failure prediction system embodiment of the present invention.According to this block diagram, the self-adaptation disk array failure prediction system 12 in a kind of embodiment of the present invention need be connected with disc array system 11 and external alarm system 13 by the internet.Disk array failure prediction system 12 is realized disk array faults analysis and forecast functions, and this system comprises information acquisition module 121, adaptive configuration module 122, data memory module 123, self-adaptation index prediction module 124, with reference to fault threshold module 125, analysis decision module 126, user management module 127 and subscriber interface module 128.
Information acquisition module 121 provides the acquisition function of and performance parameter healthy for the disc array system run duration.Information acquisition module 121 receives the collocation strategy information from adaptive configuration module 122, and according to the collocation strategy of setting disc array system 11 is carried out information acquisition, and the information of collection is transferred to data memory module 123 and handled after normalization.
The function of adaptive configuration module 122 is to determine the information acquisition strategy of disc array system.The user that adaptive configuration module 122 receives from subscriber interface module 128 imports configuration information, and from the disk array of analysis decision module 126 prediction health status information, judgement is kept strategy at the corresponding levels or is upgraded to higher one-level strategy, and sends steering order to information acquisition module 121 according to this.
Data memory module 123 is used for the significant data in the storage system.Disk array health and the performance parameter information that provides from information acquisition module 121 is provided data memory module, the disk array and health and the performance parameter threshold information relevant that receive self-reference fault threshold module 125 to provide with its fault, the user that reception provides from user management module 127, user's group and authority information, and from other system information of subscriber interface module 128 and be stored in the database, use for self-adaptation index prediction module 124, analysis decision module 126 and subscriber interface module 128.
Self-adaptation index prediction module 124 is used to make up at the go forward side by side prediction of line parameter value of the prognoses system based on artificial neural network of each parameter.Self-adaptation index prediction module 124 receives from the neural network configuration information of data memory module 123 and disk array parameter historical data and is used to train failure prediction network at this parameter, receives the data of gathering in real time the parameter value in future is predicted.Self-adaptation index prediction module 124 prediction of output results carry out the disk array failure prediction for analysis decision module 126, will predict the outcome simultaneously to export to data memory module 123 and store.
With reference to fault threshold inquiry and the configuration information of fault threshold module 125 receptions, and export to data memory module 123 and store from subscriber interface module 128.Carry out computational analysis for analysis decision module 126 with reference to fault threshold module 125 output fault parameter reference thresholds.
Analysis decision module 126 provides calculating and the analysis for disk array future health status information, and whether decision produces warning information.Analysis decision module 126 receives from the parameter prediction information of self-adaptation index prediction module 124 with reference to the parametic fault threshold information of fault threshold module 125, after carrying out computational analysis according to given computation model, be given under the current strategies information of disc array system operation health status in following a period of time.Analysis decision module 126 output analysis results are to adaptive configuration module 122, and after doping the fault of disk array, outputting alarm information is to subscriber interface module 128, and outputting alarm information is to external alarm system 13 simultaneously.
User management module 127 provides the functions such as increase, deletion, authentication, mandate and access control for user and user's group.Service management unit 127 receives user's operational order, and shows corresponding operating result by user interface elements 128.
Subscriber interface module 128 provides the interactive interface between user and the disk array failure prediction system, and the concentrated displaying interface of disk array failure prediction state is provided.Subscriber interface module 128 receives the order input from the user, and export to user management module 127, with reference to fault threshold module 125, self-adaptation index prediction module 124, adaptive configuration module 122 and data memory module 123 to be used for the configuration of system.The data message that subscriber interface module 128 receives from data memory module 123 according to user's query requests, generates corresponding statistical report form and pattern exhibiting and gives the user.
Self-adaptation disk array failure prediction method described in the invention comprises following key step:
Gather the health of disk array and performance index parameter and store according to thick parameter granularity, thick time granularity strategy as historical data;
To each index parameter that collects, train a parameter prediction network by Artificial Neural Network according to its historical data; The parameter prediction network that utilization trains is predicted the health and the performance index parameter of disk array;
To predict the outcome with thick parameter granularity, slightly pairing first reference threshold of time granularity compares, calculate and analyze according to given computation model, if it is good to judge interior health status of following a period of time of disk array, then can continue to implement prediction according to thick parameter granularity, thick time granularity strategy; Otherwise, the disk array health status will begin to worsen in following a period of time if judge, then increase detected parameters to middle parameter granularity and shortening and predict that the time granularity of fault is to middle time granularity, to obtain the more accurate failure prediction information of disc array system;
To each index parameter that newly collects after increasing detected parameters and shortening the time granularity of predicting fault, pass through Artificial Neural Network training parameter prediction network again according to its historical data; The new parameter prediction network that utilization trains is predicted the health and the performance index parameter of disk array;
With predict the outcome with corresponding in second reference threshold of parameter granularity and middle time granularity compare, calculate and analyze according to given computation model, if it is good to judge interior health status of following a period of time of disk array, then can keep current detected parameters and time granularity to implement prediction; Otherwise, the disk array health status will further worsen in following a period of time if judge, then further increase detected parameters to thin parameter granularity and shortening and predict that the time granularity of fault is to thin time granularity, to obtain the more accurate health and fitness information of disc array system;
Each is increased detected parameters and shortens the index parameter that collects behind the time granularity of prediction fault, pass through Artificial Neural Network training parameter prediction network according to its historical data; The parameter prediction network that utilization trains is predicted the health and the performance index parameter of disk array;
To predict the outcome and the 3rd reference threshold compares, calculate the residing health status of disc array system according to given analysis of calculation models.When fine-grained detection and failure prediction show that disc array system is about to break down, then produce warning message, notify related personnel or external system to handle.
Disc array system health of the present invention and performance parameter comprise data in magnetic disk reading speed, data in magnetic disk writing speed, data in magnetic disk read or write speed, mistake read rate, startup/stop number of times at least, redistribute sector number, rotation number of retries, disk calibration number of retries, ULTRA DMA parity error rate, multizone error rate etc.
In disk array failure prediction implementation method of the present invention, its concrete computing method of the given computation model of indication are as follows, and the disk array parameter of establishing the failure prediction system collection here is Ai (i=1,, n), for parameter A i (i=1, n), its corresponding threshold be respectively Ti (i=1 ... n), the actual detected value be Vi (i=1 ..., n):
For parameter A i (i=1 ..., n), ((Vi-Ti)/Ti), wherein abs () represents to take absolute value function to calculating parameter distance rates Di=abs.Distance rates is represented the distance of disk parameter actual measured value distance parameter failure prediction threshold value, and more little then distance is near more;
Calculating comprises total distance rates D=(D1+ of all detected parameters information ... + Dn)/n.
More total distance rates D with reference to total distance rates threshold value T, (T be according to the artificial value of setting of strategy), if D>T, then disk array is in health status, if D≤T, it is tactful or send the fault pre-alarming signal then to start next cover.For different operation strategies, the value of T is inequality, generally speaking, for initial parameter granularity and time granularity, can get T=0.2, for the parameter granularity and the time granularity of centre, T=0.1 can be got,, T=0.05 can be got for final parameter granularity and time granularity.
In disk array failure prediction method of the present invention, the involved structure that utilizes Artificial Neural Network model to carry out the failure prediction network.Basic thought is the variation tendency of portraying disk array health performance parameter by BP NEURAL NETWORK learning algorithm (error backpropagation algorithm), thereby parameter is made a prediction, last again according to the predicted value of comprehensive various parameters, the fault that disk array is possible is predicted.The BP learning algorithm is a kind of learning process that supervision is arranged, it according to given (input, output) sample to learning, and by adjusting the effect that the network connection weight embodies study.With regard to whole neural network, it has two states:
The firstth, learning phase: the input that learning sample is right earlier is added in the input end of network, produces output at each layer neuron by the mode of input and excitation function along forward direction (being input layer → output layer).Then the difference of neuronic real output value of output layer and desired output reverse (being output layer → input layer) is propagated into each layer neuron, and correspondingly adjust with symbol according to the size of error and respectively to be connected weights.Till this process is performed until neural network power connected mode and produces given output result with certain precision under can given input sample condition, think that promptly learning phase finishes;
The secondth, working stage: when sample to be tested is input to the neural network input end of having succeeded in school, produce the principle of similar output according to similar input, the answer that neural network is asked in the output terminal generation by the mode of interpolation or extension is come.
As shown in Figure 2, neural network is longitudinal layered, with the input of neural network k layer i element be designated as
Output is designated as
K-1 layer i element is designated as to the connection weights of k layer j element
If the neuron excitation function is f, f can select the neuron excitation function used always here, as sigmoid function.Then the neuron excitation function calculates according to neuronic input and produces neuronic output, promptly has
Wherein
Error function is defined as the quadratic sum of each element error of output layer, promptly
In the formula,
Be the real output value of m layer (be output layer) the j element relevant with input vector X, y with weight vector W
jDesired output for this element; According to the principle that gradient descends, we wish that the change amount of each layer neuron corresponding weight value is proportional to the negative derivative of error function to these weights, promptly
In the formula, η is a constant, rule of thumb specifies.
For disk array failure prediction network, its concrete learning algorithm is as follows:
At first according to application demand, set up network topology structure, promptly three-layer network is imported single networks of exporting more;
Again with each weight w of network
I, jGive little non-zero real number value at random, set learning rate η and inertial coefficient α.
Import again p (input, output) sample centering i (i=1 ..., p) individual sample is right.Annotate: for sample to, be input as the health parameters mean value of the historical data of getting according to input number of nodes, be output as the actual health parameters mean value of next time period.As network for 5 inputs, 1 output, for being the situation of time granularity with the sky, if before having had from No. 1 to No. 31 31 days historical data altogether, then first sample is to being ((No. 1 data, No. 2 data, No. 3 data, No. 4 data, No. 5 data), No. 6 data), second sample is to being ((No. 2 data, No. 3 data, No. 4 data, No. 5 data, No. 6 data), No. 7 data), by that analogy, the 26 sample is to being ((No. 26 data, No. 27 data, No. 28 data, No. 29 data, No. 30 data), No. 31 data).
It is right to import i+1 group sample at last ..., a recycling p sample is right, until w
I, jTend towards stability constant till.
Its concrete operation algorithm is as follows:
N a certain health performance index of disk array actual measured value λ constantly before the input of network input layer
1, λ
2..., λ
n
By formula
Calculate the real output value of each layer of network element.The output valve of output layer is the predicted value λ to n+1 moment performance index
N+1
The predicted value of output is reassembled into constantly predicted value of input queue calculated for subsequent (such as with λ
2..., λ
n, λ
N+1As input queue, can dope n+2 predicted value λ constantly
N+2), by that analogy, can dope the performance number in the certain hour section of back.
As shown in Figure 2, be the network topological diagram embodiment of the artificial neural network that is used to predict in the self-adaptation index prediction module 124 among Fig. 1.The BP neural network structure that is used for failure prediction is a three-layer network structure.Input layer is got 10 neurons, its input expression disk array is pressed discrete-time series (1,2,10) the sign disk array health that collects and certain parameter of performance condition, 4 neurons are got in the middle layer, output layer is got 1 neuron, its output is illustrated in the disk array parameter of following time, promptly under thick time granularity condition, predict following 1 month parameter value by preceding 10 months historical data, under middle time granularity condition, predict the parameter value in following 1 week, under thin time granularity condition, predict following 1 day parameter value by preceding 10 days historical data by the historical data in preceding 10 weeks.
As shown in Figure 3, be the main process flow diagram of the failure prediction method of self-adaptation disk array failure prediction method of the present invention and system.This method may further comprise the steps:
Step S1: the user is provided with the monitored object of disk array failure prediction system 12 by subscriber interface module 128, and the needed various systematic parameters of configuration-system, comprise number of parameters, parameter name and predicted time granularity that each acquisition strategies comprises, the prediction fault threshold of each parameter, the configuration parameter of artificial neural network etc.
Step S2: information acquisition module 121 is according to the set strategy of system, regularly gathers the health that characterizes disc array system 11 and the parameter information of performance, the professional etiquette of the going forward side by side processing of formatting.
Step S3: the information of 123 pairs of collections of data memory module is stored, and stores in the database for subsequent treatment.
Step S4: self-adaptation index prediction module 124 is set up neural network at each parameter, and forms the prediction work network of each parameter by the historical data training.
Step S5: self-adaptation index prediction module 124 is utilized nearest historical data, by the automatic calculating of artificial neural network the parameter value of next time granularity unit is predicted.
Step S6: analysis decision module 126 comprehensive each parameter prediction value and fault reference threshold values, carry out analysis and judgement according to given computation model.
Does step S7: analysis decision module 126 judge whether the predicted value of disk array under current strategies normal? if normal, then change step S2 and proceed monitoring; Otherwise change next step.
Does step S8: analysis decision module 126 judge that current strategies is the 3rd cover strategy (fine-grained policy)? if then change step S10; Otherwise change next step.
Step S9: analysis decision module 126 notice adaptive configuration modules 122 upgrading current strategies if promptly current strategies is the first cover strategy, then are upgraded to the second cover strategy; If current strategies is the second cover strategy, then be upgraded to the 3rd cover strategy.Step S2 is changeed in the upgrading back.
Step S10: analysis decision module 126 is to subscriber interface module 128 and external alarm system 13 output disk arrays prediction fault warning information.Data acquisition and prediction process finish.
The above; only for the preferable embodiment of the present invention, but protection scope of the present invention is not limited thereto, and anyly is familiar with those skilled in the art in the technical scope that the present invention discloses; the variation that can expect easily or replacement all should be encompassed within protection scope of the present invention.Therefore, protection scope of the present invention should be as the criterion with the protection domain of claim.
Claims (5)
1. self-adaptation disk array failure prediction method is characterized in that: comprise following key step:
Gather the health and the performance index parameter of the disk array in a period of time and set up the historical data base of each index parameter according to thick parameter granularity, thick time granularity strategy;
The historical data base of each index parameter to each index parameter that newly collects, is trained a parameter prediction network according to its historical data by Artificial Neural Network after setting up; And utilize the parameter prediction network that trains that the health and the performance index parameter of disk array are carried out initial forecast;
Result and certain parameter granularity, the pairing reference threshold of certain hour granularity of initial forecast are compared, calculate and analyze according to given computation model, if it is good to judge interior health status of following a period of time of disk array, then can continue to implement prediction according to this parameter granularity, time granularity strategy; The disk array health status will begin to worsen in following a period of time if judge, and then increase the time granularity of the granularity and the shortening prediction fault of detected parameters, to obtain the more accurate failure prediction information of disc array system;
Again to each index parameter that newly collects after increasing detected parameters and shortening the time granularity of predicting fault, predict network according to its historical data by Artificial Neural Network training parameter again, and utilize the new parameter prediction network that trains that the health and the performance index parameter of disk array are predicted, predicted the outcome;
Predicting the outcome once more compared with second reference threshold of the time granularity of the detected parameters granularity of corresponding increase and shortening, calculate and analyze according to given computation model, if it is good to judge interior health status of following a period of time of disk array, then can keep current detected parameters and time granularity to implement prediction; Otherwise, if when judging disc array system and being about to break down, then produce warning message, notify related personnel or external system to handle.
2. self-adaptation disk array failure prediction method according to claim 1 is characterized in that:
Aforementioned increase detected parameters granularity and the second time prediction of shortening time granularity can have repeatedly, if in prediction repeatedly once judge arbitrarily disk array in following a period of time health status good, then can keep current detected parameters and time granularity to implement prediction; The disk array health status will further worsen in following a period of time if judge, then further increase detected parameters to thin parameter granularity and shortening and predict that the time granularity of fault is to thin time granularity, to obtain the more accurate health and fitness information of disc array system; When judging disc array system and be about to break down, then produce warning message, notify related personnel or external system to handle.
3. self-adaptation disk array failure prediction method according to claim 1, it is characterized in that: described disc array system health and performance parameter comprise data in magnetic disk reading speed, data in magnetic disk writing speed, data in magnetic disk read or write speed, mistake read rate, startup/stop number of times at least, redistribute sector number, rotation number of retries, disk calibration number of retries, ULTRA DMA parity error rate, multizone error rate etc.
4. self-adaptation disk array failure prediction method according to claim 1 is characterized in that: described detected parameters granularity and detection time granularity, can carry out human configuration by system by the system manager and obtain.
5. self-adaptation disk array failure prediction system, link to each other respectively with warning system with the external system disc array system, failure prediction system is regularly gathered the health and the performance parameter of disk array, parameter prediction and analysis-by-synthesis by internal system, send alarm signal to outside warning system when doping the disk array fault, system is further processed by external alarm; This disk array failure prediction system comprises:
Information acquisition module: regularly gather the health and the performance parameter information of disk array,, be loaded into unified information object data structure, make things convenient for subsequent treatment by the regular conversion of data;
The adaptive configuration module: according at present performed acquisition strategies and to the predicting the outcome of disk array health status, whether adaptive decision changes performed acquisition strategies;
Data memory module: be responsible for various data in the system are deposited in the corresponding database;
Self-adaptation index prediction module:, utilize a plurality of failure prediction neural networks of the adaptive structure of historical data of nearest a period of time at the different parameters index according to the determined acquisition strategies of system; Wherein at first set up neural network, then by adding up the historical healthy and performance index parameter of system in a period of time, input neural network is with the weights of neural network training, reach the training termination condition after, just obtained the neural network that trains at a certain index; For the neural network that trains, after the parameter for the previous period that input is gathered in real time, output obtains the parameter prediction value of next chronomere;
With reference to the fault threshold module: the threshold value of the healthy and performance parameter of reading disk array system;
The analysis decision module: the parameter prediction value to the output of self-adaptation index prediction module is analyzed, model and correlation parameter fault threshold information according to the monitoring disk array, computation model based on system carries out comprehensive analytical calculation, draw the health status of the disc array system of monitoring, and according to next step operation of monitoring strategies decision-making that system adopted;
User management module: be responsible for authentication, have only through user's validated user of authorizing and just can consult the disk array object health status information of being monitored to the users' at different levels of system increase, deletion, mandate and user; And,
Subscriber interface module: be that user and system carry out mutual interface, the system manager carries out system configuration and system management by the system demonstration module; The system demonstration module is by the health status information of the overall situation and the development trend of disk array health status etc. in current and following a period of time of form display systems such as various forms, broken line graph, topological diagram.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN 201010611191 CN102129397A (en) | 2010-12-29 | 2010-12-29 | Method and system for predicating self-adaptive disk array failure |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN 201010611191 CN102129397A (en) | 2010-12-29 | 2010-12-29 | Method and system for predicating self-adaptive disk array failure |
Publications (1)
Publication Number | Publication Date |
---|---|
CN102129397A true CN102129397A (en) | 2011-07-20 |
Family
ID=44267487
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN 201010611191 Pending CN102129397A (en) | 2010-12-29 | 2010-12-29 | Method and system for predicating self-adaptive disk array failure |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN102129397A (en) |
Cited By (57)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103218521A (en) * | 2013-03-25 | 2013-07-24 | 南通大学 | Equipment state dynamic self-adaptive alarm method based on hidden semi-Markov model (HSMM) |
CN104426696A (en) * | 2013-08-29 | 2015-03-18 | 深圳市腾讯计算机系统有限公司 | Fault processing method and device |
CN104461824A (en) * | 2014-12-01 | 2015-03-25 | 北京同有飞骥科技股份有限公司 | Magnetic disk health information optimal management method and device |
CN104461771A (en) * | 2014-11-03 | 2015-03-25 | 北京百度网讯科技有限公司 | Data backup processing method and device |
CN105138425A (en) * | 2015-08-13 | 2015-12-09 | 浪潮(北京)电子信息产业有限公司 | Detection method and system for health status of disk |
CN105589795A (en) * | 2014-12-31 | 2016-05-18 | 中国银联股份有限公司 | Disk failure prediction method and device based on prediction model |
WO2016101786A1 (en) * | 2014-12-25 | 2016-06-30 | 华为技术有限公司 | Method and device for predicting failure of nonvolatile storage medium |
CN104050071B (en) * | 2014-06-30 | 2017-02-08 | 苏州科达科技股份有限公司 | Storage device failure early warning method and system |
CN106528377A (en) * | 2016-11-11 | 2017-03-22 | 合肥联宝信息技术有限公司 | Solid state drive health status monitoring method and device |
CN106886481A (en) * | 2017-02-28 | 2017-06-23 | 深圳市华傲数据技术有限公司 | A kind of system health degree static analysis Forecasting Methodology and device |
CN106897178A (en) * | 2017-02-21 | 2017-06-27 | 曲阜师范大学 | A kind of slow disk detection method and system based on extreme learning machine |
WO2017129032A1 (en) * | 2016-01-29 | 2017-08-03 | 阿里巴巴集团控股有限公司 | Disk failure prediction method and apparatus |
CN107179959A (en) * | 2017-05-19 | 2017-09-19 | 郑州云海信息技术有限公司 | A kind of method, device and a kind of storage medium for predicting computer operation troubles |
CN107273273A (en) * | 2017-06-27 | 2017-10-20 | 郑州云海信息技术有限公司 | A kind of distributed type assemblies hardware fault method for early warning and system |
CN107479836A (en) * | 2017-08-29 | 2017-12-15 | 郑州云海信息技术有限公司 | Disk failure monitoring method, device and storage system |
CN107643877A (en) * | 2016-07-22 | 2018-01-30 | 中国电信股份有限公司 | Disk failure detection method and device |
CN107689948A (en) * | 2016-08-22 | 2018-02-13 | 北京深鉴科技有限公司 | Efficient data memory access managing device applied to neural network hardware acceleration system |
CN107844381A (en) * | 2016-09-21 | 2018-03-27 | 中国电信股份有限公司 | The fault handling method and device of storage system |
CN107861829A (en) * | 2017-11-06 | 2018-03-30 | 郑州云海信息技术有限公司 | A kind of method, system, device and the storage medium of disk failure detection |
CN107886992A (en) * | 2017-11-06 | 2018-04-06 | 郑州云海信息技术有限公司 | A kind of RAID method for detecting health status, system and relevant apparatus |
CN108073486A (en) * | 2017-12-28 | 2018-05-25 | 新华三大数据技术有限公司 | The Forecasting Methodology and device of a kind of hard disk failure |
CN108415819A (en) * | 2018-03-15 | 2018-08-17 | 中国人民解放军国防科技大学 | Hard disk fault tracking method and device |
CN108446734A (en) * | 2018-03-20 | 2018-08-24 | 中科边缘智慧信息科技(苏州)有限公司 | Disk failure automatic prediction method based on artificial intelligence |
CN108681496A (en) * | 2018-05-09 | 2018-10-19 | 北京奇艺世纪科技有限公司 | Prediction technique, device and the electronic equipment of disk failure |
CN109918280A (en) * | 2019-01-31 | 2019-06-21 | 北京无线电测量研究所 | A kind of health control method and device of servo-system |
CN110023764A (en) * | 2016-12-02 | 2019-07-16 | 豪夫迈·罗氏有限公司 | For analyzing the malfunction prediction of the automatic analyzer of biological sample |
CN110083470A (en) * | 2018-01-26 | 2019-08-02 | 浙江宇视科技有限公司 | Disk analysis method, apparatus and computer readable storage medium |
CN110198223A (en) * | 2018-02-27 | 2019-09-03 | 中兴通讯股份有限公司 | Network failure prediction technique, device and equipment, storage medium |
CN110326004A (en) * | 2017-02-24 | 2019-10-11 | 谷歌有限责任公司 | Use consistency of path learning training strategy neural network |
CN110474799A (en) * | 2019-07-31 | 2019-11-19 | 中国联合网络通信集团有限公司 | Fault Locating Method and device |
CN110572286A (en) * | 2019-09-04 | 2019-12-13 | 深圳职业技术学院 | early warning prototype system of network equipment |
WO2020000404A1 (en) * | 2018-06-29 | 2020-01-02 | Microsoft Technology Licensing, Llc. | Multi-factor cloud service storage device error prediction |
CN110891077A (en) * | 2018-09-07 | 2020-03-17 | 阿里巴巴集团控股有限公司 | CDN node detection method and device |
US10621486B2 (en) | 2016-08-12 | 2020-04-14 | Beijing Deephi Intelligent Technology Co., Ltd. | Method for optimizing an artificial neural network (ANN) |
CN111104293A (en) * | 2018-10-26 | 2020-05-05 | 伊姆西Ip控股有限责任公司 | Method, apparatus and computer program product for supporting disk failure prediction |
US10643124B2 (en) | 2016-08-12 | 2020-05-05 | Beijing Deephi Intelligent Technology Co., Ltd. | Method and device for quantizing complex artificial neural network |
CN111124793A (en) * | 2018-11-01 | 2020-05-08 | 中国移动通信集团浙江有限公司 | Method and system for detecting performance abnormity of disk array controller |
CN111124852A (en) * | 2019-11-21 | 2020-05-08 | 北京计算机技术及应用研究所 | Fault prediction method and system based on BMC health management module |
US10698657B2 (en) | 2016-08-12 | 2020-06-30 | Xilinx, Inc. | Hardware accelerator for compressed RNN on FPGA |
WO2020134442A1 (en) * | 2018-12-28 | 2020-07-02 | 中兴通讯股份有限公司 | Disk failure prediction method and device, and storage medium |
CN111508532A (en) * | 2019-01-31 | 2020-08-07 | 马维尔亚洲私人有限公司 | Health management for magnetic storage media |
US10762426B2 (en) | 2016-08-12 | 2020-09-01 | Beijing Deephi Intelligent Technology Co., Ltd. | Multi-iteration compression for deep neural networks |
US10802992B2 (en) | 2016-08-12 | 2020-10-13 | Xilinx Technology Beijing Limited | Combining CPU and special accelerator for implementing an artificial neural network |
US10810484B2 (en) | 2016-08-12 | 2020-10-20 | Xilinx, Inc. | Hardware accelerator for compressed GRU on FPGA |
CN111813585A (en) * | 2019-04-10 | 2020-10-23 | 伊姆西Ip控股有限责任公司 | Prediction and processing of slow discs |
US10832123B2 (en) | 2016-08-12 | 2020-11-10 | Xilinx Technology Beijing Limited | Compression of deep neural networks with proper use of mask |
CN111767162B (en) * | 2020-05-20 | 2021-02-26 | 北京大学 | Fault prediction method for hard disks of different models and electronic device |
US10936941B2 (en) | 2016-08-12 | 2021-03-02 | Xilinx, Inc. | Efficient data access control device for neural network hardware acceleration system |
CN112433919A (en) * | 2020-11-25 | 2021-03-02 | 深圳前海微众银行股份有限公司 | Information warning method, equipment and storage medium |
WO2021052168A1 (en) * | 2019-09-19 | 2021-03-25 | 中兴通讯股份有限公司 | Disk fault prediction method and apparatus, computer-readable storage medium, and server |
US10984308B2 (en) | 2016-08-12 | 2021-04-20 | Xilinx Technology Beijing Limited | Compression method for deep neural networks with load balance |
CN112753020A (en) * | 2018-09-27 | 2021-05-04 | 甲骨文国际公司 | Disk drive failure prediction using neural networks |
CN114063881A (en) * | 2020-07-31 | 2022-02-18 | 阿里巴巴集团控股有限公司 | Disk management method and device of distributed system |
CN114327241A (en) * | 2020-09-29 | 2022-04-12 | 伊姆西Ip控股有限责任公司 | Method, electronic device and computer program product for managing disk |
CN114428709A (en) * | 2022-01-17 | 2022-05-03 | 广州鲁邦通物联网科技股份有限公司 | SDS state detection method and system in cloud management platform |
US11557316B2 (en) | 2018-08-21 | 2023-01-17 | Marvell Asia Pte Ltd. | Pulse-based writing for magnetic storage media |
CN115687038A (en) * | 2022-08-23 | 2023-02-03 | 江苏臻云技术有限公司 | Hard disk failure prediction system and method based on big data |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6574754B1 (en) * | 2000-02-14 | 2003-06-03 | International Business Machines Corporation | Self-monitoring storage device using neural networks |
CN1770304A (en) * | 2004-08-02 | 2006-05-10 | 日立环球储存科技荷兰有限公司 | Failure predicting method of magnetic disk device and magnetic disk device using the same |
-
2010
- 2010-12-29 CN CN 201010611191 patent/CN102129397A/en active Pending
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6574754B1 (en) * | 2000-02-14 | 2003-06-03 | International Business Machines Corporation | Self-monitoring storage device using neural networks |
CN1770304A (en) * | 2004-08-02 | 2006-05-10 | 日立环球储存科技荷兰有限公司 | Failure predicting method of magnetic disk device and magnetic disk device using the same |
Cited By (74)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103218521A (en) * | 2013-03-25 | 2013-07-24 | 南通大学 | Equipment state dynamic self-adaptive alarm method based on hidden semi-Markov model (HSMM) |
CN104426696A (en) * | 2013-08-29 | 2015-03-18 | 深圳市腾讯计算机系统有限公司 | Fault processing method and device |
CN104426696B (en) * | 2013-08-29 | 2018-09-07 | 深圳市腾讯计算机系统有限公司 | A kind of method of troubleshooting, server and system |
CN104050071B (en) * | 2014-06-30 | 2017-02-08 | 苏州科达科技股份有限公司 | Storage device failure early warning method and system |
CN104461771A (en) * | 2014-11-03 | 2015-03-25 | 北京百度网讯科技有限公司 | Data backup processing method and device |
CN104461824A (en) * | 2014-12-01 | 2015-03-25 | 北京同有飞骥科技股份有限公司 | Magnetic disk health information optimal management method and device |
CN109933448A (en) * | 2014-12-25 | 2019-06-25 | 华为技术有限公司 | A kind of method and device that prediction non-volatile memory medium breaks down |
CN105787242A (en) * | 2014-12-25 | 2016-07-20 | 华为技术有限公司 | Method and device for predicting that failure will occur in non-volatile memory medium |
WO2016101786A1 (en) * | 2014-12-25 | 2016-06-30 | 华为技术有限公司 | Method and device for predicting failure of nonvolatile storage medium |
CN109933448B (en) * | 2014-12-25 | 2021-04-20 | 华为技术有限公司 | Method and device for predicting fault of nonvolatile storage medium |
CN105787242B (en) * | 2014-12-25 | 2019-02-26 | 华为技术有限公司 | A kind of method and device that prediction non-volatile memory medium breaks down |
WO2016107402A1 (en) * | 2014-12-31 | 2016-07-07 | 中国银联股份有限公司 | Magnetic disk fault prediction method and device based on prediction model |
CN105589795A (en) * | 2014-12-31 | 2016-05-18 | 中国银联股份有限公司 | Disk failure prediction method and device based on prediction model |
CN105138425A (en) * | 2015-08-13 | 2015-12-09 | 浪潮(北京)电子信息产业有限公司 | Detection method and system for health status of disk |
WO2017129032A1 (en) * | 2016-01-29 | 2017-08-03 | 阿里巴巴集团控股有限公司 | Disk failure prediction method and apparatus |
CN107025153A (en) * | 2016-01-29 | 2017-08-08 | 阿里巴巴集团控股有限公司 | The failure prediction method and device of disk |
CN107643877A (en) * | 2016-07-22 | 2018-01-30 | 中国电信股份有限公司 | Disk failure detection method and device |
US10802992B2 (en) | 2016-08-12 | 2020-10-13 | Xilinx Technology Beijing Limited | Combining CPU and special accelerator for implementing an artificial neural network |
US10810484B2 (en) | 2016-08-12 | 2020-10-20 | Xilinx, Inc. | Hardware accelerator for compressed GRU on FPGA |
US10832123B2 (en) | 2016-08-12 | 2020-11-10 | Xilinx Technology Beijing Limited | Compression of deep neural networks with proper use of mask |
US10762426B2 (en) | 2016-08-12 | 2020-09-01 | Beijing Deephi Intelligent Technology Co., Ltd. | Multi-iteration compression for deep neural networks |
US10698657B2 (en) | 2016-08-12 | 2020-06-30 | Xilinx, Inc. | Hardware accelerator for compressed RNN on FPGA |
US10643124B2 (en) | 2016-08-12 | 2020-05-05 | Beijing Deephi Intelligent Technology Co., Ltd. | Method and device for quantizing complex artificial neural network |
US10621486B2 (en) | 2016-08-12 | 2020-04-14 | Beijing Deephi Intelligent Technology Co., Ltd. | Method for optimizing an artificial neural network (ANN) |
US10936941B2 (en) | 2016-08-12 | 2021-03-02 | Xilinx, Inc. | Efficient data access control device for neural network hardware acceleration system |
US10984308B2 (en) | 2016-08-12 | 2021-04-20 | Xilinx Technology Beijing Limited | Compression method for deep neural networks with load balance |
CN107689948A (en) * | 2016-08-22 | 2018-02-13 | 北京深鉴科技有限公司 | Efficient data memory access managing device applied to neural network hardware acceleration system |
CN107689948B (en) * | 2016-08-22 | 2020-09-01 | 赛灵思公司 | Efficient data access management device applied to neural network hardware acceleration system |
CN107844381A (en) * | 2016-09-21 | 2018-03-27 | 中国电信股份有限公司 | The fault handling method and device of storage system |
CN106528377A (en) * | 2016-11-11 | 2017-03-22 | 合肥联宝信息技术有限公司 | Solid state drive health status monitoring method and device |
CN110023764B (en) * | 2016-12-02 | 2023-12-22 | 豪夫迈·罗氏有限公司 | Fault state prediction for an automated analyzer for analyzing biological samples |
CN110023764A (en) * | 2016-12-02 | 2019-07-16 | 豪夫迈·罗氏有限公司 | For analyzing the malfunction prediction of the automatic analyzer of biological sample |
CN106897178A (en) * | 2017-02-21 | 2017-06-27 | 曲阜师范大学 | A kind of slow disk detection method and system based on extreme learning machine |
CN106897178B (en) * | 2017-02-21 | 2020-06-23 | 曲阜师范大学 | Slow disk detection method and system based on extreme learning machine |
CN110326004A (en) * | 2017-02-24 | 2019-10-11 | 谷歌有限责任公司 | Use consistency of path learning training strategy neural network |
CN106886481A (en) * | 2017-02-28 | 2017-06-23 | 深圳市华傲数据技术有限公司 | A kind of system health degree static analysis Forecasting Methodology and device |
CN107179959A (en) * | 2017-05-19 | 2017-09-19 | 郑州云海信息技术有限公司 | A kind of method, device and a kind of storage medium for predicting computer operation troubles |
CN107273273A (en) * | 2017-06-27 | 2017-10-20 | 郑州云海信息技术有限公司 | A kind of distributed type assemblies hardware fault method for early warning and system |
CN107479836A (en) * | 2017-08-29 | 2017-12-15 | 郑州云海信息技术有限公司 | Disk failure monitoring method, device and storage system |
CN107886992A (en) * | 2017-11-06 | 2018-04-06 | 郑州云海信息技术有限公司 | A kind of RAID method for detecting health status, system and relevant apparatus |
CN107861829A (en) * | 2017-11-06 | 2018-03-30 | 郑州云海信息技术有限公司 | A kind of method, system, device and the storage medium of disk failure detection |
CN108073486A (en) * | 2017-12-28 | 2018-05-25 | 新华三大数据技术有限公司 | The Forecasting Methodology and device of a kind of hard disk failure |
CN108073486B (en) * | 2017-12-28 | 2022-05-10 | 新华三大数据技术有限公司 | Hard disk fault prediction method and device |
CN110083470A (en) * | 2018-01-26 | 2019-08-02 | 浙江宇视科技有限公司 | Disk analysis method, apparatus and computer readable storage medium |
CN110198223A (en) * | 2018-02-27 | 2019-09-03 | 中兴通讯股份有限公司 | Network failure prediction technique, device and equipment, storage medium |
CN108415819B (en) * | 2018-03-15 | 2021-05-25 | 中国人民解放军国防科技大学 | Hard disk fault tracking method and device |
CN108415819A (en) * | 2018-03-15 | 2018-08-17 | 中国人民解放军国防科技大学 | Hard disk fault tracking method and device |
CN108446734A (en) * | 2018-03-20 | 2018-08-24 | 中科边缘智慧信息科技(苏州)有限公司 | Disk failure automatic prediction method based on artificial intelligence |
CN108681496A (en) * | 2018-05-09 | 2018-10-19 | 北京奇艺世纪科技有限公司 | Prediction technique, device and the electronic equipment of disk failure |
WO2020000404A1 (en) * | 2018-06-29 | 2020-01-02 | Microsoft Technology Licensing, Llc. | Multi-factor cloud service storage device error prediction |
US11748185B2 (en) | 2018-06-29 | 2023-09-05 | Microsoft Technology Licensing, Llc | Multi-factor cloud service storage device error prediction |
US11557316B2 (en) | 2018-08-21 | 2023-01-17 | Marvell Asia Pte Ltd. | Pulse-based writing for magnetic storage media |
CN110891077A (en) * | 2018-09-07 | 2020-03-17 | 阿里巴巴集团控股有限公司 | CDN node detection method and device |
CN112753020B (en) * | 2018-09-27 | 2024-05-28 | 甲骨文国际公司 | Disk Drive Failure Prediction Using Neural Networks |
CN112753020A (en) * | 2018-09-27 | 2021-05-04 | 甲骨文国际公司 | Disk drive failure prediction using neural networks |
CN111104293A (en) * | 2018-10-26 | 2020-05-05 | 伊姆西Ip控股有限责任公司 | Method, apparatus and computer program product for supporting disk failure prediction |
CN111124793A (en) * | 2018-11-01 | 2020-05-08 | 中国移动通信集团浙江有限公司 | Method and system for detecting performance abnormity of disk array controller |
WO2020134442A1 (en) * | 2018-12-28 | 2020-07-02 | 中兴通讯股份有限公司 | Disk failure prediction method and device, and storage medium |
CN111508532A (en) * | 2019-01-31 | 2020-08-07 | 马维尔亚洲私人有限公司 | Health management for magnetic storage media |
CN109918280A (en) * | 2019-01-31 | 2019-06-21 | 北京无线电测量研究所 | A kind of health control method and device of servo-system |
CN111508532B (en) * | 2019-01-31 | 2023-11-03 | 马维尔亚洲私人有限公司 | Health management for magnetic storage media |
CN111813585A (en) * | 2019-04-10 | 2020-10-23 | 伊姆西Ip控股有限责任公司 | Prediction and processing of slow discs |
CN110474799B (en) * | 2019-07-31 | 2022-03-25 | 中国联合网络通信集团有限公司 | Fault positioning method and device |
CN110474799A (en) * | 2019-07-31 | 2019-11-19 | 中国联合网络通信集团有限公司 | Fault Locating Method and device |
CN110572286A (en) * | 2019-09-04 | 2019-12-13 | 深圳职业技术学院 | early warning prototype system of network equipment |
WO2021052168A1 (en) * | 2019-09-19 | 2021-03-25 | 中兴通讯股份有限公司 | Disk fault prediction method and apparatus, computer-readable storage medium, and server |
CN111124852A (en) * | 2019-11-21 | 2020-05-08 | 北京计算机技术及应用研究所 | Fault prediction method and system based on BMC health management module |
CN111767162B (en) * | 2020-05-20 | 2021-02-26 | 北京大学 | Fault prediction method for hard disks of different models and electronic device |
CN114063881A (en) * | 2020-07-31 | 2022-02-18 | 阿里巴巴集团控股有限公司 | Disk management method and device of distributed system |
CN114327241A (en) * | 2020-09-29 | 2022-04-12 | 伊姆西Ip控股有限责任公司 | Method, electronic device and computer program product for managing disk |
CN112433919A (en) * | 2020-11-25 | 2021-03-02 | 深圳前海微众银行股份有限公司 | Information warning method, equipment and storage medium |
CN114428709A (en) * | 2022-01-17 | 2022-05-03 | 广州鲁邦通物联网科技股份有限公司 | SDS state detection method and system in cloud management platform |
CN115687038A (en) * | 2022-08-23 | 2023-02-03 | 江苏臻云技术有限公司 | Hard disk failure prediction system and method based on big data |
CN115687038B (en) * | 2022-08-23 | 2023-09-26 | 江苏臻云技术有限公司 | Hard disk fault prediction system and method based on big data |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN102129397A (en) | Method and system for predicating self-adaptive disk array failure | |
CN104809051B (en) | Method and apparatus for predicting exception and failure in computer application | |
CN108418841A (en) | Next-generation key message infrastructure network Security Situation Awareness Systems based on AI | |
CN110069810A (en) | Battery failures prediction technique, device, equipment and readable storage medium storing program for executing | |
CN104811344A (en) | Network dynamic service monitoring method and apparatus | |
CN103197983A (en) | Service component reliability online time sequence predicting method based on probability graph model | |
CN112001644B (en) | Method, device, terminal and storage medium for detecting operation reliability of power distribution network | |
CN114004102A (en) | Reservoir real-time flood control multi-objective robust optimization regulation and control method for resisting flood forecast error disturbance | |
TWI395965B (en) | Fuel cell faulty predicting system and its establishing method | |
CN109598052B (en) | Intelligent ammeter life cycle prediction method and device based on correlation coefficient analysis | |
CN115375205A (en) | Method, device and equipment for determining water user portrait | |
CN117613421B (en) | Balancing method and medium of energy storage battery pack management system based on neural network | |
CN117878979A (en) | Power balance and dynamic compensation system based on electric energy storage | |
CN117235664A (en) | Fault diagnosis method and system for power distribution communication equipment and computer equipment | |
CN117674119A (en) | Power grid operation risk assessment method, device, computer equipment and storage medium | |
CN108256693A (en) | A kind of photovoltaic power generation power prediction method, apparatus and system | |
CN111476400A (en) | Circuit fault prediction method, device, equipment and computer readable medium | |
Ma et al. | Traffic flow forecasting research based on Bayesian normalized Elman neural network | |
CN111090571B (en) | Maintenance method, device and computer storage medium for information system | |
CN112184320B (en) | AI-based intelligent operation detection and operation and maintenance method for operation and maintenance data of commercial website | |
CN111815069B (en) | Satellite lithium battery voltage prediction method based on time series analysis principle | |
CN111221479B (en) | Method, system and storage medium for judging abnormal storage capacity variation | |
Misra et al. | Tamper detection using neuro-fuzzy logic | |
IHSANE et al. | New evaluation metrics for electrical demand forecasting: application to the residential sector | |
CN102096624B (en) | Method for predicting running state of cluster and cluster platform management device |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C12 | Rejection of a patent application after its publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20110720 |