220results about "Reliability/availability analysis" patented technology

Some embodiments provide a method for a first data compute node (DCN) operating in a public datacenter. The method receives an encryption rule from a centralized network controller. The method determines that the network encryption rule requires encryption of packets between second and third DCNs operating in the public datacenter. The method requests a first key from a secure key storage. Upon receipt of the first key, the method uses the first key and additional parameters to generate second and third keys. The method distributes the second key to the second DCN and the third key to the third DCN in the public datacenter.

In memory reads, an increase in a number of correctable errors increases a likelihood that an uncorrectable error will occur. Thus, tracking of correctable errors per memory page is one technique that can be used and if the number of errors on a memory page exceeds a threshold then the memory page can be replaced through dynamic memory re-mapping. However, having a counter per memory page to track the errors is costly due to the number of memory pages. Accordingly, a system is proposed wherein a group of counters is used that acts as a cache of counters that track memory pages of interest due to a variety of factors. Additionally, an eviction policy can control which memory pages are tracked so that only the most interesting candidate pages can be tracked in accordance with a policy.

Systems and methods for detailed cloud posture remediation recommendations utilizing custom Large Language Models (LLMs). The present systems and methods are configured to perform the steps of scanning a cloud environment for posture control data; generating one or more alerts related to any of risky configurations and risky activities associated with the cloud environment; generating one or more remediation recommendations based on the one or more alerts; and providing the one or more alerts and the one or more remediation recommendations to administrators of the cloud environment.

The embodiment of the invention provides a fan service life prediction method and system, a fan, a storage medium and a program product. The fan service life prediction method comprises the steps that the environment temperature of the fan is obtained; if the environment temperature of the fan meets the first condition, the mechanical factor service life and the environment factor service life of the fan are calculated; the first condition is that the environment temperature is smaller than or equal to first preset temperature or larger than or equal to second preset temperature, and the first preset temperature is smaller than the second preset temperature; and outputting the residual life of the fan after fusing the mechanical factor life and the environmental factor life. By fusing the mechanical factor life and the environmental factor life of the fan, the prediction precision of the fan life is improved, the cost does not need to be increased, and the residual life of the fan can be output in real time; and secondly, by monitoring the residual service life of the fan, the effects of good thermal management performance of the equipment and prolonging of the service life of the equipment can be achieved.

Methods, systems, and products for sustainable Reliability, Availability, and Serviceability (RAS) balancing based on predictive models includes predicting, based on one or more predictive models, a potential failure within a computing system, and provisioning, based on the potential failure, one or more redundant resources for failover.

The invention relates to the technical field of data storage, and discloses a solid state disk compatibility and reliability testing method which comprises the steps of initializing a testing environment, installing an operating system, compiling and executing a testing script and configuring a compiling environment. Basic information of the solid state disk is recognized and classified, an automatic script is written, the classified basic information of the solid state disk is created into file clusters of different sizes according to the automatic script, continuous / random reading and writing are conducted, and file integrity is verified; and performing performance test on the classified basic information of the solid state disk by using a Fio tool and recording performance indexes of the solid state disk, wherein the performance indexes comprise sequential read-write speed, random IOPS and response delay, compiling an automatic script, and simulating a database load to perform small block random read-write operation. According to the invention, a plurality of modules for basic function and compatibility test, performance test, power management and reliability test, safety characteristic test and the like are integrated into a complete and automatic test scheme.

The invention discloses a PCBA aging test method and system, and belongs to the technical field of aging test. The method specifically comprises the following steps: S1, historical data acquisition: acquiring PCBA aging related historical data, including but not limited to voltage parameters, current parameters, core area temperature, operation duration, environment humidity, fault types and fault occurrence time during PCBA operation; s2, preprocessing the historical data: preprocessing the historical data, including removing abnormal values, filling missing values and normalizing; 24-72-hour long-time power-on loading of a traditional aging test is not needed, the test period of a single PCBA is shortened to the minute level, the production takt is greatly improved, the large-scale batch production requirement is met, and aging fault recognition is more comprehensive; and recessive aging precursors such as capacitance attenuation and welding spot microcracks can be captured, and the sudden failure risk after the product leaves the factory is reduced.

Disclosed are a service lifetime monitoring and early warning method, a memory storage device, and a memory control circuit unit. The method includes: reading a history information from a rewritable non-volatile memory module, calculating a remaining lifetime based on the history information and a user habit, generating an early warning signal, and outputting the remaining lifetime and the early warning signal in response to the remaining lifetime being lower than a preset lifetime.

An apparatus comprises a processing device configured to determine specifications for an information technology asset to be developed, and to identify, utilizing at least one machine learning model, whether at least one of the specifications for the information technology asset is defect-prone, wherein a given specification is identified as defect-prone responsive to at least one output of the at least one machine learning model indicating that the given specification has at least a threshold likelihood of resulting in one or more defects during development of the information technology asset. The processing device is also configured to establish a mapping between the one or more identified defect-prone specifications for the information technology asset and one or more components of the information technology asset, and to modify one or more development processes for the information technology asset based at least in part on the established mapping.

Methods, systems, and products for sustainable Reliability, Availability, and Serviceability (RAS) balancing based on predictive models includes predicting, based on one or more predictive models, a potential failure within a computing system, and provisioning, based on the potential failure, one or more redundant resources for failover.

The invention discloses a needle report data analysis system based on an artificial intelligence algorithm, and particularly relates to the technical field of industrial production intelligent management and data processing. The needle report data analysis system comprises a data acquisition module, an edge preprocessing module, a cloud analysis module and a decision feedback module; obtaining standardized original annotation data; features are purified through grading preprocessing and an attention gating algorithm, and a lightweight brief report is generated; the cloud integrates physical prior and an AI algorithm to realize life prediction and root cause positioning, generates a multi-dimensional report and mines association rules by improving KNN and Apriori algorithms; and a process adjustment instruction and a maintenance work order are generated based on scene adaptive federal incremental learning, and model iteration is realized through data backflow. According to the method, the problems of cross-fabric adaptation, fuzzy abnormal recognition and the like are solved, and the stitch quality and the production efficiency are improved.

Systems and methods for providing failover readiness prediction, comprise: during a training stage: selecting a first set of data from a database of historical failover data based at least in part on feature importance within the historical failover data; label encoding the first set of data; tokenizing the first set of data; converting the first set of data into a set of dense vector representations; and training a classification model using the set of dense vector representations; during a predication stage: receiving a request to predict a failover outcome for a given failover process; assessing a status of concurrent system conditions; and predicting at least one of a success probability of the failover process, or an expected time for at least one of a plurality of steps in the failover process, based at least in part on the classification model and the status of the concurrent system conditions.

The application relates to a reliability acceleration factor calculation method, device and computer equipment. The method comprises the following steps: obtaining first test parameters of an electronic device to be tested in a first test environment and second test parameters of the electronic device to be tested in a second test environment; the test parameters comprise function test parameters and interconnection test parameters; the environmental severity of the second test environment is higher than that of the first test environment; based on the first function test parameters, the second function test parameters and a preset physical model, a function failure acceleration factor of a component of the electronic device to be tested is determined; based on the first interconnection test parameters and the second interconnection test parameters, an interconnection failure acceleration factor of the electronic device to be tested is determined; and the reliability acceleration factor of the electronic device to be tested is calculated according to the function failure acceleration factor and the interconnection failure acceleration factor. The method can shorten the reliability test time of the electronic device.

A computer-based method for updating risk analysis parameters of a proactive risk analysis of a technical system, such as of an FMEA analysis for example. The method uses incident report data obtained during posterior operation of the technical system, compares these incident report data with data of the proactive risk analysis, and updates the said risk analysis parameters in function of said comparison.

Methods are provided for determining decay rates and / or shelf-lives of intellectual capital (IC) detected issues for only performing actions for expired or about to expire IC detected issues. Specifically, the methods involve performing one or more data collections that include data relating to one or more of a configuration of a computing system or an operation of the computing system and detecting an issue in the computing system based on the data. The issue relates to an anomaly in one or more of the configuration of the computing system or the operation of the computing system. The methods further involve determining a shelf-life for the issue, where the shelf-life indicates an estimated duration of the issue existing in the computing system before redetecting whether the issue is still present and adjusting a next data collection of the one or more data collections based on the shelf-life of the issue.

After generating a first customized recovery plan to resolve a first error associated with a first error message generated by a first software application, a jobs manager obtains identities of an input system, an output system or a combination thereof associated with the first software application. In response to determining that a second software application is associated with the same or similar input system, the same or similar output system, or the combination thereof as the first software application, the jobs manager determines that the first error associated with the first software application is predicted to occur relating to the second software application. Thereafter, the jobs manager generates a second customized recovery plan for the second software application based at least in part upon the first customized recovery plan generated for the first software application.

Disclosed herein are an automatic question generation apparatus and method. The automatic question generation method includes detecting uncertainty by identifying presence or non-presence of missing knowledge and content of the missing knowledge based on environment information and conversation content, as uncertainty is detected, selecting a knowledge type for the missing knowledge, selecting a question type corresponding to the selected knowledge type, and generating a question by combining the knowledge type with the question type.

The invention provides a modular operation skill library and reliability test method and system for a multifunctional mechanical arm. The method belongs to the cross technical field of intelligent robot software architecture, modular control and automatic testing. The method comprises the steps that atomization standard unit decoupling is conducted on an operation task of the multifunctional mechanical arm, and standardized operation unit data is generated; constructing a unified skill model according to the standardized operation unit data, and further forming a modular operation skill library; based on the modular operation skill library, a visual identification module, a voice instruction module and a large model module are configured. By constructing the modular operation skill library, the operation task is decoupled into a standardized atomic unit, the development efficiency of multifunctional mechanical arm operation is greatly improved, new tasks can be rapidly combined without being developed from zero and directly multiplexing existing units, and a large amount of time and labor cost are saved.

A method, computing device, and a non-transitory computer-readable medium are provided. The computing device determines an initial node-local burst buffer content at a start of a time period T. A current node-local burst buffer content is received by the computing device during the time period. For each checkpoint / restart time interval, the computing device: estimates stochastic transition rates λ and μ; estimates input flow data rates of data entering the node-local burst buffer from a compute node; and estimates drain data rates of the data leaving the node-local burst buffer to a parallel file system. The computing device models an average statistical reliability function of the node-local burst buffer within the time period T with respect to not exceeding a predetermined threshold value. When the average statistical reliability function has a value that is less than a predetermined threshold, the computing device performs an action.

To improve the reliability of nodes that are utilized by a cloud computing provider, information about the entire lifecycle of nodes can be collected and used to predict when nodes are likely to experience failures based at least in part on early lifecycle errors. In one aspect, a plurality of failure issues experienced by a plurality of production nodes in a cloud computing system during a pre-production phase can be identified. A subset of the plurality of failure issues can be selected based at least in part on correlation with service outages for the plurality of production nodes during a production phase. A comparison can be performed between the subset of the plurality of failure issues and a set of failure issues experienced by a pre-production node during the pre-production phase. A risk score for the pre-production node can be calculated based at least in part on the comparison.

The present invention relates to a system (108) and a method (600) for processing alarm data. The method (600) includes step of retrieving, historic alarm data from one or more data sources (110). The method (600) further includes step of analysing, using an Artificial Intelligence / Machine Learning (AI / ML) model (220) at least one of, patterns, trends and behaviour of one or more anomalies based on the retrieved historic alarm data. Further, receiving, real time alarm data from the one or more data sources (110). The method (600) further includes step of categorizing, one or more alarms based on the received real time alarm data using at least one of, the AI / ML model (220). Thereafter the method (600) further includes step of predicting, using the AI / ML model (220), one or more future alarms representing the one or more anomalies based on the categorization of the one or more alarms.

Described is a method for monitoring and rotating electrical and / or mechanical components, the method includes receiving a set of metrics for a plurality of components and evaluating the set of metrics for the plurality of components. The method also includes determining to perform a component rotation for a first component from the plurality of components and performing a component rotation based on a component rotation plan, where the component rotation plan indicates the first component from the plurality of components is to be rotated with a second component from the plurality of components.

The application relates to a system and method for architecture-level fault injection based on systolic array accelerator reliability analysis, relates to the technical field of CNN accelerator analysis, and solves the problems that, in the existing CNN reliability analysis method, the reliability analysis of the software level and the architecture level produces contradictory conclusions, and the analysis from the software level alone cannot meet the actual reliability characteristics of the underlying hardware application, and the like, the system comprises an execution simulator, a fault injection module and an analysis module; the execution simulator is responsible for calculation simulation, the fault injection module customizes and simulates fault generation, and the analysis module performs data analysis and result recording; by adopting the method, period-accurate fault injection can be performed on three types of registers under different systolic array sizes, a plurality of different fault types are supported, and the analysis framework can generate detailed intermediate process data information and result record information; the application is simple to implement, cross-platform, has good expansibility, and meets the application requirements.

The present disclosure provides a method for calibrating a qubit gate performed by an electronic device. The method includes: obtaining a first fidelity of a first qubit gate from qubit gates for processing a qubit; determining a first fidelity interval including the first fidelity from a plurality of preset fidelity intervals; determining a target second qubit gate from the qubit gates based on the first qubit gate and the first fidelity interval and based on a first mapping relationship between a plurality of prior-calibrated qubit gates and a corresponding fidelity interval and a plurality of preferred target qubit gates; determining a first calibration means from a plurality of first candidate calibration means based on the first qubit gate and the first fidelity interval and based on a second mapping relationship between the prior-calibrated qubit gates and calibration means; and calibrating the second qubit gate through the first calibration means.

A method and an apparatus for handling a memory failure is provided. The method includes: in response to detecting a failure occurring in memory of a host machine, acquiring (101) a failure parameter of the memory, determining (102) a crash probability of the host machine based on the failure parameter, and transferring (103) all virtual machines on the host machine to a target host machine when the crash probability is greater than or equal to a first predetermined threshold. A crash probability of the target host machine is less than a second predetermined threshold. The second predetermined threshold is less than the first predetermined threshold. Further, an electronic device and storage medium are provided.

A method for dependency task prioritization is disclosed. The method includes providing base data including task identifiers, failure values, and dependency data values, where each dependency data value is associated with a pair of tasks. The method also includes generating a probabilistic model using a directed acyclic graph, where each task is associated with a network node and each dependency data value is associated with a network edge of the directed acyclic graph, where each dependency data value represents a conditional probability related to a respective per-time-unit failure rate. Additionally, the method includes determining, for each task in the directed acyclic graph, a posterior marginal probability value representing a probability of the task experiencing a failure, and selecting, based on the posterior marginal probability values, a sequence of tasks that are most likely to fail the fastest.

The invention discloses an open source big data system software reliability model with fault removal efficiency obeying Weibull distribution, and belongs to the technical field of software reliability analysis. Performance fitting and prediction experiments are carried out through eight fault data sets and nine software reliability models, 95% confidence interval analysis is carried out, and results show that the model provided by the invention has better fault fitting and fault prediction performance than other models. The model provided by the invention can effectively and accurately evaluate the reliability of the open source big data system software, and has certain guiding significance for actual software development and testing.

Techniques for providing a centralized framework for forecasting IT component failures. The techniques include collecting raw telemetry data specific to different IT component domains, and transforming the telemetry data into structured telemetry data. The techniques include performing feature engineering on the structured telemetry data to obtain features relevant to IT component failures in each IT component domain, and, for each IT component domain, using the features to generate a customized ML model. The techniques include accessing features relevant to IT component failures in each IT component domain, accessing a customized ML model for forecasting IT component failures in the IT component domain, and forecasting IT component failures in the IT component domain using the customized ML model. By providing a centralized framework for forecasting IT component failures that combines the use of domain-specific telemetry data with customized ML models, improved fault resilience and reduced system downtime can be achieved.

To provide data management, reduction, and sampling schemes for a storage device failure.SOLUTION: A method for training a machine learning model includes the steps of: segmenting a dataset of a database into one or more datasets on the basis of time period windows by a processor; allocating one or more weights to the one or more datasets on the basis of the time period windows of the one or more datasets by the processor; generating a training dataset from the one or more datasets on the basis of the one or more weights by the processor; and training the machine learning model using the training dataset by the processor.SELECTED DRAWING: Figure 5

This disclosure relates generally to method and system for predicting batch processes. Conventional batch schedulers provide a single point of control for defining and monitoring background executions in a distributed network. The method of the present disclosure obtains a set of batch jobs from one or more users to generate a set of batch graphs by deriving a metadata. Further, a set of batch models is generated for the set of batch graphs. The set of batch models includes at least one of a forecasting model, a job-job regression model, and a job-workload regression model. Further, a batch job schedule is generated for the set of batch graphs to predict a revised batch job schedule with a real time feed and the set of batch models. Additionally, a proactive notification is sent to each user alarming one or more unexpected delays indicating the revised batch job schedule.