Techniques for determining anomalous user behavior in connection with an online application are disclosed. In one embodiment, a method is disclosed comprising obtaining user behavior data in connection with a user of an application, generating feature data using the obtained user behavior data, obtaining one or more user behavior anomaly predictions from one or more anomaly prediction models trained to output a user behavior anomaly prediction in response to the feature data. Each user behavior anomaly prediction indicates a probability that the user behavior is anomalous. A user behavior anomaly determination is made using the user behavior anomaly prediction(s).

A method for parsing log data according to an embodiment of the present disclosure includes loading a plurality of unit logs identified by first parsing log data into a memory in a two-dimensional matrix, wherein the unit log includes a plurality of items and constitutes one row, determining a target item to be second parsed among the items loaded into the memory, dividing data of the target item into a plurality of sub-items by second parsing the data of the target item, among the items loaded into the memory in the two-dimensional matrix and storing a second parsing result including the plurality of sub-items.

Systems and methods for universal auto-scaling are disclosed. In one embodiment, a method may include: (1) monitoring, by an auto-scale computer program executed by a computer processor, a utilization level at each of a plurality of data layers in a data pod, wherein each data layer comprises at least one node; (2) comparing, by the auto-scale computer program, each of the utilization levels to a threshold; (3) identifying, by the auto-scale computer program, that one of the thresholds is met or exceeded; and (4) deploying, by the auto-scale computer program, an additional node to the data layer with the met or exceeded utilization level.

A method and apparatus are described. The method includes receiving a set of data streams including data values generated by a sensor associated with the operation of a component in a system at points in time and generating an anomaly data value for the received data values. The method further includes applying a machine learning algorithm to the received data values and a subset of data values previously received to generate expected data values at points in time beyond the current point in time, generating an expected anomaly data value for each of the expected data values, and identifying an operational anomaly for the component at a point in time beyond the current time based on the expected anomaly data value. The apparatus includes an input interface for receiving the data streams and a processor for processing the received data values to identify an operational anomaly as described above.

Performance testing a data storage system includes recording operating parameters and performance data as the data storage system executes performance tests over a test period, the performance data including one or more measures of a performance characteristic (e.g., latency) across a range of I/O operation rates or I/O data rates for each of the performance tests. Subsets of recorded operating parameters and performance data are selected and applied to a machine learning model to train and use the model, and the model provides a model output indicative for each performance test of a level of validity of the corresponding performance data. Based on the model output indicating at least a predetermined level of validity for a given performance test, the performance data for the performance test are incorporated into a record of validated performance data for the data storage system, usable for benchmarking, regression analysis, hardware qualification, etc.

A method and a device for analyzing a consumption of resources in a computing infrastructure to predict a resource consumption anomaly on a computing device. The method includes determining a plurality of resource consumption modeling functions; determining a correlation between the resource consumption modeling functions; measuring a resource consumption by a measurement of a consumption value of a first resource; and predicting the resource consumption of the computing infrastructure. The predicting includes a calculation of a value of future consumption of a resource to be predicted from the consumption value of the first resource and from a previously calculated correlation between modeling functions.

A method, system and computer program product, the method comprising: obtaining computer code of an employed system comprising a plurality of components; obtaining data related to operating the plurality of components; based on the computer code and the data, identifying: a first component from the plurality of components, to be maintained; and a second component from the plurality of components, to be at least partly replaced by a machine learning component; and providing to a user an identification of the first component and the second component.

Systems, methods and computer program products for simulating workflows and activities of physical assets using digital twin models. User-defined simulations are performed by selectin digital twin components being analyzed during the simulation, concentrating the analysis on the selectively defined components and bypassing components that will not be simulated. Users can design the digital twin simulation using one or more available digital twin models. The model can be the most current digital twin model, a previous version of a model or a hybridized model comprising components or portions from multiple versions of the available digital twins. Users can further customize simulations by selecting components or sections of the digital twin model to selectively bypass during the simulation or provide overriding values for non-simulated portions of the digital twin which can be used as entry criteria inputted into the next simulated section or component of the digital twin, to complete the simulation.

Disclosed herein are systems and method for gradually updating software object instances on a plurality of computer nodes. In an exemplary aspect, in response to receiving a notification from a software object instance, a system may register the software object instance at an update server. The system may store and deploy a plurality of links, wherein each deployed link uniquely corresponds to a registered software object instance. The system may then associate two or more subsets of the plurality of links with two or more update locations, in accordance with an update policy. The system may place an update to the software object instance at the two or more update locations in accordance with an update policy. In response to receiving an update request via a link from a computing node, the system may further redirect the update request to an update location associated with the link.

Different cores of a multicore processor are used to provide diagnostics of sophisticated hardware without full redundancy by static assignment of the cores during individual cycles of the control program and comparison of the outputs. A method of automatically generating diverse programs for execution by these cores may modify one program to compile two different instructions without functionally changing that program through the use of DeMorgan equivalents and diverse compiler optimizations.