Abstract: A container-based software implementation uses separate containers for software libraries and application code. A storage system may have multiple applications executing to control various aspects of operation of the storage system, and to enable access to the storage system by hosts. These applications are containerized separately from the libraries referenced by the applications, and the libraries are commonly housed in a separate container. The libraries may be open-source libraries, proprietary libraries, or third-party dependent libraries. A vulnerability management system scans the application containers to determine dependencies between applications and libraries, including the number of containers that reference a particular library and the frequency with which microservices of the containerized application reference the library.

Abstract: Techniques are provided for machine learning-based prediction of completion time of software code changes. One method comprises obtaining events related to software code changes; applying the events to a machine learning model that predicts a completion time of the software code changes, wherein the machine learning model is trained using (i) events for historical software code changes and (ii) a completion time for each historical software code change; and performing a remedial action based on the predicted completion time. The remedial action may comprise generating a notification, and/or adjusting an allocation of resources assigned to a completion of the software code changes. The software code changes can be monitored and a new event related to the software code changes can be applied to the machine learning model to obtain an updated predicted completion time for the changes to the software code.

Abstract: A container-based software implementation uses separate containers for software libraries and application code. A storage system may have multiple applications executing to control various aspects of operation of the storage system, and to enable access to the storage system by hosts. These applications are containerized separately from the libraries referenced by the applications, and the libraries are commonly housed in a separate container. The libraries may be open-source libraries, proprietary libraries, or third-party dependent libraries. A vulnerability management system scans the application containers to determine dependencies between applications and libraries, including the number of containers that reference a particular library and the frequency with which microservices of the containerized application reference the library.

Abstract: Methods, apparatus, and processor-readable storage media for automated device power conservation using machine learning techniques are provided herein. An example computer-implemented method includes obtaining usage-related data from one or more processing devices; determining at least one usage pattern for the one or more processing devices by processing the obtained usage-related data using one or more machine learning techniques; automatically generating, based at least in part on the at least one determined usage pattern, instructions pertaining to controlling one or more power states of the one or more processing devices; and performing at least one automated action based at least in part on the generated instructions.

Abstract: A workload manager uses on-band and off-band metrics to select a host server in a cluster to handle a connection request. The on-band metrics include CPU usage, memory usage, and vulnerability metrics. The off-band metrics include hardware component error logs. Utilization and vulnerability scores are calculated for each host server from the on-band metrics. A reliability score is calculated for each host server from the off-band metrics. A health score for each host server is calculated from the vulnerability and reliability scores. The health score is used to exclude unhealthy host servers from consideration. A priority score is calculated for each host server from the utilization, vulnerability, and reliability scores. The host server that has not been excluded and has the greatest priority score is selected to handle the connection request.

Abstract: An apparatus comprises at least one processing device coupled to memory. The at least one processing device is configured to obtain a secured disk image comprising an encrypted manifest file, an encrypted install binary and a plurality of other files. The at least one processing device is further configured to obtain a certificate corresponding to the secured disk image and to derive a public key based at least in part on the certificate. The at least one processing device is further configured to decrypt the manifest file and the install binary based at least in part on the public key and to validate checksums for respective ones of the plurality of other files against corresponding checksums contained in the decrypted manifest file. The at least one processing device is further configured to execute the decrypted install binary responsive to validation of the checksums for the respective ones of the plurality of other files.

Abstract: A workload manager uses on-band and off-band metrics to select a host server in a cluster to handle a connection request. The on-band metrics include CPU usage, memory usage, and vulnerability metrics. The off-band metrics include hardware component error logs. Utilization and vulnerability scores are calculated for each host server from the on-band metrics. A reliability score is calculated for each host server from the off-band metrics. A health score for each host server is calculated from the vulnerability and reliability scores. The health score is used to exclude unhealthy host servers from consideration. A priority score is calculated for each host server from the utilization, vulnerability, and reliability scores. The host server that has not been excluded and has the greatest priority score is selected to handle the connection request.

Abstract: Methods, apparatus, and processor-readable storage media for automated device power conservation using machine learning techniques are provided herein. An example computer-implemented method includes obtaining usage-related data from one or more processing devices; determining at least one usage pattern for the one or more processing devices by processing the obtained usage-related data using one or more machine learning techniques; automatically generating, based at least in part on the at least one determined usage pattern, instructions pertaining to controlling one or more power states of the one or more processing devices; and performing at least one automated action based at least in part on the generated instructions.

Abstract: A method of providing an error occurrence estimate for a proposed software update, before the proposed software update is created, includes training a learning process to cause the learning process to learn a correlation between the complexity of the previous software updates and the error occurrences of the previous software updates. The complexity information may include the number of lines of code and the number of check-in operations that occurred in connection with creation of the previous software updates. The trained learning process is then provided with expected complexity information of a proposed software update, and used to generate an error estimate including the number of errors that are likely to occur, the severity of the errors that are likely to occur, and the amount of software developer time that should be expected to be incurred to correct the errors.

Abstract: An apparatus comprises at least one processing device coupled to memory. The at least one processing device is configured to obtain a secured disk image comprising an encrypted manifest file, an encrypted install binary and a plurality of other files. The at least one processing device is further configured to obtain a certificate corresponding to the secured disk image and to derive a public key based at least in part on the certificate. The at least one processing device is further configured to decrypt the manifest file and the install binary based at least in part on the public key and to validate checksums for respective ones of the plurality of other files against corresponding checksums contained in the decrypted manifest file. The at least one processing device is further configured to execute the decrypted install binary responsive to validation of the checksums for the respective ones of the plurality of other files.

Abstract: A method of providing an error occurrence estimate for a proposed software update, before the proposed software update is created, includes training a learning process to cause the learning process to learn a correlation between the complexity of the previous software updates and the error occurrences of the previous software updates. The complexity information may include the number of lines of code and the number of check-in operations that occurred in connection with creation of the previous software updates. The trained learning process is then provided with expected complexity information of a proposed software update, and used to generate an error estimate including the number of errors that are likely to occur, the severity of the errors that are likely to occur, and the amount of software developer time that should be expected to be incurred to correct the errors.