Emc IP Holding Company LLC

Hui Li

Shane Sullivan

Henning Rohde

Sunil V Soman

Surinder P Singh

Fangyuan Lin

Carl R Erhard

Xiaohui Wang

Yasmine Alperovitch

Hui Liu

Rui Guo

Jeff Wu

Grace Chen

Yun Qin

Ou Zhang

Boying Lu

Min Han

Li Yuan

Embodiments of the present disclosure provide a method, device, and computer program product for error evaluation. A method for error evaluation comprises in accordance with a determination that an error occurs in a data protection system, obtaining context information related to an operation of the data protection system; determining, based on the context information and using a trained deep learning model, a type of the error in the data protection system from a plurality of predetermined types, the deep learning model being trained based on training context information and a label on a ground-truth type of an error associated with the training context information; and providing the determined type of the error in the data protection system. In this way, it is possible to achieve automatic classification of errors in the data protection system, thereby improving the efficiency in error classification and saving the operation costs. Therefore, more rapid and more accurate measures can be taken to handle the errors.

Method, device, and computer program product for error evaluation

The described technology is generally directed towards automatically scaling distributed computing resources of a distributed computing system based on a system load measurement and a trend factor indicative of whether the system load is increasing or decreasing. If a computing resource load value is above a resource addition threshold value and the trend factor indicates that the computing resource load is increasing, a corresponding computing resource is added to the distributed computing system. If a computing resource load value is below a resource removal threshold value and the trend factor indicates that the computing resource load is decreasing, a corresponding computing resource is removed from the distributed computing system. The trend factor can be obtained using a moving average convergence divergence (MACD) direction indicator.

Scaling distributed computing system resources based on load and trend

Managing access to instances on a public cloud can include in responding to a request from a user to create a support account for a target instance on a public network. A process can include generating a key pair including a public key and a private key, creating, in the target instance, a support account, and associating the public key with the target instance. Access to the private key can be provided to facilitate a connection to the target instance through the support account. Temporary credentials can be generated and associated with the target instance. Upon expiration of the temporary credentials, the support account can be disabled by removing the support account from the target instance and disassociating the public key with the target instance.

Access management to instances on the cloud

A method includes determining if a workflow step has an unsatisfied dependency; setting a state of the workflow step to a blocked state if the workflow step has an unsatisfied dependency; executing the workflow step if the workflow step does not have an unsatisfied dependency; setting the state of the workflow step to a cancelled state if, after executing the workflow step; the workflow step is cancelled; setting the state of the workflow step to a success state if, after executing the workflow step, the workflow step is executed successfully; setting the state of the workflow step to an error state if, after executing the workflow step, the workflow step is executed unsuccessfully; setting a state of a workflow to a success state if the workflow is executed successfully; and notifying other nodes of the state of the workflow step and the state of the workflow using a database replication.

Distributed work flow using database replication

A computer implemented method, system, and computer program product comprising computing at a cluster manager and node manager a realized goal state by examining a goal state stack for a set of nodes of a cluster, determining differences between the realized goal state and the actual state of the cluster and nodes, and based on the differences, determining a set of actions to align the actual state with the realized goal state.

Cluster management

A method includes performing at a lifecycle manager a first process. The first process includes determining nodes in a cluster to run an application, determining a difference between a cluster goal state and a cluster actual state for the application, determining goal state for each node and sending respective goal state to agents. The method also includes performing at each agent a second process. The second process includes determining a difference between the goal state of the node and an actual state of the node, determining actions to be performed based on the difference between the goal state of the node and the actual state of the node and sending actions to resource managers to perform based on the difference between the goal state of the node and the actual state of the node.

Cluster management of distributed applications

Growing community of inventors

Bellevue, WA, United States of America

Hui Li