Rubrik, Inc.

Indian Institute of Technology Delhi

Vaibhav Bhagee

Sai Kiran Katuri

Amelia Vu

Ayush Mishra

Alok Srivastava

Sorav Bansal

Ankush Phulia

Abhishek Goyal

Nicolas Groh

Kshitiz Jain

Nitin Rathor

Toshad Salwekar

Vishnu Vardhan Itta

Sankaran Sivathanu

Methods, systems, and devices for data management are described. Software-as-a-service (SaaS) applications may host customers' information in a distributed environment that is not directly accessible to the customers. SaaS applications may store data in the form of relational tables. Snapshots for SaaS applications may maintain hierarchical relationships between computing objects and tables to enable cascading recovery. A data management system (DMS) may identify the computing object hierarchy for an organization's data as hosted by a SaaS application and also may identify the APIs associated with accessing each computing object, which may each include multiple tables. Tables that are accessed via a same API may be organized as logical entities. The DMS may store the hierarchical relationship between the tables, and as the tables are stored as logical entities, the DMS may organize the tables in the backup database based on which APIs are used to access the relevant tables.

Relational software-as-a-service data protection

Methods, systems, and devices for data management are described. Software-as-a-service (SaaS) applications may host customers information in a distributed environment that is not directly accessible to the customers. In some SaaS applications each organization may have a settings computing object, a features computing object, and a set of projects computing objects. The hierarchical relationships in each of the projects computing objects may depend on the tables in the settings and features computing objects. A data management system (DMS) may include discovery, backup, refresh, and restore frameworks that maintain hierarchical relationships between computing objects and tables including the dependency of projects computing objects on settings or features computing objects.

Backup and recovery for software-as-a-service data

Methods, systems, and computer programs are presented for backing up data to a distributed multichannel storage system. One method includes an operation for detecting, by a distributed backup system (DBS), a request for backing up data from a host to the distributed backup system. The DBS sends a request to a backup agent executing at the host to initiate the backup. The backup agent mounts an endpoint for writing to the DBS and executes a script that writes the data for the backup to the endpoint. Further, the method includes operations for creating a file system stack at the DBS associated with the endpoint, storing data received by the file system stack at the DBS, and detecting, by the DBS, an end of the backup. The DBS sends a command to the backup agent to end the backup and terminates the file system stack at the DBS.

Multi-channel backup system

A data management system (DMS) may perform techniques for snapshotting a host environment data store using a resource volume. The DMS may transmit, to the host environment a first request for a full backup of a data store of the host environment, where the first request results in mounting of a resource volume in an empty state. The DMS may generate a snapshot of the resource volume after completion of the backup at the host. The DMS may transmit one or more second requests for incremental backups, where the second requests result in mounting of the volume in a state corresponding to a previous backup of the host. The DMS may generate a second snapshot of the resource volume. For a subsequent full backup, the DMS may cause mounting of the volume in the empty state to the host rather than mounting the resource volume corresponding to a prior snapshot.

Snappable recovery chain over generic managed volume

A method, apparatus and system required to realize an optimization opportunity are disclosed. If an expression involves unsequenced evaluations involving side-effects and optionally references, then the expression is translated into optimized code that may involve optimized register usage, or optimized use of the underlying vectorization hardware. An alias-analysis logic is able to identify must-not-alias relationships through such expression tree structures. These must-not-alias relationships are helpful not just in optimizing the performance of that particular expression (from which these must-not-alias relationships were derived) but also in the performance of any other surrounding program statements that access the state elements participating in these must-not-alias relationships. More specifically, if the order of evaluation is unspecified and the sub expressions generate side effects, then must-not-alias relationships are inferred between expressions which can be used while optimizing surrounding code, such as the instructions following this expression with unsequenced evaluation.

Leveraging unspecified order of evaluation for compiler-based program optimization

Growing community of inventors

New Delhi, India

Vaibhav Bhagee