Abstract: A method and system for generating a queryable backup of database applications to allow database operations are presented. The method includes accessing a plurality of database applications; exporting data stored in the database applications to a plurality of files, each file utilizing a column-oriented storage format; generating metadata based on the plurality of exported files, wherein the metadata includes a hash map of related-columns and a list of related-tables, wherein generating the metadata comprises normalizing column values of tables in the database applications and identifying related-columns based on at least the normalized column values by computing hash-based signatures for two columns in the normalized column values and designating the columns as having similar values upon determining that the columns share a threshold overlap in their computed hash-based signatures; and generating a queryable backup based on the plurality of exported files and the metadata.

Abstract: In one embodiment, a method comprises analyzing a non-relational database comprising a plurality of documents; and generating a backup of the non-relational database, the backup comprising a table with a plurality of rows and columns, wherein a row of the table corresponds to a document of the plurality of documents, wherein the non-relational database is restorable based at least in part on the table.

Abstract: A system and method for generating a quarriable backup of database applications to allow database operations is provided. The method includes accessing a plurality of database applications deployed in a cloud environment; exporting data stored in the database applications to a plurality of files, each file utilizing a column-oriented storage format; generating metadata based on the plurality of exported files, where the metadata includes at least a hash map of related-columns and a list of related-tables; and generating a quarriable backup based on a plurality of exported files and the metadata.

Abstract: In one embodiment, a method comprises receiving, from a plurality of large language models, a first plurality of application detectors for a plurality of applications, wherein an application detector comprises executable code to detect whether an application is installed on a computing system; executing the first plurality of application detectors to determine whether the first plurality of application detectors accurately detect applications installed on a plurality of computing systems; and selecting, based on detection accuracy of the first plurality of application detectors, a second plurality of application detectors to deploy for use in detecting the plurality of applications.

Abstract: A method and system for identifying relevant database tables in a database backup are provided. The method includes receiving an input query, wherein the input query is a free text query; generating a prompt for a first language model that, when executed by a generative AI system, provides a set of column names most relevant to answering the input query; embedding a search vector for the set of relevant column names; and identifying relevant database tables by performing a similarity search between the search vector and vectors in a vector database, wherein the vectors in a vector database represent at least a portion of contents of database tables in the database backup.

Abstract: A method and system for querying multiple database tables in a database backup is presented. The method includes receiving identification of database tables to query, wherein the database tables to query are database tables of different database applications; receiving an input query in a free text format; retrieving classification data for each of the identified database tables; and generating a prompt for a first language model that, when executed by a generative AI system, provides a set of queries targeting the database tables, wherein the generated prompt is based on the input query and the classification data for the database tables.

Abstract: A method and system for generating a temporal, query-ready database backup. The method includes generating a plurality of backups of a database, each of the backups including a plurality of backup data objects, each of the backup data objects including a plurality of rows stored in a file having a column-oriented format, and each backup corresponding to a unique point in time; assigning each row in a file to a row group, the row group having a unique identifier; generating a temporal table including a column of row group identifiers for each backup of the plurality of backups; receiving a query for a backup of the plurality of backups created at different point in times; and generating a query result based on the temporal table and at least the respective backup data object, wherein results are from backups of the plurality of backups related to different point in time.

Abstract: A method and system for reducing deduplication in database backup are provided. The method includes selecting a first primary key of a database; extracting a plurality of row groups from the database based on the selected first primary key; storing the plurality of row groups arranged according to the first primary key in a file having a column-oriented data storage format as a database backup at a first time; reading at least a portion of the database backup at a second time to detect a plurality of changes; determining a position of each row corresponding to a change of the plurality of changes; selecting a second primary key of the database as a primary key; and generating a database backup of the database at the second time based on the selected second primary key.

Abstract: A method and system for reducing deduplication in database backup are provided. The method includes selecting a first primary key of a database; extracting a plurality of row groups from the database based on the selected first primary key; storing the plurality of row groups arranged according to the first primary key in a file having a column-oriented data storage format as a database backup at a first time; reading at least a portion of the database backup at a second time to detect a plurality of changes; determining a position of each row corresponding to a change of the plurality of changes; selecting a second primary key of the database as a primary key; and generating a database backup of the database at the second time based on the selected second primary key.

Abstract: In one embodiment, a method comprises receiving a request to view a portion of a file system; identifying a first file of the portion of the file system that is not likely to be corrupt, a second file of the portion of the file system that is likely to be corrupt, and a backup of the second file that is not likely to be corrupt; and generating a representation of the portion of the file system, the representation displayable in a file system explorer, wherein the representation includes an identifier of the first file, an identifier of the second file, and a point in time of the backup of the second file that is not likely to be corrupt.

Abstract: In one embodiment, a method comprises responsive to a request, accessing a storage backup log, the storage backup log comprising indications of changes to files of a plurality of storage backups relative to previous storage backups, wherein each storage backup of the plurality of storage backups is associated with a different point-in-time; and responsive to the request, utilizing the storage backup log to generate a representation of at least a portion of a file system at multiple points in time, wherein the representation includes indications of a first file in existence at a point-in-time, a second file deleted prior to the point-in-time, and a third file created after the point-in-time.

Abstract: In one embodiment, a method comprises receiving a request to view a portion of a first file system and a second file system; and generating a representation of a portion of a first file system and a second file system, the representation displayable in a file system explorer, wherein the representation includes a file at a file path, a first indication of one or more backups of the file for the first file system, and a second indication of one or more backups of the file for the second file system.

Abstract: A method includes probing one or more accounts on a cloud infrastructure so as to identify content of multiple data resources hosted by the cloud infrastructure in association with the accounts, receiving, from a user of the accounts, a backup policy specifying one or more backup parameters and at least one set of one or more data-resource properties, identifying, based on the identified content, at least some of the data resources that have the data-resource properties, and backing up the identified data resources in accordance with the backup parameters. Other embodiments are also described.

Abstract: A system and method for querying multiple database tables in a database backup are provided. The method includes receiving identification of database tables to query, wherein the database tables in a column-based format, and the database backup includes for each database table classification data for at least classifying its respective column names; generating a prompt for a first language model that provides a set of queries targeting the database tables, wherein the generated prompt is based on at least a received input query and the classification data for the database tables; executing the set of queries on the database tables, wherein each query of the set of queries is executed on a respective database table; and aggregating and displaying results returned in response to the execution of the set of queries.