A method to speed up restore times for applications that have one or more transactions logs. This is accomplished by first backing up the application, then playing the transaction logs on a separate system. Once completed, perform another backup and make that available for restore. Users are able to create recipes that automate the entire process and integrate into their environment.
An efficient method of storing and accessing Merkle Trees in any graph database. This is done by taking advantage of native graph properties like nodes and edges and applying those concepts to Merkle Trees. By taking advantage of graph query languages, finding elements within the tree is extremely resource efficient.
A process of using efficient database queries without the use of reference counts for the purpose of removing old data. This process attaches a start time and end time to each object within the database and can use native database filters for integers or time to quickly find old data for the purpose of garbage collection. This avoids the need to use reference counts as is typical in such systems.
A process and algorithm of creating static views of multi-connected graphs for the purpose of efficient queries. Graphs contain nodes, edges and vertices each of which can contain any number of properties and traversing all these properties can be an expensive operation. Views can increase performance by creating sub-graphs and static snapshots of these properties.
Ability to detect anomalies, such as ransomware, in backup and other systems by computing the change rate between backups and alerting the system when a change rate is out of norm.
Ability to backup inode based filesystems using the inode table with optimization so file and directory names, along with their stat records can be searched in constant time O(1).
Extract filesystem metadata from block devices in constant time by reverse engineering the filesystem and reading its master file or inode table and storing it efficiently for constant time O(1) search.
Extract filesystem metadata from virtual machines in constant time by reverse engineering the filesystem and reading its master file or inode table and storing it efficiently for constant time O(1) search.
A system where source side data can be index for their metadata properties and stored directly in a scale out search system combined with data protection system with deduplication; all done in an efficient manner resulting in constant time O(1) search and insertion.
A website where a user provides input and selects one or more products to deploy and the site produces a set of automated scripts that deploys and configures each of the products in a simple shopping cart manner that can work across many deployment environments.
A system that determines the optimal storage mediume based on the data properties of each file and object with a goal to reduce storage costs or improve data avaiblity and IO.
A system using blockchain and ledge technology where data integrity can be recorded and verified to a distributed ledger where actors have the ability to prove data integrity by a series of crediability scores.
A deduplication storage system that uses peer to peer communication in a hyper-scale environement to facilitate data storage. The underlying technology uses a distributed ledger to record and catalog all events such as create, update, modify and delete events.
A system of using or creating data labels, assigning those data labels to backups and being able to define data rules based on the data labels. The result is that rules can be applied based on data labels which then get translated to backups.
A system of using or creating multiple data labels, assigning those data labels to backups and being able to define data rules based on the data labels. The result is that rules can be applied based on multiple data labels which then get translated to backups.
Designing a system that is efficient in storing document based data such as office documents from Office 365 in a cost efficient manner on object storage while offering deduplication.
Designing a system that is efficient in storing document based data such as email from Office 365 in a cost efficient manner on object storage while offering deduplication.