Guepard / Overview
How Guepard Works
Deep dive into Guepard's virtual data management
Guepard implements virtual data by capturing an initial copy of a database and continuously recording all changes to that database over time. All captured data blocks and transaction logs are compressed.
Snapshots and Database Versions
A database version is represented by a snapshot, created by taking an incremental backup of the source database. Guepard stores compressed copies of all changed blocks in the snapshot. The snapshot is structured as a tree of pointers to the blocks consistent with that point in time. Guepard keeps track of the distinct set of data blocks for all snapshots associated with a specific database. Each data block in Guepard is referenced by all snapshots that depend on it.
Virtual Databases
A Virtual Database can be created from any point in time since the source database was first copied into Guepard. If the Virtual Database is based on an existing snapshot, it will share the same tree of pointers. If the Virtual Database needs provisioning from a different point in time, Guepard performs point-in-time recovery, creates new versions of required data blocks, and assigns its own tree of pointers to the Virtual Database snapshot.
When a Virtual Database is initially created, the only additional storage required is for the unique blocks of that Virtual Database. Any subsequent changes made through DML and DDL activity on the Virtual Database are specific to that Virtual Database.
Efficient Data Management
This efficient approach allows Guepard to provision multiple database copies instantly with minimal storage overhead. Guepard also applies a similar method to files and folders, enabling customers to virtualize all their data.
Guepard's method ensures efficient data management, enabling businesses to maximize their data utility while minimizing resource consumption.
- Minimal Storage Overhead: Only unique data blocks require additional storage.
- Instant Provisioning: Create multiple database copies instantly.
- Comprehensive Data Virtualization: Applies to both databases and file systems.