System And Method for High Performance Enterprise Data Protection

An enterprise image destination for backup images, wherein such images may be consolidated, nearlined and versioned, comprising:

• one or more data storage units accessible by said enterprise image destination;
• software adapted for operating said enterprise image destination, said software comprising support for a journaling file system with sparse file support for accommodating multiple persistent snapshots;
• an interface providing support for LUN creation/deletion, LUN cloning, and snapshot creation and deletion;
• and
• a database associating snapshot IDs with block allocation bitmaps and incremental bitmaps.

A method for creating or updating a tertiary backup image on a tertiary node from a primary backup image stored in a node constituting an enterprise image destination, the enterprise image destination comprising one or more data storage units accessible by said enterprise image destination, software adapted for operating said enterprise image destination, said software comprising support for a journaling file system with sparse file support for accommodating multiple persistent snapshots, an interface providing support for LUN creation/deletion, LUN cloning, and snapshot creation and deletion, and a database associating snapshot IDs with block allocation bitmaps and incremental bitmaps, said method comprising:

• using software on the enterprise image destination to contact the tertiary node and pass to it a message to request data from the enterprise image destination node;
• using software on the tertiary node to request said data;
• receiving said request at the enterprise image destination node and recognizing it as a request to create or update a tertiary backup image;
• designating the appropriate snapshot image using saved bitmaps;
• and
• sending the appropriate snapshot image to the tertiary node.

A method for rolling back a volume to an earlier state, comprising:

• addressing the volume by a unique job identifier and locking and dismounting so the volume;
• taking a snapshot of the volume to flush the change journal;
• mapping the snapshot as a mountable volume;
• logically ANDing all change journals since the time of the desired state that is being restored to, to create a bitmap file;
• adding the contents of the bitmap file to the content file associated with the snapshot;
• using the bitmap file to restore only a set of blocks from the mapped volume.

A method for providing an immediately available restoration of a read-only backup image to a target volume, comprising:

• transforming the read-only backup image into an iSCSI addressable read-write LUN by creating a sparse file backed by the backup image;
• exposing the read-write LUN to create a restore volume containing both backup data and application-specific metadata;
• mapping the restore volume to the local namespace;
• determining whether the target volume supports addressing by a unique job identifier;
• determining that the restore volume and the target volume are logically related;
• and
• calling a back-end API with the job identifiers of the restore volume and the target volume.

A method for mirroring a primary volume onto a secondary volume, comprising:

• creating a list of blocks to be copied from the primary volume to the secondary volume;
• writing new blocks to both the primary and secondary volumes, as they arrive;
• as blocks are written, removing those blocks from said list of blocks;
• traversing said list, and copying blocks encountered as a result of such traversal from the primary volume to the secondary volume;
• and
• continuing the preceding two steps until all blocks on said list have been copied.

A method for reducing redundancy in a primary data set being backed up on a changed block basis to a backup server used to back up a plurality of data sets, comprising:

• taking a digest of each block that has changed in said primary data set;
• saving the digest data;
• comparing the block digests of said blocks to the digests of blocks previously saved by the server, and determining which blocks the server does not already have;
• sending to the server only those blocks that the server has determined that it does not already have, together with digests associated with those blocks;
• updating the collection of digests saved on the server by adding said associated digests;
• and
• constructing a catalog of the backup from the list of blocks sent to the backup server plus those blocks which the server determined that it already had.

A method for reducing redundancy in a file system, comprising:

• taking a digest of each block as it is to be written to the file system;
• comparing the digest of said block with the digests of the blocks already stored in the file system;
• if the identical content block already exists on file system, using the existing directory pointer and not writing the duplicate block;
• when a file is deleted, deallocating the associated blocks from that file;
• and
• freeing any block when no files reference it.

A method for rapidly restoring an application from a set of one or more backup images to a target node, comprising:

• selecting said backup images;
• creating addressable virtual storage LUNs from said backup images on an on-demand basis;
• making the LUNs visible to the target node;
• attaching the LUNs as local disks on the restore target via iSCSI protocols;
• and
• programmatically recovering the application using appropriate application-specific APIs.