|
How Does The Backup Process Work? |
|
|
|
|
Click here for a free trial of our Online Data Backup software!A JAVA GUI is loaded in order to select data for backup. Single files, folders, or entire directory
structures can be selected for backup either manually or by using customisable inclusion and
exclusion filters. These can be based on file types (e.g. .doc) or file modification dates (e.g. no
files older than 01/01/1997).
Each backup initiates a new data scan to update the backup set with any additions or deletions.
There is no software limit on individual or collective file sizes. The nature of the backup process
does require disk workspace equivalent to roughly 50% of the backup set size.
This workspace requirement to create the patches is compulsory in the Backup Lite Edition.
Two patching techniques are however available in Server Backup Edition, the default Binary
Patching and Delta Blocking, which reduces this requirement.
Backups can be scheduled (once a day, multiple times daily, once a week, etc.) and can also be
started manually via the GUI or the system tray icon.
The Backup Process
Initial Backup: All data selected for backup is compressed in a backup file. A separate index
file is also created detailing files and their hashes. The maximum size of this backup file can be
limited, in which case the Backup Client will compress until the limit is reached and then flag any
further files for backup at the next schedule time. Thus on the first backup only the first gigabyte
of data (for example) will be backed up. The next time (which could be an hour or a day later)
the first gigabyte will be patched and the next gigabyte will be added, and so on until all files
have been backed up and only patches need to be sent. This process can automatically be
cycled in Server Backup Edition to speed up the process.
Patching: Subsequent backups use a proprietary patching process to reduce the size of the
backup. This works as follows:
1. Backup Client does a scan of the files and folders and filters selected for backup.
2. Any new files are compressed and form the new backup file.
3. Any deleted or removed files are added to a deletion list.
4. Any modified files (we use the modification date and do not reset the archive flag) are first
checked against the local cache to see if a previous version of that file is in the cache. If
so then the files are compared on a binary level (and thus the process is file type
independent) and any changes are stored in a file patch and added to the backup file.
Thus the greater the modification date setting of the cache the greater the chances of a
previous version of a file being present to allow for the patching process.
Backup Transfer: An SSL connection is established to the NameServer (NS) which verifies the
account and passes the IP address of the FileServer (FS) where that accounts’ data is stored.
The connection is dropped and a new SSL connection is established to the FS. The files are
then sent to the FS where they are stored in a dedicated account directory. File hashes are
checked to ensure file integrity. Files are stored using 448-bit Blowfish encryption in CBC mode.
Local SnapShot: Should the initial backup be too large for the available bandwidth, a portable
FS (or DR Box) is temporarily installed on the client LAN and the initial backup is done to this
unit. The DR Box is then physically transported to the remote FS to which the backup is
uploaded. Subsequent backups, being incremental and thus smaller in size, will revert to using
the available bandwidth. This process is known as a Local SnapShot. The same logical
process applies to recovering large amounts of data to the server. Proxy settings can be
specified during these processes.
Local Cache: Once the FS has confirmed the successful transfer of the initial backup, the
Backup Client drops the connection and creates a local cache. This is a compressed folder
containing copies of all files that are backed up that had been modified in the previous 14
(default setting) days. The size of this folder is dependent on the type and size of the files
concerned and also the modification date setting. To reduce the size of the cache the date can
be reduced to files changed in the last 7 days etc. or turned off altogether. This will impact on
the patching process.
The Restore Process
When a user wants to restore files, the Backup Client connects to the Storage Platform and
retrieves the list of all the stored backup dates. The user then selects a date and requests the
index file for that date from the SP. The index file contains a list of all the files that were
protected with that backup. Once the index file has been transmitted, it is displayed in the client
interface where the user can select the file(s) to be restored. The user can also use the Find
option to search for certain filenames or file types. The list of files to be restored is then sent to
the server. The server goes through a process a finding the right files in previous backups,
applying any applicable patches, and then compressing the files after which the requested files
are transmitted. Compression can be disabled in the Backup Client. This speeds up the restore
process on a LAN as the Storage Platform will not compress any patched files before
transferring them to the Backup Client.
For all your Online Backup, Data
backup, Computer backup, Data storage and Data restore
needs go to SaveAndSecure.com
|
