Dear Bacula developers,

A few years ago I deployed IBM Tivoli Storage Manager in the company I
worked for at the time. Since then I've never really encountered a
backup product that could match it in terms of speed or resource
utilization, which can be entirely attributed to the "progressive
incremental" backup strategy it uses. A file is never transferred over
the network twice, TSM always takes incremental backups and keeps track
of files it already has somewhere in its backup pool by using a
relational database (IBM DB2 since TSM version 6). For a longer
explanation, see

https://agora.cs.illinois.edu/display/tsg/Progressive+Incremental+Backups+explained

Only new files are transferred to backup storage, files which have
disappeared from the host since the previous backup are marked inactive
and eventually purged from backup storage depending on the defined
retention policies. It backs up to disk at night for speed, and
transfers from disk to tape during working hours in a FIFO manner. This
way restores are often almost instantaneous because the recent backup
data is still on disk. Other daytime tape maintenance operations involve
the creation of an off-site copy of the primary storage pool tapes
(which are always on-line (!), your tape library must be large enough to
accomodate this), reclamation (freeing tapes with mostly expired data)
and collocation (moving data from a specific host on as least tapes as
possible).

The system works really well. Unfortunately no other backup product that
I know of implements the same backup strategy. As a side effect, there
is no real competition in this space and the licensing costs of TSM
aren't pretty... My current employer isn't a TSM shop and as I'm not
exactly thrilled with our current backup solution, I'm looking at
affordable TSM alternatives but it appears that there just aren't any.

I hoped that Bacula's new "basejob" deduplication feature would start
offering something in this direction (as files in a basejob are only
backed up once), but now that I've read a bit more about it, it doesn't
seem to do what I hoped for. :-(

What could be the reason no other companies or open source projects go
in this direction? There are great open source databases, there are
great open source backup projects, but there are none which attempt to
forge these technologies into an "always incremental" backup product (or
"enterprise class data management system" as some prefer to call it...).

Best regards,

Pascal

P.S. The following document is a great introduction to TSM concepts:
http://www.redbooks.ibm.com/redpapers/pdfs/redp0044.pdf



Attachment: signature.asc
Description: OpenPGP digital signature

------------------------------------------------------------------------------
The Planet: dedicated and managed hosting, cloud storage, colocation
Stay online with enterprise data centers and the best network in the business
Choose flexible plans and management services without long-term contracts
Personal 24x7 support from experience hosting pros just a phone call away.
http://p.sf.net/sfu/theplanet-com
_______________________________________________
Bacula-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/bacula-devel

Reply via email to