From tapes on trucks to storage in the sky - cloud backup explained
6th February 2012
Back to Databarracks Blog
Cloud backup, online backup or disk-to-disk backup (whichever
name you want to call it) works in a very different way to
traditional backup to tape. When you backup to tapes, it's
easy to visualise where your backups are for each day - you can
point to your tape or tapes and say "that backup is from Tuesday
night". Even if you run incremental or differential backups
to tape, it's still very easy to envisage where the data is.
For people who are new to cloud backup (and even those who have
been using it for a long time) this is often one of the biggest
changes to get their head around. The data is backed up and
then sent off somewhere else - "into the cloud" - and rather than
having lots of small discrete packages of backups sitting on
individual tapes, what you've got is a big storage pot.
Within that single pot is all of the data and the potential to
access hundreds or thousands of versions of files, all held
differently to tape backups.
The reason cloud backup has become so popular so quickly -
despite the fact we are all storing more and more data - is because
cloud backup software is designed to transfer the smallest amount
of data offsite. It also stores data efficiently through the
use of deduplication.
When you begin using a cloud backup service, your first backup
will be the big one and the only real "full backup" ever
needed. Your data will be compressed and encrypted and in
most cases it will need to be physically transported to a data
centre for upload. From that moment onwards you won't
be transferring the equivalent of a tape (or tapes) offsite over
your internet connection each day.
Some of the original backup data may never change again.
Imagine a 4GB corporate movie file that was created a year ago. You
will want to hold on to that video but it is unlikely that you will
ever modify the file again. If you are backing up onto a tape
every day using a full backup, this single file will be taking up
4GBs worth of space and time every backup. With a cloud
backup service, this data can be backed up once and even archived
off away from your changing data.
One of the questions we frequently get asked is how much will
the backup data storage grow as the number of backups increases?
Again, the technology cloud backup uses, such as deduplication and
compression, means that the total volume of storage is kept
low. If you keep two versions of a file/database/VM image,
your backup data doesn't double in size - it increases by just a
small increment. To take a Word document as an example - most
modifications are very small. You might edit just a few
sentences and then save the document again. Rather than
needing storage for two full backups, you have a master backup and
small incremental backups.
Another big difference in cloud backup technology compared with
traditional tape backup methods is that you are able to set
different backup and retention policies for different types of
data. You might have a database that it is important to keep
month-end copies of for 3 years or a finance team who use big
spread sheets they save thirty times a day, which may also need to
roll back to any saved version from the last four weeks. You
are no longer limited by needing to conform to a single
grandfather, father, son tape rotation for all of your backups.
With that being said, it is important to be aware of what affect
this has on backup storage. The ability of software to
perform deduplication and take incremental backups is dependent on
the type of data that is being backed up.
If you chose to keep hundreds of versions of financial spread
sheets, it won't have a big impact on your storage. If,
however, you decide to keep hundreds of versions of your PST files,
it will increase your backup storage significantly. Due to
the nature of the file type, each time they are backed up, there is
a more significant change to the file and a bigger incremental
backup. Over time these will add up.
On a similar basis, if you are backing up at a server image
level - for virtual servers, the incremental backups for those
server images will be larger. Again this has an effect on
both the backup window and the total backup storage.
These VM backups are great for fast and simple restores - but
you must take into account what you are sacrificing in terms of
individual file versioning and storage space. In many cases,
the best approach is a combination of VM level backups and backups
that allow for the specific file versioning and retention
policies.
The obvious benefit of cloud backup is that you can get your
backups off-site to another location automatically without anyone
moving tapes from one site to another. Some of the less
obvious benefits are the greater flexibility of backup retention
policies and the frequency of backups. Once you can get to
grips with the differences and the new options cloud backup allows
- you may find that you can completely rethink your backup
strategy.
Claudiu Craciun, IT Technical Manager