Unsolved
This post is more than 5 years old
1 Rookie
•
31 Posts
0
1270
December 15th, 2014 14:00
Archiving vs Long Term Retention
Hi ,
1) I wanted to know the difference between archiving and Long term retention ? is there actually a difference or these can be used interchangeably ?
2) Just trying to understand how archiving assumption fits in a backup design. If a customer says he backs up 2PB of Data every month. Dividing this by 4 ( 4 weeks ) gives me his full backup size of 500TB ( baseline data). I back this up to Datadomain( retention policy daily inc for 21 days , weekly full for 4 weeks and monthly full for 3 months) and if every month he archives 10% of data to Isilon/Atmos ( I am assuming Archiving is equivalent to a cut and paste activity i.e 10% of 500TB =50TB every month is moved to Isilon/ATMOS), would'nt he run out of data to backup at the end of 11th month ? is this how it works, Please explain ?
Thanks in advance
ble1
4 Operator
•
14.4K Posts
1
December 16th, 2014 00:00
Long-term backup is backup using long(er) retention. Archive is archive which normally would use longer retention, but this is not necessary. Key difference is term backup and archive as they both have their own set of requirements (normally those comes from protection layer and query layer). Typically, many companies will ignore this and to cut the cost use backup application to do archiving. In some cases this is feasible; in majority not if we stick to definitions and strict requirements.
Your second question seems to focus on target which in your case would be either filer or cloud device. If using de-dupe, you must know how good does this de-dupe to do math. If you already backup this, use existing data to learn this (though to be honest it won't say much as you might have different targets for backup and archiving and if case of de-dupe they may have different content). Note that archiving by some definitions (or approaches) is not repeatable operation (meaning you remove data from source and therefore there is no repeatitive pattern as in backup).
I assume you need to do archiving sizing for customer which is not so funny for your customer given that you don't have experience here. I would suggest to use logic, get familiar with dataflow and get requirements for archiving first from customer. With that you will be able to create model yourself.