Hi,
When I back data up from an EMC Centera system, is the data on tape
also optimized using the content addressing scheme?

Dvy

[ Post a follow-up to this message ]

Dvy,

First, I'm assuming that you are referring to 'optimization' - as the
blob - de-duping / single instance storage right?

Inside the Centera - lot's of misunderstanding about what is de-duped
and what isn't.
CDFs are never duplicated.

Blobs even though the same, are stored in Centera as duplicates if -
they are embedded in a CDF, or the naming scheme used is anything
other than M or M++.  GM naming kinda does sometimes - depending on
parameters.  Lot's of machines - configed so that de-duping doesn't
occur.

If the app does anything with the blob - it can mess up de-dupe.
(i.e. Legato EX and Symantec EV put  data into a container, and even
though objects in the container are the same - Centera can't de-dupe
them.

All that being said, and assuming that you've got dupe data in the
Centera, that has been de-duped by the Centera - when it goes to tape,
with either of the EMC tools for backing up a Centera (CASBAR or
CBRM), your data will be re-populated.  That is any effect of de-
duping is removed.

At StorageSwitch, we have some tools for BU of a Centera, that are
currently in test, that will give you the option of maintaining the de-
dupe, or not.  Feel free to give me a shout if you have any questions
clark @ storageswitch . com.



On Jan 30, 9:19 am, dvymil...@yahoo.com wrote:
> Hi,
> When I back data up from an EMC Centera system, is the data on tape
> also optimized using the content addressing scheme?
>
> Dvy

[ Post a follow-up to this message ]

On Feb 1, 8:12 am, clark.ho...@gmail.com wrote:[vbcol=seagreen]
> Dvy,
>
> First, I'm assuming that you are referring to 'optimization' - as the
> blob - de-duping / single instance storage right?
>
> Inside the Centera - lot's of misunderstanding about what is de-duped
> and what isn't.
> CDFs are never duplicated.
>
> Blobs even though the same, are stored in Centera as duplicates if -
> they are embedded in a CDF, or the naming scheme used is anything
> other than M or M++.  GM naming kinda does sometimes - depending on
> parameters.  Lot's of machines - configed so that de-duping doesn't
> occur.
>
> If the app does anything with the blob - it can mess up de-dupe.
> (i.e. Legato EX and Symantec EV put  data into a container, and even
> though objects in the container are the same - Centera can't de-dupe
> them.
>
> All that being said, and assuming that you've got dupe data in the
> Centera, that has been de-duped by the Centera - when it goes to tape,
> with either of the EMC tools for backing up a Centera (CASBAR or
> CBRM), your data will be re-populated.  That is any effect of de-
> duping is removed.
>
> At StorageSwitch, we have some tools for BU of a Centera, that are
> currently in test, that will give you the option of maintaining the de-
> dupe, or not.  Feel free to give me a shout if you have any questions
> clark @ storageswitch . com.
>
> On Jan 30, 9:19 am, dvymil...@yahoo.com wrote:
> 
> 

Thankyou.  How about on the NetApp NearStor platform?  How well does
it perform De-Dup?  And will files that I backup from it via my Media
Server to tape also be optimized?

Dvy

[ Post a follow-up to this message ]

On Feb 1, 11:26 am, dvymil...@yahoo.com wrote:
> On Feb 1, 8:12 am, clark.ho...@gmail.com wrote:
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
>
> Thankyou.  How about on the NetApp NearStor platform?  How well does
> it perform De-Dup?  And will files that I backup from it via my Media
> Server to tape also be optimized?
>
> Dvy

You bet,

I don't know about de-dupe in NetApp product.  Heard about the
capability last week, but haven't seen anything on it yet.  Unless the
backup app has a 'backdoor' into the NearStor, I would presume that it
will not be able to maintain the de-duplication in the backups.

Gimme a yell if you go the Centera route.  I have a lot of experience
there.  303.859.3321.

..clark

[ Post a follow-up to this message ]