2015-09-14 22:30:00 UTC

Can anyone share their real-life deduplication ratios using Data Domain?


I know there are a number of variables but EMC is quoting us 6:1, we are seeing 7:1 in a POC but a colleague tells me he is getting 2:1.

Thanks in advance!

Eric

Guest
1414 Answers

author avatar
Top 5Real User

I'm getting about a 20:1 on my DD2500. It all depends on what types of files are being backed up. Some are better at deduping than others. Graphics and Videos are lousy at deduping, but other files are really good.

2015-09-15 16:34:00 UTC
author avatar
Top 20Real User

It depends of what you are backing up, how, when and for how long.
You will see better dedup ratio in filesystem. Databases depends of DB mutability. Media files are not deduped.
The data quality is a big variable on dedup. If EMC is giving you 6:1 its because you will get better dedup ratio, they have a big database of clients on witch they get ratios.

2015-09-15 12:15:26 UTC
author avatar
Vendor

It all depends how many copies they are maintaining. The Dedupe ratio
really differs based on the type of data.
For Oracle DB, I noticed around 4:1 ratio.
For OS backups, the dedupe will be higher as same files exists on
different hosts.

Thanks!
Satish

2015-10-10 15:44:01 UTC
author avatar
Vendor

As Ankur said, data retention longer, get higher ratio. From my experience, it is better to test on your required whole backup data will get better result on POC. It is very difficult to tell exact ratio fit on your company data set.

2015-09-17 10:39:16 UTC
author avatar
Consultant

dedupe ratio depends on various factors. ur colleague told u dedupe ratio is 2:1. but could be for a server where file changes are very frequent. find dedupe ratio for numerous server backups then take a final decision. datadomain dedupe ratio even goes upto 20:1(only for file system)

2015-09-16 18:36:12 UTC
author avatar
Vendor

Hi,

We have just done a PoC with DD. And it really depends on if you enable dd boost and how many round of full backup have been completed.
The dedupe ratio started from 3:1 and ended with 11:1 after we have done 3 cycle of full backup within 2 weeks time for the PoC.

Regards,
Mak CY
Skype! makcy@hotmail.com
Facebook! makcy@hotmail.comWechat! dar3d3v11

From: site-admin@itcentralstation.com
Subject: Seeking expertise in Data Domain
To: makcy@hotmail.com
Date: Tue, 15 Sep 2015 11:22:28 +0000

Hi,
Can you help out a colleague with the following question about Data Domain:
-----

Can anyone share their real-life deduplication ratios using Data Domain?

I know there are a number of variables but EMC is quoting us 6:1, we are seeing 7:1 in a POC but a colleague tells me he is getting 2:1.

Thanks in advance!

Eric
-----
Please reply to this email with your response or post your answer here.

You've been a member of IT Central Station since July 2015.

In a market full of vendor hype, we provide free connections between real users of Data Domain to share advice and make better buying decisions.

Thanks for being part of the community,

Ariel Lindenfeld
Community Director, IT Central Station

IT Central Station
244 Fifth Avenue
New York, NY 10001
(646) 328-1944

To change notifications, please visit your subscription page or unsubscribe completely from similar emails.um-nqn
gitbdelTFFjd0p2bDVVb3pmYU9ucUlMWEFYK1ZqMWI2ZnJ5WndodmlhK3JYWk4xY0pZRmxmVG4wKzZ2OWNHRUVQSWoyTS0telFJVEU4aDFHcjFEbVdYbVRFMHdUZz09--ceec30da700db2115171858eeb1a7fa15007656dgitbdel

2015-09-16 08:11:40 UTC
author avatar
Vendor

Data Domain will gives average 5:1-20:1 depends on the data type. If you are only getting 2:1, you may be putting compressed data which cannot do global deduplication that well.

2015-09-15 15:21:26 UTC
author avatar
Vendor

Hi Eric,

Mostly depends on data volume, data type and retention time . Usually vendors have a very conservative approach when it comes to dedup and compression ratios because there are so many variables. I have seen ratios of 7:1 and above on several installations, but again, it largely depends on the costumers data environment.

2015-09-15 14:17:20 UTC
author avatar
Top 20Real User

EMC figures for marketing:

Filesystem 70% dedup
Normal DB 30% dedup

This figure are given by EMC when they want to performe an assessment.

2015-09-15 13:26:01 UTC
author avatar
Consultant

From DB, normally dedupe ratio is 2:1, for VM, I think 7:1 is possible, for DB, please stop all compression from the host. If image, video, compressed DB, pdf, most of such data are no good to use dedupe appliance for backup.

2015-09-15 12:54:48 UTC
author avatar
Consultant

May I know what you backed up from your POC? VM, DB, File Server, scanned image or?

2015-09-15 12:51:59 UTC
author avatar
Real User

The dedup ratio you will get in your environment is totally dependent on the type of data you are backing up. In our environment we are getting a 9.5:1 dedup ratio. We get a very good dedup on our backup exec and veeam backups, but our Oracle backups the ratio is much lower.

2015-09-15 11:40:51 UTC
author avatar
Top 5LeaderboardIntegrator

Hi Eric,
As you said, many variables. "Mileage varies" is the common saying when discussing things like dedup, compression, and even IOPS and throughput. If you are seeing 7:1 in a POC, and that is a copy of your production data, I would expect that in production. From what I've heard in the past from engineers is that dedup algorithms arent drastically different from vendor to vendor, data domain providing inline dedup is what will set it apart (vs post process). So if your colleague is seeing 2:1, that's likely a result of his data, not the technology.

2015-09-15 11:35:31 UTC
author avatar
ExpertTop 5Real User

I guess it also depends on the model of appliance you are using. We upgraded from a DD670 to a DD2500 and are seeing at least the quoted ratio or higher. Sometimes we get 10:1 but it is also dependent on what data you are sending to the array as well since some data compresses better than others - for example SQL flat file backups. Hope this helps.

2015-09-15 11:32:09 UTC
Learn what your peers think about Dell EMC PowerProtect DD (Data Domain). Get advice and tips from experienced pros sharing their opinions. Updated: April 2020.
419,794 professionals have used our research since 2012.