Deduplication Appliances - Exagrid vs Data Domain

[paps40]

I am a long time Veeam customer currently researching deduplication appliances specifically Exagrid and Data Domain. I am currently using a 32TB Dell R720xd server with Server 2012R2 deduplication as my Veeam Repository. That solution worked great for years but due to unexpected growth it no longer meets our needs as we have several vms over 1 TB which windows dedupe chokes on.

I have been on demos of each solution and received pricing. I wanted to reach out to the Veeam Forums and get feedback / real world experiences of each deduplication appliance. They both integrate well with Veeam but Exagrid does post process dedupe while Data Domain has inline dedupe. Does the data domain inline dedupe affect the speed of sure backup jobs, restores, or writing weekly fulls to tape as it needs to rehydrate the data before doing those? What are some pros and cons of each solution?

I would appreciate any feedback from current admins of those devices on how they like them and if they are happy with their purchase.

Thanks,

Pete

[chris.childerhose]

We currently use a Data Domain DD2500 as our main backup appliance and while the inline dedupe, compression and amount of data you can throw at it is great restores are still an issue for me on length of time to rehydrate the backups. Also lately I have been seeing some corruption of files as well.

From what I hear about Exagrid the Data Mover is built in to the appliance and there is the Hot landing zone which would make restoring way faster. I have never used one but hear great things and may look that direction at our next refresh possibly.

[JaySt]

i've done a few implementations with ExaGrid and i can say they're solid units. I'm not too familiar with DD to really be able to compare based on experience, but it's clear that the main difference is the inline vs post-process. ExaGrid Post Process dedupe and the landing zone will give you great throughput during backups AND restores. Restore performance is best if your restores are coming from your latest backup cycle most of the time, which is what i see the most.
Doing tape-out on data still in the landing zone is also quite fast compared to having to re-hydrated data. Surebackup is fast on an ExaGrid unit as well because the data used by surebackup is not yet deduplicated.

Generally speaking, dedupe appliances are best suited for the longterm storage (due to re-hydration and capacity efficiency etc.). However, ExaGrid has an architecture and some features that still let you do things (like surebackup) with your data in the shortterm, while also having the deduplication within the box. Maybe best of both worlds?

Scaling is also quite different between the two. DD is more on scale up and ExaGrid is all about scale-out. It depends on your (future)needs and data volume which fits better.

ExaGrid support is great and responsive. it's quite rare to find EMC customers saying those two words when talking about support these days...

[tommyo]

If you have the option and the time another appliance you may want to look at is the DXi4700 by Quantum. It has the Veeam Datamover Service built in.

[BartP]

I can imagine various use cases where each of the appliances would be 'prefered' over the other.
If you are planning to utilize Surebackup, make sure you have tested on both appliances and i can make an educated guess that you will like the Exagrid over the DD.
Is bandwidth a possible problem, then a DD shines due being inline deduplication.

[Luca82]

Hi everyone, sorry for bumping up this old thread, but I need your help to understand what "veeam appliance" means. I mean: with both Exagrid and DD the datamover is installed into the appliance, but while I can understand the advantage on DD (inline deduplication), I don't understand what datamover does on Exagrid. I'll try to explain me better. Suppose to have a traditional repository: the deduplication is done by proxy and it is per job. If the repository does deduplication, it will do an offline deduplication...With DD the deduplication is done in sinergy by both proxy and repository: a sort of wan accellerator between proxy and appliance. But Exagrid doesn't do this... it makes an offline deduplication...therefore I don't understand the difference between add Exagrid as an appliance or throught a windows/linux machine wich acts as datamover.

[foggy]

The difference between integrations with Data Domain and Exagrid is whether the target data mover is running on a gateway server (as it does in case of Data Domain, where the DD Boost library is built into Veeam data mover) or directly on the storage (as in case of Exagrid, where data mover itself is installed directly on the appliance). If you add ExaGrid as a Linux server, there will be no Veeam specific sauce in terms of recommended job/repository settings and access control on the storage.