Exagrid recently announced the new EX13000E which they claim is the “fastest system per TB” on the market. I am not sure how exactly which math they are using, but it is clear that they are off a bit. The Channel Register posted some comparisons showing Exagrid is slower than both Data Domain and Sepaton and said “The EX13000E is not the fastest deduplicating box in the industry by any means” (http://www.channelregister.co.uk/2011/05/24/exagrid_ex13000e/).
If you compare Data Domain and Exagrid head to head, you can see their claims break down even further. The Data Domain DD670 is almost 9tb usable and can ingest data at nearly 9tb/hr when using BOOST, and is all inline, not post process. This inline approach leads to faster time to recovery in the DR site because data doesn’t have to sit on disk waiting to be deduplicated before it is moved to the remote location. The Data Domain line supports very flexible N:N replication while the Exagrid supports only a Hub and Spoke approach. When scaling out the Exagrid solution, you add controller after controller in order to provide the performance required to drive the capacity. Data Domain on the otherhand already has all of the horsepower it needs, which allows you to scale by adding only additional capacity. The largest Exagrid system supports 130TB of capacity and can ingest at 24tb/hr. The largest Data Domain system on the other hand supports 570TB of capacity and can still ingest faster at over 26tb/hr.
Lastly, Exagrid has a price match program that I believe shows a level of desperation. It appears they are simply trying to buy your business. I know I have always found that I get what I pay for…and pricing hardware to simply move product doesn’t instill confidence in the the long term relationship I look for in a vendor.
Digital IRL 
Eric,
I am writing on behalf of ExaGrid Systems in response to your blog post. The confusion you fell into that led to questioning ExaGrid’s performance numbers is understandable based on the architectural differences between the products. We have addressed this matter and responded to your post in our own blog post found at: http://hub.am/j3zr8V
As anyone reading our press release will note, the release is entitled, “Launch of ExaGrid EX13000E Sets New Standard In Backup Price/Performance and Backup Economics.”
For some reason, this post at “Got Dedupe?” failed to mention anything about the key message of our announcement, which was ExaGrid’s superior price/performance. Since the ExaGrid and EMC Data Domain architectures are so thoroughly different, an explanation is warranted of the performance characteristics, which you will find here:
http://blog.exagrid.com/data_deduplication_blog/tabid/16489/bid/54051/ExaGrid-Disk-Backup-Performance.aspx?source=BlogTwitter_%5BExaGrid%20Disk%20Backup%20%5D
We hope our respective commentaries on the architectural and performance differences will help serve the greater good of educating IT professionals so they can make the best decision to fit their technical and business requirements.
LikeLike
Bill —
I am curious. Why on your blog do you have multiple (20 to be exact) EX13000E to match the performance of the DD890? You failed to compare anything to the Data Domain GDA…
So as I read it, If I aggregate the top end speed of 20 Ford Pintos, I could state that a Pinto is faster than a Lamborghini Murcielago?
I think that your numbers are extremely inaccurate (not to mention that you are only quoting ingest speed…When is a backup job fully ingested? Is it when it hits disk or when the dedupe is complete? Do we even need to start talking about replication?). Everything that I see points to EMC having the #1 Purpose Built Backup Appliances, and I am not the only one saying it:
http://www.datadomain.com/news/press_rel_051011.html
http://searchstorage.techtarget.com/product-reviews-2010/Backup-Hardware
Looks like Eric was right when he said, “I know I have always found that I get what I pay for…and pricing hardware to simply move product doesn’t instill confidence in the the long term relationship I look for in a vendor.”
LikeLike
I found more helpful to compare both posts…
The quantity of 20 EX13000E was used for 260 TB. If you read both sides you can see that both exaggerate on different areas…
LikeLike