Deduplication Technology: Does it Really Work? It Depends.

AUGUST 28TH, 2015

The Arcserve competitive intelligence team recently conducted an analysis of vendors who offer deduplication technology, as well as spoke with end-users, partners and industry experts who’ve had hands-on experience with a variety of products. The capabilities of one vendor came up time and again, so to maintain transparency, we’ve included a summary of this analysis below.

Q: We hear from Veeam that “It just works” when it comes to data protection. Is this true for deduplication?

Arcserve Team: There’s a lot more than meets the eye. When it comes to this specific vendor and how well its deduplication performs, we have to consider the facts and data we’ve seen first-hand, through interactions with our partners, and through their own user community.

Q: Can you be more specific?

Arcserve Team: A 2015 report from Gartner points out that their native deduplication technology is a poor performer. This also appears to be a common concern amongst IT Administrators who use Veeam.

One explanation is that their deduplication only works per backup job. So to get high deduplication levels, you’d have to put all of your VMs into the same backup job, which you can imagine is unrealistic and bad practice.

For improved deduplication, Veeam recommends to combine their backup product with a “non-native” solution (e.g. Windows Server 2012 R2), but here’s the catch: Windows Server 2012 dedupe is not without its own issues regarding scalability and performance, and we understand that it’s not designed for high data ingest rates.

Compare Veeam vs Arcserve

Q: How does Arcserve deduplication compare?

Arcserve Team: To analyze the deduplication technology apples-to-apples, we ran two backup tests using the same data set in the Arcserve engineering lab. In the first run, the results were similar for both (see table below). However, in the second run, the results were quite different. Here, Arcserve UDP shows a 94% reduction in raw data, while Veeam only shows a 2% reduction. This is a clear indicator that Veeam deduplication only works per backup job.

UDP and Veeam

Q: This looks very compelling, but these are lab results. What about real-world proof?

Arcserve Team: We encourage end-users and resellers to investigate each product and do comprehensive research; not to take our word for it. But for real-world proof, the below table demonstrates how an actual UDP customer is seeing overall data reduction results of 56-92%. Commonly, UDP users see roughly 70% decrease in their data size, which is a 3X reduction in backup capacity.

In other words, if you need to protect 24 TBs of source data (before deduplication and compression), you only require the Arcserve 9TB appliance or an equivalent server and storage capacity if you are using our software option.

UDP customer screenshot

This is a very important capability to note when you consider the additional cost of backup storage or a dedicated backup appliance. Arcserve UDP provides proven deduplication without any additional cost.

Q: In your analysis, what are the top weaknesses of Veeam?

Arcserve Team:

The first is its lack of scalable backup outside of VMs:

  • They lack support for physical servers, other hypervisor environments, and the ability to backup from major cloud services.
  • The software requires the use of Veeam’s file copy functionality for physical servers, which is not an automated or scalable solution.
  • The software lacks the ability to backup from major cloud services like Azure, AWS, or VFX.

Secondly, their retention periods are very inflexible, meaning that clients run into roadblocks if they want to create their own retention period and backup strategies. For customers with greater than 10 TBs of storage, the deduplication ratio is very low (as noted above). This means you’d need to spend more money to purchase more backup storage or a dedicated backup appliance.

Finally, there’s no backup appliance option. Many end-users like to consume backup technology in this form-factor and connect to a cloud backup service like the Arcserve Cloud.

Q: Any final thoughts?

Arcserve Team: Deduplication is an extremely valuable and effective technology for data protection, so organizations should be vigilant when selecting a backup vendor. The wrong choice could mean they’re required to spend additional money on backup storage of a dedicated backup appliance to get the storage efficiency they require.