Backup Deduplication

IBM Storage Announcement

As previously posted, I was confused about the muted launch of IBM’s XIV disk platform. Well, the formal launch finally occurred at IBM Storage Symposium in Montpelier, France. Congratulations to IBM, although I am still left scratching my head why they informally announced the product a month ago!

Another part of the announcement was the TS7650G which is Diligent’s software running on an IBM server. Surprisingly, there is not much new; it appears that they are banking on the IBM brand and salesforce to jumpstart Diligent’s sales. Judging by the lack of success in selling the TS75xx series, it will be interesting to see whether they will have any more success with this platform.

From a VTL perspective, IBM has backed themselves into a box. Like EMC, they have a historic relationship with FalconStor and have chosen a different supplier for deduplication. This creates an interesting dichotomy. Let’s look at the specs of their existing FalconStor-based VTL and newly announced technology.

4,400 MB/sec
900 MB/sec
Max Capacity
1.3 PB
1.0 PB
VTL Software Provider
Deduplication Enabled
Source: IBM datasheets, emphasis added by the author

The most obvious difference is that the TS7520 is about 5x faster than the TS7650G. Customers have to make a difficult trade off: Do you want to meet your backup and restore requirements or deduplicate? The problem is that the benefit of disk-based backup is performance and the slow performance of the TS7650G limits its usability in many larger accounts. At SEPATON, we believe that this is not a workable trade off since end users need both. We architected our DeltaStor deduplication software to provide both performance AND scalability to meet this need.

The other challenge is that single stream performance for Diligent’s software is poor. Here is a chart from an ESG whitepaper, sponsored by Diligent, entitled “ProtecTier Data Protection Platform, A validation study dated July 2006” highlighting their performance.

Diligent Performance
See Page 11, Figure 9

This is a bit older and lists the maximum performance is 250 MB/sec. That said, look at the single stream performance. The graph suggests that it is about 25 MB/sec. Let’s be generous and assume that single stream performance increased linearly with overall performance. (e.g. the currently released product performs at 450 MB/sec which is 1.8X 250) Thus current single stream performance would be 45 MB/sec (25 MB/sec * 1.8). This represents the fastest that an individual backup job can go. Your single stream Oracle backup will never go faster and so if your Oracle database is 4 TB, it will take you 25 hours to complete the backup. Of course, you can enable multi-streaming, but that adds cost and complexity to your environment and thus reduces the ROI of your solution.

On a side note, did you see the hardware configuration of this solution (page 18)? In order to achieve this performance, Diligent required a very powerful server and 110 300GB 10,000 RPM Fibre Channel drives. So much for a lightweight footprint! This hardware is required because the Diligent application is extremely random I/O intensive and performance suffers with lower cost and higher density SATA drives.

The other challenge for this solution is that by moving to an entirely new architecture (to Diligent from FalconStor), they have left their customers with no migration path. If you bought a TS7520 and want deduplication, your only option is a forklift upgrade. This is particularly painful since you will need to purchase more storage and simultaneously support both systems. Forget the idea of data in place upgrade, a better term would be forklift in place.

If you are a FalconStor/IBM customer, you have to wonder what this means to you. IBM spent between $100 and $200 million on Diligent and now owns technology that directly competes with FalconStor. What does this mean to IBM’s existing FalconStor based products? How long will the IBM/FalconStor solutions be offered and/or supported? How do IBM/FalconStor customers gain access to deduplication technology? As an existing or potential IBM customer, you need to obtain answers to these questions. If the only future path is to a forklift upgrade, then you owe it to yourself to consider other options.

Be Sociable, Share!
  • Twitter
  • Facebook
  • email
  • StumbleUpon
  • Delicious
  • LinkedIn

Leave a Reply

Your email address will not be published. Required fields are marked *