Talking with Data Domain

Today I have the pleasure of interviewing Saradhi Sreegiriraju, a Product Manager from Data Domain. Data Domain is a partner in our OpenStorage initiative. If you aren't familiar with OpenStorage you can read all about it in section 3.5 of this Symantec whitepaper on "NetBackup Disk Based Data Protection".

Q. Tell us a little more about yourself. How long have you been with Data Domain and what is your role there?

A. I’m the Product Manager at Data Domain for the NetBackup OpenStorage product. I also have Product Management responsibilities for our System Management and Quality of Service products. I’ve been with Data Domain for the past 22 months and prior to that I was with NetApp and Fujitsu.

Q. I'm interviewing you specifically about the OpenStorage program. You were the first vendor to release an OpenStorage solution. From your perspective what is the problem the OpenStorage program is trying to solve in the backup and recovery workflow?

A. Data Domain was the first vendor to release a product and the first to pass the Symantec OpenStorage

hardware compatibility testing program. And to the best of our knowledge, Data Domain is the only vendor that has the solution in many production deployments.

The biggest challenge NetBackup administrators have is the lack of single pane of visibility from NetBackup into the disaster recovery copies of their backup data. This complicates and lengthens the recovery process as NetBackup needs to import the images at the DR site before proceeding with recovery. OpenStorage elegantly solves this problem by providing a mechanism for seamless integration between NetBackup and Data Domain Storage systems.

With OpenStorage, NetBackup controls the backup of data to the Data Domain systems and the duplication of the backup data to another Data Domain system at the DR site and keeps track of both the copies in its catalog. Data Domain systems use WAN efficient Replicator Software to duplicate the data to DR site. This allows for an improved and faster disaster recovery process because NetBackup knows about the copy at the DR site. Additionally, NetBackup’s Storage Life Cycle policies allow administrators to fully automate the retention and duplication of the images stored on Data Domain systems.

OpenStorage also provides the ability for NetBackup to treat disk as disk and take full advantage of the advanced capabilities offered by Data Domain systems while eliminating the management complexities and liabilities of tape emulation with disk based storage systems. For example, VTL users are forced to provision imaginary tape drives, slots and cartridges for no added benefit. The inability of data management tools to handle multiple instances of the same barcode or image when making DR copies becomes a breaking point.

Q. What has the customer reaction been to the Open Storage devices that you have shipped so far?

A. The customer reaction has been very positive. Customers immediately see the value proposition OpenStorage brings to table and how it can help them by providing a single pane of management from NetBackup for all the copies of their backup data. The value proposition of OpenStorage is clear from the adoption we’ve seen from large enterprise customers with multiple sites.

Q. What enhancements are you planning to implement in the near future?

A. Data Domain and Symantec are collaborating on a number of additional features and Data Domain plans to support the new features that are being planned in the coming NetBackup releases. These include Hierarchical Duplication, OpenStorage Direct Copy-to-tape, and Virtual Synthetics.

Data Domain currently supports Solaris, Linux (RedHat and SuSE), and Windows Server (2003 and 2008) media servers today and we are planning to expand platform support to other operating systems.

Also, we would like to leverage the tighter integration OpenStorage allows between Data Domain systems and NetBackup to make optimizations that improve throughput performance.

Q. You company more than any other has linked itself to deduplication. What sets your deduplication solutions apart from the competition?

A. The differentiation between the Data Domain offering and competitive offerings starts at the architecture level. Data Domain considers data de-duplication to be a storage fundamental, not a "bolt-on" technology to be added to an existing storage system. It’s an integral element of our product design and our architectural approach. In addition, Data Domain’s SISL (Stream Informed Segment Layout) architecture uniquely exploits the continued scaling of CPU speed to provide high throughput inline deduplication with economical storage hardware. To put it in perspective, CPU processing, according to Intel research, will have grown by 175x between 1996 and 2009 where as HDD performance has scaled by about 1.3x over the same period.

The high performance inline de-duplication provides an ability to greatly reduce the amount of data that needs to get transferred and the time to create a DR copy using WAN efficient replication. This allows for effective of use of existing network, minimizes the use of tape for the purpose of getting data offsite, and minimizes the time needed to get the data offsite. Enterprise customers are still at risk until the backups are safely replicated to the DR site. Post-processing deduplication systems cannot achieve this level of efficient time to DR this because the creation of off-site copy cannot start until the backup followed by de-duplication are complete in sequence. Finally, because our systems "de-dupe" data in-line, there’s no need to overprovision massive amounts of disk to cache data before de-duplicating it.

Data Domain’s field-proven, reliable, fast and easy to manage deduplication storage systems provide a disruptive economic differentiation for backup deployments seeking greater retention, replication and recovery speed. The same system provides efficient deduplication storage for many other Nearline applications, such as archiving, bringing similar impact to the economic equation. Further, Data Domain’s Retention Lock software also makes archiving on Data Domain systems useful for IT organizations challenged with compliance to corporate governance policies.

Q. Among those companies who are implementing Open Storage, what differentiates Data Domain from the competition?

A. Data Domain actively collaborated with Symantec on the development and is the first vendor to have a Symantec certified OpenStorage solution. We clearly saw the additional benefits of increased integration with NetBackup using OpenStorage in terms of simplifying disaster recovery and as a foundation for additional integration going forward.

Data Domain’s in-line deduplication architecture lends itself to integration with optimized-duplication, an anchor feature of OpenStorage solution and storage lifecycle policies. Unlike post-processing de-duplication storage systems, which have to wait until the deduplication is complete before optimized duplication can be initiated, Data Domain systems can greatly reduce the time-to-DR as the duplication to DR site can be initiated as soon as the backup process is initiated.

Data Domain’s commitment to OpenStorage solution is reflected in the support of the broad range of media

server platforms.

Q. Your company has a number of products related to backup, archiving and file storage. Where do you see the balance between tape and disk in the future?

A. Tape has been losing ground to disk based de-duplication storage systems, which are drastically simpler to use and over time can deliver an order of magnitude greater data reduction than traditional tape-based compression. We see disk based backup systems becoming more and more main stream as tape is relegated for retaining copies for regulatory compliance or completely eliminated from customer environments. This is certainly reflected in the decline in revenue of tape based systems from vendors such as Quantum and Overland. For example, Quantum’s tape automation systems revenue declined by 22.5% YoY for the most recent quarter and this has been a continuing trend. We’ve already made tape consolidation a reality for our multi-site customers, but some larger enterprises still have compliance policies that, many years ago, were built around tape.

Q. Do you use NetBackup and OpenStorage internally for production backups?

A. Internally we use NetBackup in our backup infrastructure with NFS and it has been very stable for us. Our internal IT team is currently evaluating a migration to OpenStorage from an NFS Data Domain disk based solution. Our IT team is very much interested in realizing the benefits of the joint solution and to provide valuable product feedback.

Q. If a customer wants to know more about NetBackup and Data Domain's OpenStorage plug-in where should they go?

• * Customers can visit Data Domain website for more information on Data Domain’s OpenStorage software 
• * A whitepaper on Data Domain’s OpenStorage software is also available
• * They can also get in touch with their local Data Domain sales representative for more information and demos.

Thanks for taking the time to answer my questions. It’s great to hear more about how our collaboration is paying off for our customers.