Data Domain Lauches a Faster Inline Dedup Appliance

Data Domain announced the DD690, the industry’s highest performance inline deduplication storage system for backup and other nearline applications. With aggregate throughput up to 1.4 TB/hour, and single-stream throughput up to 600 GB/hour to enable protection of large databases in short backup windows, the DD690 establishes consistently high benchmarks across the spectrum of common datacenter backup metrics. A fully configured Data Domain DDX Array with 16 DD690 controllers increases aggregate throughput performance to up to 22 TB/hour and offers up to 28 petabytes of usable capacity, delivering the capability for long-term online retention to large, consolidated datacenters.

The DD690’s price/performance is enabled by the Data Domain SISL (Stream-Informed Segment Layout) scaling architecture to minimize the number of disk accesses required in the deduplication process. Unlike many competing approaches, Data Domain systems ride the price/performance wave of multi-core processor architectures rather than depending on over-provisioned storage subsystems for throughput. As a result, Data Domain systems based on dual-socket controllers and minimum SATA-based storage configurations have improved in throughput by nearly a factor of ten from 2004 (150 GB/hour) to 2008 (1.4 TB/hour).

With the Data Domain Replicator Software option, the DD690 can automate WAN vaulting for use in disaster recovery (DR), remote office backup, or multi-site tape consolidation. A single DD690 system can support a replication fan-in from up to 60 remote offices using smaller Data Domain systems such as the recently announced DD120. The DD690 can deduplicate globally across remote sites, further minimizing required bandwidth since only the first instance of data is transferred across any of the WAN segments.

"Data Domain continues to scale its systems to address the demands of ever-growing environments," said Heidi Biggar, analyst at the Enterprise Strategy Group. "The DD690 raises the bar significantly in dedupe performance, capacity and replication functionality. These are just the type of enhancements that users — particularly larger ones — tell us they are looking for."

Like all Data Domain systems, the new DD690 is simple to install and flexible enough to be implemented into existing user environments without disruption. Backed by Data Domain’s available 24x7x365 enterprise class service, the DD690 supports leading backup and nearline software products on any standard storage fabric, including a new 10Gb Ethernet option. The DD690 is available as an easy-to-deploy appliance, or as the DD690g Gateway, which supports external SAN disk array storage.

The J. Craig Venter Institute (JCVI), a genomics research organization based in Maryland, has been beta testing the new Data Domain DD690 since early March, and recently moved forward with the purchase of the system. "We had a very positive experience with the DD690," said Eddy Navarro, Storage Team Lead for JCVI. "JCVI runs a Linux based high performance grid computing environment. We’ve been very pleased with the performance and the strong deduplication ratios we’ve achieved, and the DD690 works very well with our Symantec NetBackup software. In addition, the IP-based replication capabilities and the availability of a gateway model of the DD690, which we have also now ordered, were definitely an attraction for us."

Why Architecture Matters
The DD690 is based on the same CPU-centric approach to inline data deduplication as all Data Domain systems. Unlike most deduplication approaches that are added as afterthoughts to existing disk arrays, Virtual Tape Libraries (VTLs) or backup software, Data Domain’s combined efficiencies include:

• SISL scaling architecture leverages CPU improvements to increase deduplication speed inline while minimizing reliance on disk accesses for performance.   Data Domain systems have delivered consistent improvement in throughput performance by nearly 10x and in capacity by nearly 30x over the last 4 years.  Based on Intel’s CPU roadmap, increased throughput is expected to continue growing in the future.
• High Performance Inline Deduplication for simplicity, to minimize system resources, administration, and internal system process contention.
• Green Storage Efficiency means a smaller system footprint, lower power consumption and less hassle.
• Easy Infrastructure Integration with support for nearline, backup or archive workloads, as a local consolidated storage tier and for remote networked DR.
• Data Invulnerability Architecture defends against data integrity issues by providing continuous verification during storage and recovery of data.

   
"The DD690’s performance is phenomenal," said Brian Biles, VP of Product Management at Data Domain. "While most of our competitors will compete by highlighting disk subsystems that are vastly larger than deduplication storage requires, the DD690 shows that with the right architecture, the simplest configuration can also be the most powerful."

The DD690 will be generally available in the second quarter of 2008 through authorized Data Domain channel partners.