Frank Slootman, President, EMC Backup
Recovery Systems, addresses several of the
key issues that customers face when considering deduplication storage.

Architecture Matters
Deduplication is a simple concept, but building a system that is enterprise class is not. (~3 minutes)

CPU vs. Disk I/O for Deduplication Throughput
In the past 15 years, disk I/O speeds have increased 10-12x, while CPU performance has scaled by a factor of 2 million. Enough said. (~2 minutes)

Why Data Domain?
Data Domain provides a mature technology that has been hardened in customer production environments for years. (~2.5 minutes)

Inline deduplication removes redundant data during the backup process before any data is written to disk. Any method of deduplication that must first cache all of the original data to disk, then deduplicate it as a secondary step, falls into the category of post-process deduplication. Post-process deduplication forces users to buy, provision and manage more disk, thereby eliminating the benefits of reduced costs and simplified management that inline deduplication provides.

Data Domain systems were designed specifically for deduplicating data inline. Data Domain customers can reduce their storage footprint and reduce or eliminate the need for tape backups. Backup data can be reduced by a factor of 10-30x and greater during the backup process and can then be replicated offsite over existing WAN infrastructures for network-efficient disaster recovery protection – all at the same time, i.e. inline.

Post-process deduplication methods present the same disk I/O bottlenecks as traditional VTLs, which require a disproportionately high number of disk spindles just to land the native data to a disk cache. That data must then be read back from disk and deduplicated before any additional processes, such as replication or verification, can occur.

Get the facts about deduplication. Then get inline.