PowerEdge: DR4000 Overview

The DR4000 is a high-performance, disk-based backup and recovery target appliance that is simple to deploy and manage. The appliance is available in 2.7 TB, 5.4 TB , and 9 TB post-RAID storage capacities (with future support for an 18TB post-RAID model) and is ideal for deployment in small enterprise and remote office environments.

The DR4000 boasts an impressive feature set, which includes:

• Inline deduplication and compression
• Deduplicated replication
• Advanced data protection
• Replication licenses included
• Nondisruptive deployment
• NFS, CIFS, and future OST (licenses included) protocol support
• Simplified management
• All-inclusive licensing

DR4000 Building Blocks
The DR4000 firmware implements an inline, content-aware deduplication and compression engine. The software architecture consists of the following building blocks:

• Frontend protocol layer: This layer implements protocol support for incoming I/O. The supported protocols are NFS and CIFS. (OST and NDMP will be supported in future releases.)

• Dedupe/compression engine: This is the key component of the architecture. The deduplication engine implements advanced chunking and deduplication algorithms to achieve space savings.

• Ingest buffers: The DR4000 implements inline deduplication and compression. Ingest buffers, which reside in high-speed NVRAM, receive incoming data flows, and the deduplication engine works on the data in the ingested buffers.

• Disk storage structures: Several key data structures are stored on disk, which enable efficient deduplication and compression operations. These data structures include:

• Dictionary: The firmware maintains a list of references to all unique data chunks that have been encountered while processing the incoming user data. The dictionary is a dynamic list of the all unique patterns and forms the basis for deduplication.
• Data Suitcase: This is the list of unique data chunks. The dictionary contains references to the unique data chunks in the data suitcase.
• Block map/ Object map Suitcase: This is a mapping file that maintains the mapping of user data to chunks in the suitcase. During the processing of user data, if a chunk matches a reference in the suitcase, reference counts to those chunks in the data suitcase are incremented, rather than storing a duplicate copy of the data.