Components
IBM Deep Archive is a long-term archival solution optimized for Deep-but-Accessible data bringing S3 functionality on-premise with tape software innovation. This solution is now available on Diamondback through either the entry configuration and standard configuration. All configuration consist of an IBM Diamondback Tape Library, IBM TapeCloud Hardware, and IBM TapeCloud Software.
Hardware
The system is made up of the following high-level hardware components.
- IBM Diamondback Tape Library – The Diamondback Tape Library provides preserves and protects your data with sustainable, cyber-resilient, ultra-high-density tape storage. The library is populated with up to 14 tape drives in a standard configuration and up to 1584 LTO 9 tape cartridges providing 27 PB of native capacity. The TapeCloud nodes are connected to the library internally through a private set of Serial attached SCSI (SAS) connections providing up to 400 MB/sec native data rates per drive.
- IBM TapeCloud Hardware – The hardware include in the top rack of the Diamondback Tape Library includes the TapeCloud nodes that provide a landing area for data being written to or read from the S3 Glacier interface. These TapeCloud nodes contain all TapeCloud software required to manage S3 communications as well as the intelligence to optimally tier data to and from the tape library in response to S3 requests. Each TapeCloud node has up to 50 TB of RAID 5 protected NVMe flash storage and provides S3 access through 25 Gigabit Ethernet connections.

Configuration options
This choice determines the number of IBM LTO 9 Tape Drives and TapeCloud nodes installed in the library top rack.
Entry configuration
The entry configuration provides 4.1 TB / hour data transfer speeds.
- IBM Diamondback tape Library
- Up to 27.8 PB (9 PB lowest entry recommended)
- 4 LTO-9 SAS Tape drives
- 5U Top rack
- Inside 5U Top Rack
- 1 TapeCloud node
- Internally managed SAS cabling connecting to 4 tape drives
Capacity options
For both the entry and standard configurations, this choice determines the number of IBM LTO 9 tape cartridges installed in the library and whether these are provided at the time the library is shipped or by the user in preparation for installation. The IBM Diamondback Tape Library has a total of 1584 storage slots for tape cartridges. These storage slots are arranged as 176 (in 4 columns and 44 rows) ultra high-density slots each of which stores 9 tiers of cartridges stacked one after the other. The cartridge accessor must run a shuffle operation to access the cartridges that are stored in tiers other than the top tier (Tier 1). A shuffle is the process of moving cartridges in lower tiers into the gripper (or other available storage slots) to access cartridges in higher tiers.
- Half capacity with prepopulated media (800 tape cartridges, 14 PB native capacity)
- Full capacity with prepopulated media (1548 tape cartridges, 28 PB native capacity)
- Provide your own media in any quantity
Software
The TapeCloud software running on the TapeCloud nodes consists of the following components…
- S3 interface – This handles S3 requests via port 443 securely with HTTPS. Each S3 tape library represents a single S3 namespace. By default, self-signed certificates are setup but can be modified later to match your security requirements.
- Distributed file system - All data entering the system via S3 PutObject requests is staged onto a sharded file system that is distributed over the flash storage on all TapeCloud nodes prior to being written to the tape library. This space is also used to stage data in response to S3 RestoreObject requests.
- Library manager – The management of the tape library is controlled by the TapeCloud nodes using the library manager. This component ensure data is optimally moved to and from the tape library ensuring correct layout on tape cartridges and optimal utilization of tape drives. Data written by the library manager is stored on cartridges using the LTFS format.