IBM is expanding its own DiamondBack tape library with an on-prem S3 object storage archive. This makes more possible in terms of storage using magnetic tapes.
Back in October, it came out with the DiamondBack TS6000, which can house up to 14 tape drives and 1458 cartridges for 28.8 petabytes of data. There is now also an x86 server present in the DiamondBack S3, Blocks & Files reports. This provides the S3 interface and object-to-tape cartridge/track mapping. DiamondBack S3 could also give object-storage an advantage that already applies to file storage on the medium, with tape being cheaper to store than hard drives or flash.
IBM has long stored data on tapes: the company’s PCs were already using tapes similar to what was common for storing audio at the time. For storage, tape has always been an inexpensive alternative to hard drives. As far as IBM is concerned, the rise of big data is reason enough for hyperscalers to stay true to tape. In addition, the technology is said to be a lot more ecological: tape media reportedly emits only 3 percent of CO2 compared to hard drives.
Early access
Julien Demeulenaere of Tape & High-End Storage EMEA at IBM announces an early access program to deploy tape via S3. He argues that IBM’s solution should remove the complexity of tape technology. For long-term use, he commends the advantage of this single appliance that can be set up in less than an hour at a colocation. The early-access program lasts 14 days on a prototype and offers “unlimited storage.”
Demeulenaere told Blocks & Files, “It’s not a miracle, we can’t store buckets on tape natively. It’s a software abstraction layer on the server that displays the data as an S3 object for the user.” The server can be accessed exclusively through Ethernet. It cannot be used as a file-repository target.
Also read: IBM customers can get their hands on Meta’s LLaMA 2 model