IBM: S3 Glacier Storage Classes Bring Benefits of Tape into Mainstream

By Shawn Brume, tape evangelism and strategy, IBM Corp.

Every industry stores massive amounts of archival data, with new creation of data accelerating.

Although every organization has unique data access requirements, a large amount of archive data quickly moves to a rarely if ever accessed state. End users and the applications they utilize demand access to data; the challenge has been how to integrate data access with long-term storage and data resiliency. This has given rise to the deep-but-accessible archive.

Modern data centers have been specifically focused on SSD and HDD implementations at the lowest possible operational impact. Object Storage has been the crucial implementation to supporting billions of data assets on HDD at the lowest possible cost. Even lowering costs with object storage have not been enough to keep up with the growth of data and more importantly the growth of infrequently accessed but retention required data, deep archives.

Tape is so hard to use.
As a tape evangelist, I can clearly say this is a valid statement for organizations and storage admins that have never deployed tape. Although nearly all users of tape will also say that once tape infrastructure is running, it is easier to manage per petabyte of data than active storage (SSD or HDD). 

A big part of the difficulty in using tape has been the requirement for independent Software Vendors (ISV) to manage the data and interface to tape. Enter the state of complexity: servers, software management, storage administrators with specific skill sets, and no direct end user or application access to data. One of the reason ISVs have been required is because tape is a linear access device, data streams faster from tape than from HDD, but the time to random data has a high latency. That is all solved with IBM Storage Deep Archive with S3 Glacier Flexible Retrieval storage class.

IBM Storage Deep Archive provides a standard S3 interface, with full data storage placement on tape, without special tape skills, the application and the administrators simply must support the S3 APIs. S3 Glacier Flexible Retrieval storage class handles the latency of tape for users and supporting applications.

All this comes back to reducing the impact of retaining deep archival data. IBM Storage Deep Archive reduces energy consumption by up to 97% and lowers carbon footprint by up to 85%, when compared to object storage on HDD. Compared to AWS S3 Glacier Deep Archive IBM Storage Deep Archive results in a storage savings of up to 85%, and that is before adding in retrieval costs! Since deployment of tape has been viewed as difficult, the company has delivered IBM Storage Deep Archive in an install-and-go configuration. The IBM Diamondback can be installed in less than 30mn, once the services are brought on-line the system is ready to use, no different than the requirements for bringing an HDD object storage on-line.

IBM Storage Deep Archive on Diamondback provides the S3 Glacier compliant interface, up to 27 Petabytes of uncompressed data archives and up to 16.1TB per hour of uncompressed data transfer.

If your organization is considering enhancing storage data placement, or have a Cloud-First initiative, the CFO will thank you for bringing IBM Storage Deep Archive as a solution to reduce the cost of information technology, the IT team is going to get the benefit of reduced complexity for storing data in the hybrid infrastructure, and the Information Security and Governance team will thank you for the security and control of on-premises data storage.

More at: IBM Storage Deep Archive 

1.     S3 Glacier Flexible retrieval storage classes
2.     Subject to future product release, not a commitment
3.     Compared to comparable HDD Object Storage
4.     When compared to AWS S3 Glacier Deep Archive