IBM cloud will reach back into tape for low-cost storage

At its Edge conference in Las Vegas this week, the company will preview an archiving architecture that can span all tiers of storage from server-based flash cache to tape, moving data to the best and most cost-effective tier at any time based on enterprise policies.

Enterprises are accumulating growing volumes of data, including new types such as surveillance video that may never be used on a regular basis but need to be stored for a long time. At the same time, new big-data analytics tools are making old and little-used data useful for gleaning new insights into business and government. IBM is going after customers in health care, social media, oil and gas, government and other sectors that want to get to all of their data no matter where it's stored.

IBM's system, which it calls Project Big Storage, puts all tiers of storage under one namespace, creating a single pool of data that users can manage through folders and directories without worrying about where it's stored. It incorporates both file and object storage.

The system can be implemented on a company's own premises, as a cloud service, or as a hybrid. With it, IBM says it will be able to offer a higher quality of service than established cloud providers like Amazon or Google for a lower price. Key to the low cost is the use of tape, which is part of IBM's long heritage in storage and is being incorporated in a multi-tier, active cloud storage service for the first time, said Bernie Spang, vice president of software-defined infrastructure at IBM.

Specifically, Project Big Storage will cut the cost of retrieving your data from the cloud and make those costs more predictable, he said. Though there's nothing simple about calculating the cost of storage as data automatically moves across multiple tiers, IBM will let clients see the costs as they're incurred.

Project Big Storage is entering a pilot phase with select customers who will help to shape the eventual product. The pilot should take just a few months, Spang said. When it comes out, it will be offered as a cloud service through the IBM Cloud Marketplace and the BlueMix cloud development environment.

Tape costs less per bit of stored data than flash or hard drives do, though it can take longer to deliver the information. Project Big Storage will have multiple tiers of tape storage, some with retrieval times measured in minutes and some that can take a few hours.

If the system works as intended, only data that's rarely needed will end up in that coldest tier of cold storage. It uses IBM's http://www.computerworld.com/article/2885129/ibm-puts-software-and-cloud-at-the-center-of-storage.html Scale software, an evolution of GPFS (General Parallel File System), a distributed storage technology that includes automatic tiering based on enterprise policies. Another key piece is IBM's Spectrum Archive tape technology, formerly LTFS (Linear Tape File System). Within their own premises, customers can include non-IBM storage products in the system.

The cloud capability is provided through the company's SoftLayer infrastructure. To run the tape portion of the cloud service, IBM partnered with Iron Mountain for its storage facilities and expertise in compliance and tape handling, Spang said.

Stephen Lawson covers mobile, storage and networking technologies for The IDG News Service. Follow Stephen on Twitter at @sdlawsonmedia. Stephen's e-mail address is stephen_lawson@idg.com