Manage Learn to apply best practices and optimize your operations.

Does RAID have a place in object-oriented storage environments?

Expert Brien Posey explains whether RAID technology can bring any benefits to an object storage environment.

Brien PoseyBrien Posey

Does RAID have a place in object-oriented storage environments?

Object-oriented storage is quickly becoming the standard for storage in the cloud and in the enterprise, but does RAID have a place in object storage environments?

Object storage was actually created as a RAID replacement. Although RAID has been the de facto standard for data protection for decades, RAID has trouble scaling to match the needs of today's larger environments.

One problem with RAID is the time it takes to rebuild the contents of a failed multi-terabyte drive. In addition, when using RAID 5 or RAID 6, storage performance can be significantly diminished while the array is rebuilt.

Fault-tolerant RAID arrays are designed to protect against disk failure, but as the number of drives in an array increases, the likelihood of a failure also increases.

Object-oriented storage is designed to address these problems by dealing with objects rather than file systems. In an object storage environment, each disk's contents can be mirrored multiple times. That means if a disk were to fail, the storage environment can simply redirect requests to a different copy of the disk. The performance impact involved is minimal and this approach also reduces the urgency of replacing failed disks because several other replicas of the disk may be readily accessible.

Going back to the original question, I think an argument could be made that object-level storage uses a variation of policy-driven RAID, even if vendors do not explicitly use the term RAID. After all, most RAID offerings have two-way or three-way mirroring capabilities. Object storage also uses multi-copy mirroring, but this is usually referred to as replication.

Dig Deeper on Object storage