This article can also be found in the Premium Editorial Download "Modern Infrastructure: Cloud storage at your service."
Download it now to read this article plus other related content.
Some -- not all -- cloud storage services are based on object storage back ends that don't bear much resemblance to the block and file storage devices normally seen in the enterprise. What is object storage, and why do cloud providers rely so heavily upon it?
Unlike other storage technologies, object storage does not store data as hierarchical files, but packages that data into an "object" with a unique identifier or key, plus any associated metadata, that are then placed in to a flat namespace or "bucket."
The large cloud providers love object storage because it is consummately scalable. To add capacity, just add extra nodes; the underlying object storage platform takes care of replicating the objects to meet the necessary redundancy. Cloud providers also love it because it's cheap. Most object storage systems are built on top of commodity servers and disk drives, without expensive RAID and disk controllers.
But object storage is new enough that traditional applications such as backup, collaboration, document management and the like have not yet been updated to recognize it as a target. Organizations that want to leverage object storage need to rewrite their applications for it, or deploy cloud storage gateways that can translate between file, block and object formats.
Nor is object storage particularly fast, making it a poor fit for applications such as databases that need to make frequent updates to the data. As such, object storage is generally used for data that is not often updated, such as backup, archives and media files.
About the Author:
This was first published in November 2013