Object storage is a means of storing data, that organizes data into flexible-sized data containers called objects. These objects contain data as well as its associated metadata (an extensible set of attributes describing the object). Instead of a hierarchical collection of directories, folders, and file names, objects are organized as a flat collection of unique Object IDs that are assigned on creation and used to access the object later. Object storage can be dispersed across local area and wide area networks. Some object storage “slices” across a number of storage devices to enhance availability and security. With this approach, object data is highly available because slices include redundancy and are spread across devices and locations so data is highly resilient from component failures and even site disasters. Many object storage solutions accomplish this resilience by simply making duplicate copies of the entire data across storage devices as well as across geographic locations.
Disk archives that combine object storage with industry-leading erasure code to deliver maximum availability.
When traditional RAID can't meet demands for petabyte-scale capacity with both high reliability and fast access, the answer is object storage.