Object storage is a way of structuring stored data so that it's characterized as objects that can be manipulated in different ways by hardware and network storage systems. One important difference between file and object systems is that in object storage systems, the objects are not in a file-folder hierarchy. Some describe objects as a series of buckets that can be delivered through multiple storage nodes, or zones. In this locational system, the metadata is stored with the object. It may help to think of each object as a singe block or chunk of data that does not get stored within a tree design as with more traditional file-folder systems but instead has a label attached to it that shows what is inside.
Like traditional files, objects may have specific signifiers attached to them in order to help a system understand what they contain. This information, often called metadata, allows the system to handle each data object more capably. Building this kind of versatility into systems can help network managers:
- Track data object generation
- Understand how information flows through an IP or network pipeline
- Monitor document conversion and other forms of data handling
- Handle data disposal when data is no longer needed
Object storage models are also useful for the kinds of new data handling procedures attached to fault-tolerant and redundant systems. For example, many systems use a RAID system to write parts of files to a variety of drives or storage locations. New technologies allow for partitioning one physical drive into many storage areas. When object storage models are compatible, this can allow for a much more versatile use of chunks of data that can be easily backed up, evaluated and recalled on demand. Object storage is also very useful in the new distributed file systems that support the enormous data centers being built around the world by government agencies and large private businesses.
0 Comments