How simple objects are stored

An IBM Cloud Object Storage System™ links systems to store data that is divided into multiple slices by using an Information Dispersal Algorithm (IDA) before sending each slice to a storage node. The storage node is called an IBM Cloud Object Storage Slicestor® node. A client application sends object data to a device that slices the data, called an IBM Cloud Object Storage Accesser® node, over an HTTP-based protocol.

When the system stores an object by using the SOH, it assigns the object a unique 18-byte object ID (OID). This OID contains enough information to allow the system to perform a quick read without consulting a centralized routing system.

From the application’s perspective, objects are stored in a large capacity storage system without a hierarchical structure.

If the object changes, it must be written again as a brand new object.