Why Object Storage
Cleversafe utilizes a simple or named object storage approach that efficiently stores billions of data objects in a single name space and exposes the data through REST, a standard HTTP based protocol.
The Old Way: File Based
Traditional storage systems typically use an underlying file system and expose the data via NAS-based protocols like NFS and CIFS. The file system approach is ideal for human users but not for big data applications that must manage billions of objects. File systems allow humans to organize content in an understandable hierarchy where access speed is not critical and directory contention is minimized. Once you are dealing with billions of objects, a file system approach does not scale and incurs performance breakdowns and bottlenecks.
The New Way: Object Based
Cleversafe has made it easy and painless to integrate your applications with our object-store interface through Simple or Named Object HTTP. The client application houses the metadata and the Simple or Named Object interface is accessed using HTTP/REST API. Similar to retrieving a URL instead of asking for a Web page, you are asking for an object. The process tolerates Internet latency, provides for programmable storage and efficient metadata management.
Our method for managing metadata uses storage nodes working in conjunction with every other node in the system to address massive amounts of data in a small footprint. Essentially, each storage node can keep the entire namespace mapped within its memory and know precisely which storage nodes contain slices of an individual object.
Unlike the Hadoop File System which has the concept of a master metadata node which knows where all the objects reside, a Cleversafe dsNet storage system utilizes the Yottaspace technique to reference individual slices of an object and can recreate the entire object segment quickly.
Cleversafe offers a superior approach to the old way of doing things: one that ties your applications more closely to your storage, avoiding the scalability limitations of the file system. Our approach provides easier storage provisioning and allows for access control to be set at the object level.
By leveraging an underlying object store, several advantages are realized:
Referencing objects by IDs versus file names enables easier movement and migration of data, freeing it from underlying hardware
Scalability of namespace
The namespace is independent of the file and operating system and isn’t impacted by their size limitations
Object storage enables unlimited simultaneous readers and writers directly to the objects
Simplified integration & development
Tighter coupling between applications and storage yields enhanced features
Objects take up only the space they need, without having to pre-allocate storage for the storage container
Learn more about Cleversafe’s Object-based access methods.
For more information, please visit Cleversafe's Library