Object storage is a highly-scalable storage model that manages data as key and value pairs in a single flat namespace. All storage schemes employ the concept of an individual identifier for a specific piece of data, such as a physical disk address in a block-based system like a storage area network (SAN), or a file name in a network attached storage (NAS) or file system. In object storage, the keys are the identifiers for the data objects (values) stored within the system. The values are the actual data payloads, or to reinforce the nomenclature, they are the “objects” in the object storage system.
Object storage gives us a way to store and retrieve objects via keys— to illustrate, think of object storage as a valet parking system where you hand over your ticket (the identifier), and what is retrieved to you is your car (the value). You never need to know the physical, or even the logical location of your data object; you need only to have the key to access or retrieve your data.
The flat namespace part of the object storage equation refers to the lack of complex naming hierarchies found in a typical file system, where we have the concepts of folders, sub-folders, and files—but you must know how to navigate the file system to find specific files. In our parking lot analogy, if you were to locate your car yourself without help from the valet (object storage), you would have to remember where the parking structure is, the floor you parked on, and the individual parking space you parked in. A flat namespace within an object store removes the complexity and management overhead of this complex naming structure, making it much more scalable and far easier for the user to find his data (or car in our analogy).
Object storage is most advantageous for use-cases where:
With these use case criteria in mind, object storage is now effectively deployed in massive scale data archives, across private and public clouds, and in large media repositories where the file count can be very high and access to data may require cloud-based streaming.
Object storage is deployed both on-premises and in the cloud. Enterprises are creating object storage-based private storage clouds to capitalise on the scalability benefits discussed above. Service providers and cloud services use it for easy access and for massive capacity scaling, driven by burgeoning numbers of end-users.
Object storage systems should primarily be reliable and durable in terms of protecting data. Next, the solution should be highly available even when failures occur with hardware components such as servers and disk drives, but also with system management software. In addition, object storage solutions should be proven to bring extreme scalability for billions of objects, in support of up to hundreds of petabytes of storage capacity.
Ideally, customers should be able to combine object and file storage within the same platform, this helping enterprises transition to next-generation cloud applications. This means that both legacy applications and new applications can coexist in a single solution to make data management easier. And finally, the solution should be suitable for a variety of data sizes, from small objects and files typical of email/webmail use-cases, to massively sized objects found in backup & archive, financial services, medical imaging, public sector, video surveillance, and media use-cases where massive, high-resolution 4K and 8K video files are the norm.
So the next time you valet park, after you surrender that ticket to the parking attendant and your car magically appears without you ever knowing or caring where it was parked, you may be reminded of the power of object storage at work!