Object storage is a method to store data. Most IT professionals are familiar with the more traditional Portable Operating System Interface (POSIX) file system approach. Both are a way of organizing and storing data on a storage system. POSIX mimics pre-computer method of storing data, typically hard copy papers, into folders and filing cabinets or volumes or directories. Object works like a computer, objects (a.k.a files) are given an ID instead of being placed into a hierarchical directory structure. The difference in the storage and organization of data is what makes object storage so ideal for unstructured data.
The most noticeable difference between a POSIX compliant file system and an object storage system is the retrieval of data. A POSIX file system requires navigation through a series of volumes, folders and sub-directories. An object store merely requires the unique ID. A request for the ID goes to the metadata engine and it points the request to the exact location of the file.
Back when users did most of their data retrieval, navigating through a directory structure made the most sense. But today applications do most data retrieval, not humans. Even in the most basic of retrievals, accessing a document is often done through a content management or web application.
The Unique ID Advantage
The unique ID is created by performing a checksum or hash on the file. A file with the same data should always create the same unique ID. This means that the object storage system can eliminate data redundancy and insure a single instance of each file. In addition, the object can physically move anywhere within the object store and it is still found by the same unique ID. The application does not need an update with the new location as it would in a POSIX file system.
From a data protection perspective an object store can use the unique ID to confirm data integrity, the ID should always generate the same result. It can also use the ID to make sure that multiple copies of the object are available on different nodes, clusters or geographic locations.
Finally it provides a universal method of placing and retrieving data. For example, most cloud providers use object storage. Once the organization updates its application to support object storage, extending to a cloud provider is an easier step. Object storage methods are also extending to tape devices. Several tape vendors can present their tape library as a disconnected, offline object store.
Object Storage Challenges
The challenge facing object storage adoption is that in theory the application needs updating to access objects through IDs instead of mapping to a file. These code changes should be relatively straightforward but can take some time. And of course there are plenty of situations where the organization simply does not “own” the code. To help give organizations time to update applications and to support applications that can not be updated, IT professionals should look for object storage systems that support a variety of legacy protocols, giving them a bridge to object storage.
The best approach to understanding object storage is to not overthink it. It is simply another way to access data, then leverage unique IDs generated from the data it stores. While this method requires application updates there are legacy gateways that can bridge the customer to the future. With the “what” of object storage addressed the next question is “why”, which is the focus of our next entry in this series.
Sponsored by Dell EMC®
Dell EMC®’s Elastic Cloud Storage (ECS™) will bridge the gap for traditional and next-gen applications on a single, multi-purpose object-based platform. ECS can manage your workloads for today, and better ready your organization for the digital transformation of tomorrow.