Mind the Tipping Point: Object Storage versus File Storage

Reading time for this article .

Data is a central part of our day-to-day as a society, from our personal electronic devices, like smart phones and tablets, to our professional lives. Major industries like agriculture, transportation, energy, healthcare and finance, to name a few, all rely on it. Whether you’re working behind a computer or at a job site building houses, digital information is an intricate part of your environment. The intrinsic value of data, be it the intelligence, communication or analysis of it, points to a revolution in how we access, manage and consume it. In order to enable organizations to thrive within the unsurpassed data growth of modern times, storage technologies have evolved accordingly.

File vs. Object

Depending on use case, organizations most often face a choice between file storage and object storage for long-term data storage or archive. The route they choose will influence overall cost of implementation and management. Considering archived data is often stored for years, decades or even forever, any associated costs or complexities are significantly magnified over the course of that extended timeframe.

One could say that file-based storage is almost as ubiquitous across the world as electricity. Open systems in data centers rely on standard file access to their data. Every laptop in a typical office building will connect to shared network attached storage (NAS). Furthermore, standard file-based protocols like CIFS/SMB and NFS make it simple for operating systems and storage vendors alike to build plug-and-play file storage solutions for a wide variety of markets and verticals. For data that’s changing regularly, file systems are ideal as they provide benefits such as controlled access and strong versioning.

Object storage, on the other hand, manipulates data in the form of objects. In simple terms, it combines the pieces of data that make up a file, adds relevant metadata, and places everything into a flat address space, also known as a storage pool – a limitless range of metadata that can be assigned to each object, providing unmatched data analysis capabilities. Object storage systems allow users to select the desired level of data protection, setting preferred data tiering policies, and enabling digital assets to migrate across multiple storage targets and geographic regions in an automated manner. The flat address space of the storage pool eliminates the hierarchical folder structure used in file storage, enabling infinite scalability. A custom identifier, or unique object ID, is associated with each object to ensure fast retrieval. These traits found in object storage are the attributes most ideal for data storage rather than data editing or manipulation.

file storage versus object storage - metadata, scalability and TCO tip the scale

Object storage is also not without its challenges. Ingesting tiny objects in an object-based system incurs higher overhead penalties per object (e.g., an email file vs. an 8K resolution video clip), making it more suited to larger objects and managing unstructured data at scale. Another hurdle for object storage systems is the interface protocol. Object storage uses an HTTP-based REST application programming interface to access objects. While it’s advantageous because of its simplicity, this interface is unlike file-based protocols and may require changes to the application code.

The tide is shifting…

“If it’s not broken, don’t fix it,” they say. The problem is, file systems weren’t designed for long-term storage of “fixed” data or content that’s not being manipulated. The overhead associated with file systems is completely justified for dynamic, changing data. That same overhead is nothing but burden when it’s not needed – such as over the course of long-term storage. File systems perform best for smaller files, data that typically needs to be shared locally, as scaling file systems past a certain threshold can quickly become a nightmare. Moreover, finding a file that is months old fixed in a single location in a hierarchy of folders can often feel like navigating a maze due to a file’s limited number of metadata tags.

Object storage is the more efficient way to manage data over time. From vendors that build storage systems, to applications that use and consume data storage, to streaming services that provide most entertainment, object storage is being adopted across the board to streamline data movement. The metadata enables applications and users to browse and see what is available without actually touching the data. Object storage de-couples the valuable information about the data from the actual data itself, typically storing this information as metadata in a separate faster tier. This means it is much easier (and faster) to search through an unfathomable amount of data within an object storage system than a file-based system. Finally, an object-based system’s ability to automatically store multiple copies of data over a distributed system makes it easy to manage cost and enables extremely high data reliability.

Spectra Logic's hybrid storage ecosystem leverages BlackPearl's object storage capacities to manage data in multiple targets.

The idea and concept of object storage has been built into proprietary applications for a long time. However, with the advent of public cloud storage and cloud services, various object storage interfaces have become prevalent. The most popular interface today is S3, created by the largest cloud service provider by market share. Any application that adopts a REST interface will, at a minimum, use S3, but may eventually adopt other object interfaces as well. As more applications adopt S3 and it becomes the de facto standard, it will be easier for vendors to build plug-and-play object storage systems, providing customers with storage platform options where their data is well managed for its entire lifecycle.Organizations should look for storage solutions that provide the right set of features to achieve their goals while managing costs. Object storage’s gaining momentum makes it the clear choice for future-looking systems. Whether your organization is ready to transition to object storage now, or predicts your data will scale beyond your current storage’s capabilities in the future, Spectra Logic has you covered with the object-based Spectra® BlackPearl® Converged Storage System and upgradeable BlackPearl NAS.

BlackPearl Network Attached Storage - a NAS solution that easily scales to an object-based platform to leverage a multi-tier architecture

2 Responses to “Mind the Tipping Point: Object Storage versus File Storage”