Object Storage: A Data Storage Technology Whose Time Has Come

October 18, 2017
Cloud, Storage

Object Storage: A Data Storage Technology Whose Time Has Come

Object storage isn’t exactly a new technology; it’s been around since the ‘90s. But, it is a very timely data storage option — and one that is now available from US Signal.

Object storage offers a cost-effective way to store large quantities of unstructured data.  It’s “timely” because unstructured data is growing exponentially. According to analysts at Gartner, we’ll likely see an 800% increase in data volume over the next five years with 80% of it residing as unstructured data.

Much of it will be in the form of emails, videos, social media, RSS, medical records, legal documents, web pages and other content types, as well as information produced by Internet of Things (IoT) devices. Increasingly, organizations are relying on this data for predictive analytics, automated reasoning and other forms of business intelligence.

Many, especially those in regulated industries such as healthcare, retail and financial services, are also required to comply with laws or policies governing how long unstructured data must be retained, where it resides, how accessible it is and who has access to it. Failure to do so subjects them to potential legal issues, penalties and fines.

The Challenges of Unstructured Data

Among the problems with storing unstructured data is that it tends to be written only once and accessed infrequently. As such, it can take up a lot of expensive block and file storage capacity. Tape is cheaper per terabyte. The tradeoff is unacceptably high latencies.

Distributed IT environments also create concerns. Storing unstructured data at multiple sites and through public cloud deployments increases risk. Data center resources become fragmented. Local storage is not efficiently planned or used, and is uncontrollably spread across multiple storage devices.

There are also classification issues. With many types of unstructured data, such as jpegs and video files, there is no way to know what the data is other than by its filename and extension. It’s difficult to determine if it’s something that should be retained or disposed of; stored indefinitely or kept readily available; or placed in a set storage tier or policy-based, tiered storage system. Plus, searching for specific content stored in a large file hierarchy consumes considerable time and CPU resources.

The Benefits of Object Storage

Unlike storage systems that organize data in files or blocks, object storage uses objects. Each object contains the actual data, along with metadata — the contextual information such as policies for retention and deletion or data protection.

The objects are stored in flat address spaces called buckets. When the object is stored, a unique identifier is created for it so it can be located within the bucket. Because objects aren’t restricted by a hierarchical file structure, the number of object identifiers can be increased for almost limitless scalability. And, if an object is changed, it will be stored as a new object to prevent corruption through simultaneous changes.

Applications can quickly retrieve any desired object (data) via the Internet from anywhere and any device using an HTTP-based REST application programming interface (API). It’s simply a matter of using the object’s identifier or querying the metadata with simple commands such as “get,” “put” and “delete.”  

Object storage also employs erasure coding to help maintain data integrity. It restores lost or corrupted data by using redundant chunks, like how RAID works at the disk drive level.

The Best Use Cases for Object Storage

Object storage works best for storing unstructured data that will likely not change and is non-transactional. It won’t be needed often, but may require anytime, anywhere access. It won’t require low latency or high performance, but will need high resilience and durability against disk, node or site failures, as well as bit rot. Use case examples include backup and archiving and storage of large data sets or large size files. 

Object Storage Costs and Cost Savings

The cost for object storage is usually a recurring monthly charge for a specified storage amount per data center, bandwidth usage between data centers and API transaction fees. However, some providers, like US Signal, don’t charge API transaction fees or bandwidth charges in their base pricing.

US Signal Has Your Data Storage Solutions

US Signal can help you assess your data storage needs and select the storage solution that can best meet your business requirements and budget. To learn about US Signal’s new object storage service or other storage solutions, call 866.2. SIGNAL or email [email protected].