Storage

Why data center managers shouldn't overlook object storage backends

Object storage is mainly associated with web-based applications and the cloud. But, the backend is proving to meet the needs of traditional NFS use cases as well.

dcstorage.jpg
Image: iStockphoto/Vladimir_Timofeev

When hearing the term object storage, the first thing that comes to mind may be cloud. Amazon has made S3 synonymous with the term, and it's common to see object storage in cloud-native applications. However, outside of cloud-native applications, there is plenty of interest to the enterprise. Enterprise data center managers would do well to pay attention to object storage backends.

A matter of perspective

It's tough to make a case for object storage at the presentation layer in the enterprise. Solutions such as Dropbox utilize object storage on the backend. However, the front end is a traditional, tree-based presentation layer. Even the web interface for Dropbox presents the common file/folder view of your data.

It's the backend that offers an interesting objection for enterprises. I asked storage expert and consultant Enrico Signoretti of Juku.IT about enterprise use cases, and he said that the API nature of object storage is an awkward fit within the enterprise. Signoretti advised to look beyond the front-end, as object storage offers many advantages for backend systems. And, IT vendors have started to pay attention to the qualities of object storage for enterprise solutions.

Scalable backend

Object storage doesn't carry with it the burdens of a file system. Object storage relies on the application to present data in an organized manner to end users. As such, object storage systems don't have the overhead associated with file systems. Much of the complexity of modern day, scale-out storage solutions originates in balancing metadata and data across nodes and sites. In object storage solutions, the metadata is a part of each object.

See: How to deploy a Ceph storage cluster (TechRepublic)

The storage of the metadata along with the data object results in a robust system for unstructured big data elements. It also means expansion and replication are simple. The shared metadata between nodes and sites is just the table storing the location and replication policy of the entire data store. The methods used to distribute that central data varies from solution to solution.

Most object storage platforms produce objects at a fixed size. Having a fixed size creates predictability in performance and replication. Network engineers can leverage the predictability of storage replication in their designs. If IT combines the attributes of object storage, the result is a system that can scale into the thousands of petabytes at a predictable cost.

Enterprise solutions

With all of the backend advantages of object storage, we are starting to see solutions catering to enterprise use cases. The primary enterprise use for storage is still NFS and SMB. Many solutions build backend systems with object storage and an NFS or SMB proxy to present the storage to users.

The approach combines the best of both worlds. End users get the consistent experience of file/folder access to data, and data center managers can create scalable, cost-effective solutions.

Also see

About Keith Townsend

Keith Townsend is a technology management consultant with more than 15 years of related experience designing, implementing, and managing data center technologies. His areas of expertise include virtualization, networking, and storage solutions for Fo...

Editor's Picks

Free Newsletters, In your Inbox