Vox Media Case Study

Company automates archive process and eliminates steps to accelerate workflow by 10X

Vox Media is a digital media company known for its high-profile brands including Vox, SB Nation, The Verge, Racked, Eater, Curbed, Recode, and Polygon. Headquartered in New York, the company also has offices in D.C, Los Angeles, Chicago, Austin, and San Francisco.

The Challenge

Vox Media creates digital content that caters to technology, sports, and video game enthusiasts. At any given time, the firm has multiple active projects, including videos, podcasts and other digital content for a diverse audience across multiple sites. Together, these digital projects drive demand for multiple petabytes of data storage.

Quantum StorNext provides the central storage management point for Vox, connecting users to both primary and Linear Tape Open (LTO) archival storage. In their legacy workflow, the digital assets currently in use were stored on SAN. Once a project was completed, those assets were migrated to LTO tape. A NAS system served as a staging area between the SAN and the LTO archives.

Figure 1: Legacy media workflow at Vox

The legacy workflow had multiple issues which prevented the team from meeting its goals. First, it required many manual steps, making it cumbersome and susceptible to failure. With over 150 content creators producing rich media assets and only two media managers to move content to archive, the system could not keep up with the growing data volumes. As a result, completed projects remained on expensive SAN storage, increasing costs and exposing media managers to potential data loss in the case of SAN failure.

An additional issue was the inability to search media. Although Vox had an Evolphin Media Asset Management (MAM) system, this system could not be integrated with their LTO tape library. Lack of an effective search tool made it time-consuming to find archived assets.

What was needed was a storage system that could accelerate their workflow, help get data off the SAN faster, and accelerate search in conjunction with the Evolphin MAM systems. To accomplish this, it was also imperative that the new solution integrate seamlessly their Quantum StorNext.

The Solution

Vox conducted a thorough search and ultimately chose Cloudian HyperStore object storage as its next-generation archive system. In their evaluations, Vox found HyperStore to be an extremely scalable and flexible solution. And because HyperStore was certified interoperable with Quantum StorNext and Evolphin, it would therefore integrate seamlessly with the existing environment.

Figure 2: Current media workflow at Vox

The team deployed a 1.3 petabyte cluster at its New York headquarters and began moving data to it. Because HyperStore interfaced directly with the StorNext SAN, the new workflow eliminated the need for the NAS staging area, immediately accelerating the archive process.

“With Cloudian HyperStore we have sped up our data archiving workflow by 10X, allowing us to save cost and increase productivity by freeing up our primary storage for active projects,” said Sarah Semlear, director of post-production and technical operations at Vox Media.

The new workflow also eliminated the manual steps required with the LTO system, enabling prompt archival of completed projects onto HyperStore and freeing up expensive primary storage. Sarah added, “Implementation was straightforward. The Cloudian system integrated easily and was up and running with our StorNext right away. The Cloudian team was great to work with. They stuck with us through the process, even helping to configure the workflow to further increase performance.”

Looking Forward

In the near future, Vox plans to integrate the Evolphin MAM into the workflow. Cloudian is proven interoperable with Evolphin, and this integration will allow Vox to rapidly search media assets across all storage tiers.

Figure 3: Future media workflow at Vox Media

Vox also plans to capitalize on HyperStore’s multi-site capabilities to provide disaster recovery capabilities. HyperStore clusters in San Francisco and Washington DC will allow the team to automate data replication for DR and for content distribution. This data fabric will provide economical data protection, and provide local storage access at remote locations, all under a single management console.

Finally, the team plans to implement a cold storage option, to which the existing HyperStore can push older content for long-term archive. Sarah concluded, “Implementing Cloudian HyperStore has already produced great results, enabling a better workflow that helps us save time and money. Down the road, we see lots of opportunity for growth as we further capitalize on the system’s data management and easy integration with our tools.”