Content value is a puzzling thing. It is difficult to make any sense of what content the public will want and when, and then proceed to put a value on it.

Consumers demand a greater amount of content delivered across a larger variety of distribution points, expecting content anytime and anyplace on devices ranging from smartphones and tablets to TVs and computers. If content is not readily available, then people will look elsewhere, or for the next thing. Content life is cyclical, and to take best advantage of the cycle, content owners must have content ready to go when it is requested.

Even content relationships are strange. Often when a certain type of content is popular, associated content also gains popularity — and quickly. Many hours of analytical study have been and will be performed to evaluate this behavior and better predict it. So far, the results are inconclusive, and content owners are trying a wide variety of tactics to capture the greatest share of viewers.

Despite these uncertainties, one thing is sure: Keeping content archived and quickly accessible is critical.

The Changing Role of Archive

Adapting a content archive to current demands is not always as easy as it seems. Standard definition (SD) video and terrestrial broadcast content were easy to archive, but gone are the days of racks of video tape numbered and bar coded into a database for easy access from a shelf to a tape machine. Today, most content is archived on digital tape with a vast majority “vaulted away” with a directory maintained by media asset managers. As higher-resolution content takes over, more and more tapes are generated for the same program time. As these valuable assets sit on a shelf, they deteriorate. Periodic rewriting of tapes on a scheduled basis is a necessity.

Every content owner knows that retrieval from tape to meet critical consumer demand timeframes is lacking, but up until now, there has been no good alternative. All that is changing, and changing rapidly, thanks to technology first developed for long-distance communication in space. This technology, known as object storage, when used with erasure coding, delivers a superior archive experience. A proven technology that has been in use for more than 20 years, object storage today powers many of the world’s largest cloud storage providers.

While object storage is not a magic bullet that can solve every archive problem, it is much easier to maintain and more effective than tape-based models in today’s “always on” content-driven society. It’s not just that object storage with erasure coding is reliable; it also offers (depending on the system chosen) virtually unlimited scaling, geographical independence and significantly reduced maintenance. Put it all together, and it is a pretty compelling archive platform for the foreseeable future.

The basics of object storage with erasure coding are pretty simple. Using the lowest cost, largest capacity hard disk drives available, combined with some smart software and common off-the-shelf server hardware, a content owner can realize the fastest and most durable archive experience. Whether one chooses a fully integrated and supported solution or “roll-your-own” open source software and third-party hardware, building object storage-based archives can be straightforward.

That said, it’s going to take a lot to replace well-established and proven tape archives. Tape is still the lowest cost means of long-term preservation, and while many RAID-based archives have been tried in the past, they either have not really caught on or ultimately failed due to cost or lack of efficiency in scaling. But object storage is different. As a technology, it gets cheaper as it scales, and faster too. Content can be retrieved from object storage systems up to 20 times faster than from tape.

Think of object storage as a better RAID than RAID. Object storage not only offers greatly enhanced data durability (in some cases this can be less than 1 bit lost in 1 trillion bits over 20 years) but also significantly lower cost in the petabyte scale, and less complexity of management. The real clincher is that when properly architected, object storage can virtually eliminate those aggravating migration cycles associated with any spinning-disk archive of the past.

Though object storage is technically spinning disk storage, the replacement of its spinning disks can be performed in a safe and considered manner, without “forklift upgrades” and user downtime. Performance-wise, this model is more like nearline storage than vaulted tape, making content retrieval fast and painless.

One large advantage of object storage is versatility. In a well-designed object storage-enabled workflow, object storage can play a bigger part in the day-to-day operations, serving as more than a long-term archive. Without compromising its value as long-term archive, it can greatly enhance overall workflow efficiency, improve the user experience, facilitate media asset management and help the content owner to deliver content faster and more efficiently.

Modern High-Resolution Workflow

To understand object storage better, let’s look at a modern high-resolution HD or 4K workflow. See Figure 1.

While modern workflow is considered a nonlinear, file-based, tapeless process, there are many round-trip operations and possible bottlenecks across this process. The load on the various workflow stages can be highly unbalanced, depending on the amount of work being performed simultaneously across these stages. This lack of balance places greater loads on the underlying storage infrastructure. For instance, if content being ingested resides on the same physical workspace (disks) as content being edited, the two will “fight” for disk time. The resulting physical movement of the disk components leads to increased latency, causing dropped frames and general slow-downs. This situation can lead to bottlenecks and require limits on how users interact with the workflow, a demand that can compromise the overall collaborative abilities of the pipeline. This issue is especially pressing when physical workspaces become full and system headroom is minimized.

At this point, management often becomes a manual process, and it results in the growth of the workspace to an unruly size or, as is now more common, the breaking up of the workspace and addition of expensive solid-state disk (flash) to the mix. This approach can work, but it adds complexity and cost. Though it may help to resolve production slow-downs, this model does nothing to help the movement of content on or off archive media.

This lack of overall efficiency is compounded in high-resolution workflows when content is being moved back in the work-in-progress workspace for re-monetization, increasing the load on the storage infrastructure. As content owners strive to deliver “anywhere, anytime” content, this problem is becoming commonplace.

A Promising Alternative: Object Storage-Enhanced Workflow

Object storage affords the most promising way to resolve these bottlenecks and build a better long-term content archive. In this case, object storage is beneficial for both high-resolution workflow efficiency and long-term archive resiliency.

When the modern high-resolution workflow is modified with the addition of object storage as both long-term archive and a new nearline concept (using object storage as disk to enhance production operations), the workflow shows much greater production efficiency with less complexity, as well as enhanced access and reduced time to retrieve archive content.

While modern HD and higher-resolution pipelines are not simple by any means, the addition of properly designed and implemented object storage can make growth in resolution and larger content archive manageable while helping to eliminate slow-downs and bottlenecks.

In the example illustrated in Figure 2, raw content is moved directly into object storage through the main interface (high-speed Fibre Channel). Content and assets thus can be archived on ingest while  simultaneously being prepared for use in production. Transcoders are directly attached to the object storage via IP connections, allowing the assets that have been moved to the resilient object storage to park for long-term retention, as well as be available for creation of a working mezzanine and proxy formats. This content is generated and delivered to work-in-process disk directly from the object storage over high-speed Fibre Channel, lowering the amount of online storage required and ensuring maintenance of a safe archive of the raw assets right at ingest time. IP-connected clients and even remote users over a REST interface (Internet) with proper permissions can have access to these assets without impacting the overall work-in-process storage performance.

At the same time, the assets on the work-in-progress volume move through the production pipeline with no interruptions or chance for slow-downs or dropped frames. Once they are ready to deliver, the object storage volume is again ready to provide content for transcoding, QA and subsequent delivery. Finished content remains on the object storage in the customer-desired format for long-term archive retention while transient content is purged. In a well-designed system, the majority of the steps in the entire pipeline would be automated through the media asset and archive managers with little to no human intervention other than the planned QA steps required.

The Results

Adding object storage to the workflow can result in improved efficiency and predictability by automating and offloading many time-consuming and tedious operations that could otherwise slow down the overall workflow. At the same time, archive is accomplished with a very high level of resiliency that also provides disk access performance with costs that are far less overall than those of traditional media-class disk systems. By matching the correct storage to a particular workflow stage of the production and archive lifecycle, users can achieve the greatest efficiency and performance for their storage infrastructure spend.