The Future is Looking a Bit Cloudy, or Is it?
The Role of Cloud Storage in Post-Production
Perform a query with your favorite search engine using “cloud storage and post-production” and you’ll see increased chatter about the potential role of cloud-based storage for post-production. Some of the bigger cloud-storage providers are starting to promote the benefits of their cloud offers in relation to Post workflows, and quite a few articles and blog entries are floating various use cases and success stories, too. Not many of them are getting down to any technical level of detail, but the vast majority are citing the usual benefits, in true sales-speak, of cloud storage: easy of accessibility, ability to scale up or down as needed, and of course cost savings.
What’s the truth behind some of these claims that cloud is the next big frontier for post-production outfits to explore when considering their storage infrastructure needs? As with most complex technical questions, the safe answer is “it depends.” So what I’d like to do is talk about the applicability of cloud storage to post-production—and to be clear, I do see applicability and promise—while also framing the question in terms of what aspects of production, in general, most directly benefit from cloud, if at all. Of course, I will also focus on what areas don’t and won’t benefit, and as a matter of fact, will hinder Post activities and timelines.
Let me start by pointing out that cloud storage, whether public, private, or hybrid, presents many conveniences and cost reductions, but only when used to support the right activities and applications. I don’t think that anybody would throw shade on the entire concept of cloud, especially when you consider its core benefit of “access from anywhere.” It comes down to what specific problems cloud uniquely addresses versus other problems with which cloud-based storage doesn’t fare so well. Again, it just depends on the use case. Overall, though, cloud does serve a purpose and does that intended purpose quite well.
This raises the question: what am I really comparing cloud storage to? To simplify matters, let’s just boil it down to a dichotomy between public cloud storage and on-premise enterprise-grade storage. We could complicate matters hopelessly by getting into the multitude of variations of both, so let’s just keep these two binary choices as the focus of discussion.
Another concept I’d like to tee up before I really get down to an assessment of cloud-based storage for post-production is to list out the various aspects of production generally, just for clarity’s sake, and try to map cloud to the most appropriate phase of production activity. The relevant stages of production include:
- Development—ideation and financing.
- Pre-Production—logistical planning and casting.
- Production—the capture of raw footage either on-location or in a studio.
- Post-Production—editing and refinement to produce a master.
- Post-Services, or Distribution—transcoding, marketing, and distribution of finalized production.
I really want to focus on the last two, because of the primary question at hand—is cloud storage appropriate for post-production, and the follow-up question, for which activity is cloud storage truly most appropriate?
Given the performance and latency requirements associated with editing video and audio—especially when you get into 4K editorial and beyond—I can safely say that cloud storage is not the best option (or any option at all) for post-production. This comes down to a simple question of latency. For efficient and frustration-free post-production, you need to have incredibly low latency for jitter- and lag-free playback and scrubbing. You simply can’t have a team of highly paid creatives all struggling to perform their various functions when the project files are slow to load and they can’t navigate around without obstacles.
Can cloud-based storage support this high-performance demand? Probably not, but let’s take the focus off the storage solution for the moment and look at another critical piece of the puzzle: the network infrastructure between client workstations and the storage itself. In the US, the reality of the situation is that the vast majority of our public Internet infrastructure simply isn’t designed to sustain the types of symmetrical, high-speed point-to-point traffic necessary to support super-low latency requirements. So to summarize, I’m really not condemning cloud storage in and of itself (though performance and throughput of public cloud storage is definitely a concern), but rather I’m making an assessment of the entire network-storage infrastructure as a whole. It just won’t work to anybody’s satisfaction.
What post-production really depends on is a storage solution that is on-premise, enterprise-grade, with very high speed and performance capabilities, along with data integrity checking and a host of other must-have capabilities. All of this must be married up with an equally high-performance network (typically a local Ethernet-based one) that fully supports the low-latency, high-speed data transfer requirements to allow unencumbered, delay-free post-production. The final call on the primary question? Cloud storage in its present state is not a viable option for the vast majority of post-production outfits.
Which brings us to that ancillary question, where does cloud storage best fit into the entire production lifecycle, if not at the post-production stage of activity? The answer might be obvious at this point. Production post-service. Cloud is definitely a winner when an organization needs to scale up Compute capabilities on-the-fly, for tasks such as computational animation (increased render power), transcoding, and of course long-term storage needs.
Transcoding is a particularly interesting application for cloud storage, because the fact of the matter is that potentially hundreds of versions of a master asset need to be created, in an automated manner, prior to final distribution of all these deliverables. Keep in mind that this has nothing to do with the visual editing and the intense performance requirements which Post requires. What we are talking about here is leveraging the cloud for the conversion of the master asset in the cloud-based transcoding farm into all those necessary deliverables (variants with different languages, audio dubs, bandwidth requirements, etc.). In this instance, cloud is a particularly good choice, freeing up on-premise storage resources for low latency real-time editing tasks.
When it comes down to it, you can say that on-premise storage and cloud storage each solves a different problem (or set of problems). For post-production, you need that ultra-high-performance on-premise storage and network infrastructure to enable very low latency data transfers. For “offline” transcoding, computing, or other long-term storage needs, sure, cloud storage is something that any organization should probably explore. Just make sure that you’re using each tool for its intended purpose to maximize potential and minimize costly inefficiencies and frustrations. Using a screw driver to drive nails into wood? Doesn’t make sense.
Authored by Sean Lee, OpenDrives Product Evangelist