Moving Media into the Cloud

As discussed in the December 2020 issue of Newswatch, cloud-based media production has made rapid advances recently due to a range of industry initiatives and technological breakthroughs. One of the important drivers in moving media production into the cloud on a grand scale, however, has been the ongoing thrust by major technology companies to hone their proprietary business cloud computing platforms to make them compatible with the entertainment space specifically. For example, Hanno Basse, chief technology officer of media and entertainment for Microsoft’s Azure cloud computing service, says Microsoft is working strategically to understand the needs of the entertainment industry in order to build tools for media production, and as part of that agenda, to join wider industry initiatives to design standards and protocols for cloud-based production of media content. Basse, a SMPTE Fellow and member of the recently launched SMPTE Media in the Cloud Advisory Group who came to Microsoft after seven years as CTO for 20^th Century Fox, is a firm believer that the path to wider and more successful adoption of cloud-based media workflows is by helping the industry achieve specific standards for cloud-based production platforms.

“Microsoft interviewed close to 50 industry insiders, directors, producers, editors, cinematographers, visual effects supervisors, and others about their thoughts on production in the cloud,” Basse says. “We gained a lot of great insights into what content creators are most interested in, such as collaboration of globally distributed teams, and concerned about, like security of pre-release content.

“The company has refocused on media and entertainment solutions within the last few years, and has been in the process of acquiring direct industry expertise ever since—that’s why I joined Microsoft to begin with. We are especially focused on content production, and now have multiple partnerships with Hollywood studios and [major media technology providers], and now we are developing an open industry media cloud platform built on [the foundation of] our Azure technology.” 

Basse was among the industry experts who participated in crafting the recent MovieLabs white paper outlining a potential roadmap for converting media production to an industry-standard, software-based methodology by the year 2030. As discussed in the December Newswatch article, that initiative portends a radical shift in how production could be done. Microsoft, along with others, has thus been busy pushing to bring the entertainment industry into this new cloud-based universe with a series of maneuvers.

In particular, according to Basse, Microsoft has entered into partnerships with Disney’s StudioLAB and with Universal Pictures/Dreamworks to migrate studio workflows into the cloud—initiatives Basse says will form “the foundation for the open cloud-based production Azure would like to build for the industry.” Microsoft also has been involved in partnerships in recent years with Avid and Adobe to allow editors to run professional editing tools through the Azure platform for remote work. Likewise, Microsoft has a partnership with Technicolor to utilize set-to-cloud workflow techniques to streamline and automate the ability to enable cloud-based solutions from the camera head onward. 

Basse says “the goal for the platform is to create a truly open ecosystem where artists can use tools they are familiar with, whether for modeling, animation, compositing, lighting, editing, sound mixing or whatever else, and have them all run seamlessly on the cloud while being integrated into a single seamless overall workflow. So artists can transition into working in the cloud, but still have a total familiarity with the tools and process they are using. At the same time, working this way would make it easier to perform new functions, because [companies or developers] could create new apps and simply just plug them into the same platform.”

From a technical perspective, this open cloud-based production platform would be built on top of the storage, computing, and networking capabilities of Microsoft’s existing Azure platform, which is already used by a wide range of industries for all sorts of cloud computing functions. But, more importantly, he says the goal is to make such a platform adhere to wider industry-defined asset and workflow standards being proposed by MovieLabs through its aforementioned 2030 Vision initiative, as previously discussed in Newswatch. And that, he emphasizes, is the larger point—companies need to, at a minimum, adhere to the key principles outlined in the MovieLabs paper in order to show the industry and vendors “what the core parameters and requirements for production in the cloud should look like,” Basse explains.

In order to get what Basse calls the current world of “fragmented workflows” to organically come together so that most processes can become automated using cloud computing techniques, this means the first principle outlined in the MovieLabs approach is crucial—the creation of an industry standard ontology for this kind of work. 

“Basically, this means that in order for assets to be stored and processed in the cloud, they must be described in a uniform, standardized manner,” Basse explains. “So we are supporting the industry as they create the asset ontology. This ontology would describe what assets are stored in the cloud for a particular project, the metadata associated with those assets, and how those assets relate to other assets in the project or elsewhere that you might bring into the project. The end goal for the industry is to use one name and one description for the same things. Right now, all vendors and studios use their own nomenclature and their own naming conventions, their own workflows and pipelines for their production processes. As the industry migrates to the cloud, we think it necessary to agree on common nomenclatures, or a common ontology, as we call it.”

Secondly, and just as important, the industry will need a standard API (Application Programming Interface) so that all existing and new production-related applications can seamlessly run on all cloud platforms across the industry.

“The API is what allows your application to interface with the open cloud-based production platform,” Basse says. “It’s what lets creative applications access the assets, process them, and utilize cloud resources to work on them, various types of computing, storage, networking, and what have you—the API provides all that capability. So anyone building a platform like this will need to provide an API so that third-party applications can consistently and easily interface with it. It must be ideally standardized, but at a minimum, it should be published and the community allowed to contribute so that anybody who wants to participate in this ecosystem will know how to plug into a cloud and run properly.”

While these principles are widely proliferating across the industry thanks to the MovieLabs initiative and participation by large studios and major technology companies, this approach does beg the question—why is it advantageous for studios, production companies, and other content makers to entrust their precious assets to a major cloud computing provider like a Microsoft, Google, Amazon, or others? Why trust the cloud at all, for that matter, given concerns and hacking incidents that have occurred across the wider landscape in the past?

There are initiatives and information across the industry that address such security concerns, including the Trusted Partner Network (TPN) initiative, and MovieLabs’ recent security paper. But more generally, Basse suggests the conceptual answers to these questions lie in the fact that major cloud providers can offer wider scale, bandwidth, and security for sensitive major projects, while at the same time, “some of the security concerns our industry has relative to cloud workflows are not applicable anymore. Cloud technology has been proven to be significantly more secure than on-premises environments.” In other words, he suggests, major cloud providers are best suited to making sure cloud systems are redundant and resistant to human failure. “The only major exposure area lies in misconfigurations and mishandling of credentials, but that’s a problem both on-premises and in the cloud,” he adds.

“But the biggest argument [for security] is scale [from major cloud providers],” he continues. “If you compare the scale of the cybersecurity efforts that a Microsoft undertakes every day—compare that with what any studio or post production company can do. Its just orders of magnitudes larger. And the proof is that we can basically apply this scale across many different industry applications. It is not an issue unique to just media and entertainment applications. We have health-care customers, government customers, banking customers all dealing with tons of very sensitive personally identifiable information each day. So we can basically apply cyber-security efforts across all those verticals at the same time, which just makes it a lot more powerful than what a post house, let’s say, can possibly do. Any hyper-scale cloud provider could do that at a much higher scale, and therefore make your assets more secure than any individual [entertainment] company could do.”

Related to this issue, Basse adds, is the issue of data integrity generally.

“The cloud has so much redundancy built in—that is why we almost never lose data,” he says. “If there is an issue in any given data center, the data can be replicated somewhere else. That is totally transparent to our customer if we do have to do any type of redundant data recovery. A hyper-scale cloud just doesn’t lose data. So I think people are now coming to realize that data integrity and security are actually much better taken care of in a cloud environment than they would be otherwise. But people need to make sure they don’t confuse the cloud with the so-called ‘open Internet.’ They are two very different things.”

Basse emphasizes that this initiative is aimed at “bridging the gap” between the IT world and the needs of the entertainment industry. The hope, he says, is for the technology industry to more fundamentally “understand the creative process and what it takes to enable creativity” while the entertainment industry learns “how [IT-based] technology can be a tool that helps, rather than a hindrance—that is super important.”

“We are telling partners [in the entertainment industry] that we want to free them from the low-level commodity technology stuff that they normally deal with,” he says. “In a cloud production environment, we tell them we want them to only think about their creative process, the stories they want to tell, the characters and the worlds they are creating. This platform we are creating is designed to do that—to free them from wondering where their data is, do they have all their files, do they have the right files, which versions should they be using, how far along are they in their workflow, and so on. That is the goal we are working toward.”

How long this might take to achieve, of course, is a harder question. Basse says there are infrastructure-related questions on the artistic side that will need to be answered, as the rush to make remote workflows more possible in the last year due to the Covid-19 pandemic has already illustrated. Will artists have the broadband pipe they need at home to take full advantage of such sophisticated cloud computing platforms? Will they have properly calibrated monitors as they would inside a post-production facility? How long will it take to set up remote infrastructures for artists so that they can fully take advantage of cloud computing platforms for any production task, such as rigorous mastering processes, for example?

“The other issue is that the content sits in the cloud, gets rendered in the cloud, and then gets played back to a device that is miles away,” he relates. “For that, we need high-resolution, wide color gamut, high bit-depth transport mechanisms. That is something we are still working on. It exists right now, but is not ubiquitous or widely available yet. From a fundamental standpoint, though, we know what to do. This issue is more of an implementation problem at this stage.”

Therefore, neither Basse nor anyone else can give a detailed timeline of when there will be industry-wide, interoperable, cloud-based production platforms widely available. But that is not a deep concern at this time, he adds, because “we already have major bits and pieces of it existing today” and the evolution of this industry transition is proceeding unusually rapidly and relatively smoothly.

“It’s a complex ecosystem, obviously,” he says. “It will take a little bit of time to really get all the tools and all the different functionality that you need in your pipeline all migrated to the cloud. But at the same time, the MovieLabs initiative calls for it to happen by the year 2030. Many of us are already joking we should have called it ‘2023.’ That’s because, for one thing, the pandemic accelerated the shift to cloud production—it exposed and accelerated all kinds of societal and technology processes that were already in development. So it’s hard to pin down a timeline, but I think it will be here quicker than we think.”