Where the Work Happens: Rethinking Storage for a Truly Distributed Industry

A little while ago, we caught up with Post Perspective to talk about why storage still becomes a bottleneck once teams start working across multiple locations.

Together with our partners at Resilio, we talked about what we’re seeing every day, studios keeping their existing storage, spreading work across teams and locations, and figuring out how to get data to the work instead of the other way around.

Sharing here:

What’s the biggest storage challenge you anticipate over the next 2-3 years?

Chuck Parker (CEO Sohonet): The core problem is that production has gone truly global, but a lot of storage is still tied to specific locations or clouds. Teams are spread across facilities, homes, and regions; files keep getting bigger; schedules keep getting shorter. Moving that data to where the work is actually happening is still too slow, too manual, and often too expensive.

Over the next few years, hybrid won’t be optional - it will be the default. Facilities will need storage architectures that treat on‑prem, private cloud, and public cloud as one logical pool, with automation deciding what lives where and when. The winners will be the ones who turn data location into an advantage, using intelligent connectivity and data‑movement orchestration – like what we’re building with Resilio – so distributed teams can work as if they’re on the same volume.

Michael Masson (Global BD & Partnership Manager at Resilio): The gap between where data lives and where the work happens is only getting wider. Productions are more distributed than ever, budgets are tighter, and yet file sizes continue to grow. The traditional answer has been “just put everything in the cloud,” but egress costs and latency make that impractical for heavy editorial and VFX work at scale.

What we think we’ll see is a real reckoning around hybrid storage architectures. Not just some on-premises and some cloud, but intelligent systems that can actually move the right data to the right place at the right time without human intervention. The facilities that figure this out will have a real competitive advantage. Those who are still manually shuttling drives or waiting for overnight syncs will struggle to meet the new turnaround expectations clients now have.

‍

Do you see users more and more embracing workflows to the cloud, or bringing work back on-prem? Or both? Why?

Chuck: It’s absolutely both - but in a much more deliberate way than during the pandemic rush to “put everything in the cloud.” What we see now is customers placing each part of the workflow where it makes the most sense: cloud for elasticity, global reach, and spinning up new vendors or teams quickly; on‑prem for latency‑sensitive editorial, colour, and VFX where performance and cost predictability matter.

The real ask is flexibility. Customers want pipelines that let them shift workloads between on‑prem and cloud without re‑architecting every show. That’s why we’re so focused together with Resilio on giving them a consistent connectivity and data‑movement layer, so they can make placement decisions based on creative and business needs, not infrastructure constraints.

Michael: Both, and I don’t think that’s a contradiction. Cloud wins for collaboration, accessibility, and elastic capacity. When you need to spin up artists quickly or have global teams reviewing the same content, cloud storage and cloud native tools are hard to beat. But for the actual heavy lifting for color, conform, VFX rendering, most facilities have found that on-prem is still faster and more cost-effective once you factor in egress and compute costs.

The smart money is on architectures that let you move fluidly between both. Start a project with centralized on-premises storage, push select elements to the cloud for distributed review or vendor handoffs, and pull it all back together for finishing. That’s the workflow we’re built for.

‍

Which departments or stages of the pipeline put the most pressure on storage today?

Michael: VFX, without question. The combination of massive EXR or DPX sequences, multiple shot versions, and the constant back-and-forth between facilities creates storage pressure that compounds throughout a project. A single show can generate PBs of data.

Dailies and onset data management of Original Camera Files are also increasingly stressful points. The expectation now is that editorial can start cutting almost immediately, which means media needs to flow from set to post in near real- time. The old model of waiting for drives to be shipped just doesn’t work for the turnaround expectations on episodic television now.

The other emerging pressure point is archive and retrieval. Studios are realizing that their vaults, full of older projects, represent potential value for rereleases, AI training, ​​and derivative content, but only if they can actually find and access what they need without a multi- week restoration project.

Chuck: VFX is still the obvious heavy hitter. But the real stress we’re seeing comes from the combination of size and speed. File sizes have exploded, and now those same huge assets are expected to be available to multiple teams, in multiple locations, almost immediately. Size on its own isn’t new; it’s the pace and the distribution that have changed.

Modern productions run several workflows in parallel: editorial needs updates within hours, VFX vendors need consistent access to shared assets, marketing and localisation start earlier, and delivery teams want versions sooner. When all of those groups are pulling on the same material at the same time, both the storage and the network feel it.

That’s exactly where a high‑performance network and sync layer earn their keep. Using Sohonet’s media network and collaboration services together with Resilio’s platform, we’re trying to take the friction out of keeping those departments in step, instead of forcing them to fight over where the “real” master lives.

‍

How do you see storage architectures evolving as AI and real-time engines become more central to the pipeline?

Chuck: AI and real‑time engines amplify everything we’ve been talking about. They demand low‑latency access to very large datasets, often from multiple locations at once, and they generate a lot of derivative media and metadata along the way. The bottleneck stops being “do we have enough terabytes?” and becomes “can we get the right data to the right compute, fast, and keep it in sync as things change?”

Architecturally, that pushes you towards more distributed, hybrid designs – with on‑prem, cloud, and edge each doing what they’re best at – tied together by intelligent connectivity and data‑movement. That’s where we see Sohonet and Resilio playing: turning storage from a passive utility in the background into an active part of the creative fabric that can keep up with AI‑driven and real‑time workflows.

Michael: The biggest shift is going to be around data locality and intelligent caching. AI and real time engines are hungry, they need fast access to large datasets, often simultaneously. That doesn’t play well with traditional centralized storage models where everything funnels through one bottleneck.

We expect to see more distributed architectures where hot data lives close to where it’s being processed, with intelligent systems managing what gets cached where. This is already happening with virtual production stages that need instant access to massive texture libraries while also syncing changes back to central storage for the broader production team.

The other piece is metadata. AI generates enormous amounts of derived metadata. Like transcriptions, object detection, scene analysis. And that needs to live somewhere accessible across the pipeline. Storage systems that treat metadata as a first class citizen rather than an afterthought are going to have an advantage.

‍

Why did Sohonet and Resilio decide to partner, and what problem does that solve for studios?

Chuck: What brought us together is simple: customers need their storage and workflows to function seamlessly across multiple locations, and they don’t want to rebuild their infrastructure to get there. Resilio gives us a powerful way to extend the storage they already rely on and make it behave globally.

By combining Sohonet’s managed media network services with Resilio’s sync technology, we provide teams a simple, real-time way to keep files and artists aligned across sites without adding complexity or cost. Together, we turn existing NAS and SAN into a distributed, hybrid‑ready fabric that can span on‑prem and cloud.

Michael: Resilio is storage-agnostic by design, which means facilities can keep using their current storage investment. What we solve is high-performance movement and synchronization of data across those systems when teams are distributed

The sweet spot is any workflow where you have multiple locations that need to work from the same source files without the latency of traditional cloud sync or the complexity of managing data pipelines manually. VFX studios, for example, with their multiple studios, satellite offices, or remote artists who need to pull shots, work locally at full speed, and push changes back without stepping on each other. However, we’re also seeing heavy use in animation, episodic post-production, and audio post-production where onset data needs to flow to multiple downstream facilities simultaneously.

The partnership with Sohonet is particularly relevant in this context. Sohonet handles the network connectivity and secure delivery infrastructure that studios and facilities trust while also offering a managed service element and month to month pricing. Resilio handles the intelligent sync and replication layer. Together, we’re essentially giving facilities an enterprise grade hybrid cloud architecture without having to rearchitect everything they already have.

‍

Does it include a MAM?

Michael: No, and that’s intentional. We integrate with MAMs rather than compete with them. Facilities have already invested in tools like ShotGrid, ftrack, Iconik, or whatever works for their pipeline. Resilio handles the data orchestration underneath, making sure the files those MAMs are tracking are actually where they need to be, when they need to be there.

‍

If distributed teams and hybrid storage are part of your world, this is worth a look: https://info.sohonet.com/resilioxsohonet