The Real Cost of Flat Storage in Post-Production
I spent years as a post supervisor before moving to the vendor side — including running a post-production facility before the strikes. I worked across features, episodic, and commercial production, and one thing stayed consistent, at facilities of every size: the same storage strategy. Buy more, plug it in, deal with the consequences later.
It works for a while. Until you’re running a DI on a locked picture while your colorist in London is pulling selects from the same NAS, two VFX vendors are uploading renders, and someone from production wants to know why review files aren’t accessible. At that point, it’s not a storage capacity problem. It’s an architecture problem.
Post-production workflows are uniquely brutal on storage infrastructure. A single project can generate terabytes of raw camera originals before proxies have even started. Multiple projects at different lifecycle stages compete for the same infrastructure simultaneously. And the people who need that data are increasingly distributed across time zones and facilities.
Tiered storage is the structural answer to that problem — not a new one, but an increasingly necessary one as file sizes grow, teams get more distributed, and the gap between high-performance storage costs and archive-tier costs continues to widen.
What Tiered Storage Actually Means in a Post Environment
Storage tiering is the practice of organizing data by how frequently it’s accessed and routing it to the appropriate storage tier. In theory, it’s simple. In a post-production environment, the execution requires mapping those tiers to how a project actually moves through the pipeline.
Most frameworks break this into three tiers:
- Cold tier: Low-cost, high-capacity storage for completed projects, raw originals, and anything unlikely to be touched for months. LTO tape, object storage, and cloud archival tiers like AWS Glacier.
- Hot tier: High-performance, low-latency storage for actively used data. Flash, NVMe, SSD-backed NAS. This is where your editors live.
- Warm tier: Mid-tier storage for data that’s still in circulation but no longer needs peak performance. Sequences in review, assets being handed off, and deliverables awaiting final approval.
The reason this maps well to post is that projects have a natural lifecycle with distinct phases. Data starts hot, cools as the project progresses, and eventually belongs in cold storage once it’s locked. The challenge is that those phases rarely line up cleanly. A feature in finishing still has VFX shots coming in. A locked episode might get reopened for a network revision. Your archival workflow has to account for real production, not an idealized version.
Following the Data Through the Pipeline
Stage 1: Ingest and Active Editorial
This is when storage demands peak. Camera originals are landing, transcodes are running, and editors are hammering the storage with constant reads and writes. Any latency here is immediately visible in the edit suite. Dropped frames, sluggish scrubbing, stalled exports — all of it traces back to hot tier performance.
For high-resolution camera formats, sequential read speed isn’t a nice-to-have; it’s a hard requirement. This stage belongs on fast, local, or on-premises storage with the throughput to support real-time playback across multiple streams simultaneously.
Stage 2: Collaboration, Review, and Finishing
Once a cut is in review, the access pattern changes. Fewer people are writing to the project. More people need read access from more locations. The performance bar relaxes somewhat, but the reliability and accessibility bar goes up.
This is where distributed teams feel infrastructure decisions most acutely. If a client reviewer in New York can’t pull a frame-accurate reference file, or a colorist in London is waiting on a conform that should have been available hours ago, the warm tier is usually where the problem lives. Hybrid architectures, on-premises storage syncing to a cloud environment, often work well here, provided the sync is reliable and the access pattern is predictable.
Stage 3: Delivery and Archival
The project is locked. Deliverables have shipped. The raw originals, project files, and completed renders still need to be retained contractually, or because projects get reopened more often than anyone expects, but they don’t need to live on expensive primary storage.
This is where tiering pays for itself most clearly. Moving locked projects to cold storage frees capacity on your hot tier for active work, reduces costs, and creates a cleaner operational environment. The key is having a defined policy that governs when data moves, rather than making that call manually, project by project.
What Each Tier Needs to Handle
Hot Tier: Performance Is Non-Negotiable
The hot tier has one job: support active editorial without introducing latency. For color grading, VFX work, and audio finishing, the I/O load is significant. The threshold for what belongs here is simple: if an editor, colorist, or finishing artist is actively working with it, it stays on hot storage.
What belongs in hot storage:
- Raw camera originals during active editorial
- Proxy files and transcodes
- Active project files and auto-saves
- Shared bins and collaboration assets in active use
For most studios, the hot tier is on-premises: a high-performance NAS, shared SAN, or local server with flash or NVMe storage. The economics of cloud at this tier get expensive fast when you’re constantly moving large media files. Caching layers can help remote editors, but the core hot tier typically stays local.
Warm Tier: Reliable Shared Access Across Locations
The warm tier is where distributed collaboration actually happens. The primary requirement isn’t peak performance; it’s reliable, consistent access for multiple users across multiple locations, often simultaneously.
What belongs in warm storage:
- Sequences and cuts under client or internal review
- Deliverables awaiting final approval
- Assets being handed off between editorial, color, audio, or VFX
- Reference files and approved assets
Hybrid architectures work well here: on-premises storage that replicates to a cloud environment to support remote access, or shared cloud storage that eliminates the need for VPN or manual file transfers. The warm tier is also where having a storage-agnostic sync layer matters most. If your architecture ties you to a single cloud vendor or proprietary hardware, you lose flexibility when team structures or project requirements change.
Cold Tier: Long-Term Retention at Minimal Cost
The cold tier is where completed projects are deployed at the lowest possible cost. This tier directly controls your long-term storage spend, which is significant for most studios.
What belongs in cold storage:
- Locked project files and finished deliverables
- Raw camera originals after editorial wrap
- Completed VFX renders and composites
- Archival data required for compliance or contractual retention
Cold storage options include LTO tape (still widely used for deep archival, particularly at facilities managing multi-petabyte libraries), on-premises object storage, and cloud archival tiers such as AWS Glacier, Azure Blob Storage, or Google Cloud Coldline.
Two factors that get underestimated here: egress costs and retrieval latency. Cloud archival tiers can have retrieval windows ranging from minutes to hours, and egress fees for large media datasets add up fast if a project gets reopened for revisions. Model your retrieval scenarios before committing to a cold storage provider. If you think there’s any meaningful chance of needing data back within a few months, it probably belongs on warm storage, not cold.
What to Look for in a Post-Production Storage Architecture
Not all storage architectures are designed to support tiering across a distributed team. When evaluating your options, a few capabilities matter more than anything else:
Multi-site and remote team support. If your team spans multiple locations, and at this point, most do, your storage architecture needs to support reliable file access across all of them without requiring people to babysit file transfers or troubleshoot VPN access. This means some combination of on-premises and cloud storage with a reliable sync layer.
Automated data movement between tiers. Manual tiering doesn’t work at scale. You need policy-based automation that moves data based on defined rules, time since last access, file type, project status, and folder location, without requiring IT to intervene on each project.
Granular file filtering. The ability to define exactly which files get tiered, and when, matters. Post-production lifecycles don’t map cleanly to simple rules. A project folder might contain active VFX shots alongside locked editorial assets. Granular filtering by modification time, access time, file name patterns, or explicit file lists gives you control that generic tiering tools don’t.
No vendor lock-in. The tiering layer should integrate with your existing NAS, object, and archive targets, rather than forcing proprietary dependencies. That flexibility allows you to switch storage classes, regions, or providers as economics change, without reworking your workflows.
Mistakes I’ve Seen Teams Make Repeatedly
Even teams that understand tiered storage in theory tend to run into the same problems in practice.
Keeping everything on hot storage too long. This is the most common and most expensive mistake. Completed projects and locked deliverables sitting on high-performance NAS consume capacity and cost that should be reserved for active work. Without a defined archival policy, hot storage fills up faster than anyone expects, and the default response is to buy more — which just delays the same problem.
No defined tiering policy. Tiering only works if there are clear rules governing when data moves. Without them, data management becomes reactive, and IT ends up making judgment calls manually on a project-by-project basis. That’s not scalable.
Siloed storage that breaks collaboration. When different departments or locations manage storage independently, you get file duplication, access problems, and version confusion. I’ve seen this break workflows in both directions — VFX vendors working off stale exports because editorial had already moved to a new version on a different share, or finishing artists missing reference files because archive jobs had run prematurely. A unified architecture with consistent tiering policies reduces that friction significantly.
Underestimating cloud egress costs. Cloud storage looks economical until you start retrieving large media datasets with any regularity. Projects get reopened. Directors want selections from three years ago. Standards evolve, and deliverables need to be regenerated. If you haven’t modeled your retrieval scenarios, the egress costs can easily erode the savings you expected from cloud archival.
Treating tiering as a one-time project. It isn’t. As data volumes grow and team structures evolve, tiering policies need to be revisited. What made sense at 50TB doesn’t necessarily hold at 500TB, and a policy built around a centralized facility team needs to be rethought when you’re running distributed workflows across three continents.
How Resilio Active Everywhere Handles This in Practice
Resilio Active Everywhere is the data movement and automation layer built for workflows like this. It’s storage-agnostic — built on a distributed peer-to-peer architecture — so it works with whatever combination of on-premises systems and cloud targets you already have: AWS Glacier, Azure Blob Storage, Google Cloud Storage, any S3-compatible target. No proprietary hardware required, no vendor lock-in.
A few things that matter specifically for post-production:
Archival that runs without disrupting active work. Tiering jobs run on custom schedules — weekly pushes to Glacier, monthly compliance retention runs — without touching real-time sync or editor access. Your team stays focused on the project. The storage takes care of itself.
Granular file filtering that maps to production lifecycles. Policies can be defined by modification time, last access time, regex patterns, or explicit file lists. If a project hasn’t been touched in 90 days, move it. If a specific folder contains locked deliverables, archive it on a defined schedule. The filtering is specific enough to handle the messiness of real production workflows.
Primary storage that stays clean automatically. After archiving to lower-cost cloud storage, Active Everywhere removes the data from primary storage without requiring manual intervention. For studios managing expensive NAS or SAN capacity, that’s reclaimed headroom — immediately available for the next active project.
Distributed architecture for multi-site teams. Because Resilio is built on a distributed P2P architecture rather than a centralized hub, it reduces dependence on a single central choke point when your London colorist and your LA editor are both pulling from the same project. Data moves efficiently regardless of where team members are located.
The Bottom Line
Post production storage doesn’t have to be a perpetual cycle of buying more capacity and hoping it holds through the next project. A well-designed tiered architecture matches the right storage to the right data at the right point in the production lifecycle: fast local storage for active editorial, reliable shared access for review and finishing, and cost-efficient archival for completed work.
The teams that get this right don’t just reduce storage costs. They create a more predictable, more scalable infrastructure foundation — one that handles the complexity of real production workflows rather than an idealized version of them. That means less time firefighting storage problems and more time focused on the work itself.
If you’re evaluating how to bring automated tiering into your post-production infrastructure, Resilio Active Everywhere is worth a closer look. Schedule a demo to see how it fits into your existing storage environment.
