The Problem with Tape Today
LTO tape has been around since 2000. The hardware is mature, the media is cheap, and the technology is rock-solid. So why does tape still have a reputation as something that only enterprise IT departments bother with?
The honest answer is that the tooling around tape has never caught up with how people actually work. The typical home lab or small business tape workflow looks something like this:
Every step after "put it on a shelf" is painful. And that pain discourages people from reaching for tape at all — even when it's clearly the right tool economically. The question isn't whether tape is cheap enough. It is. The question is whether it's accessible enough to use as a real, day-to-day storage tier. HuskHoard is built to answer that question with a yes.
What LTFS Gets Right — and Wrong
LTFS (Linear Tape File System) was a genuine leap forward when it arrived. For the first time, you could plug in a tape cartridge and have it show up like a USB drive. No proprietary software, no restore jobs — just drag and drop. For studios and broadcast houses that need a standard interchange format, it's still extremely useful.
But LTFS was designed for a specific workflow: you write a reel, hand it to someone, they read the reel. It's a transport format, not a living archive. When you start using it as your primary long-term storage system, the cracks show quickly.
LTFS / Manual Tape
HuskHoard
fanotify — applications request files transparentlyThe core issue with LTFS is that it's volume-centric rather than archive-centric. Each tape is its own island. HuskHoard treats your entire collection of tapes as a single, coherent archive where the physical location of a file (which cartridge, which shelf) is an implementation detail you shouldn't need to think about.
The 2026 Cost Case: Tape vs. Spinning Rust
This is the number that stops most conversations cold. We're in the middle of a real storage crunch. Hard drive prices have been trending up for the first time in a decade, and for anyone building out a 50–200 TB archive, the math has flipped decisively in tape's favor.
| Storage Type | Unit | Unit Price | Cost / GB | 50 TB | 100 TB | 200 TB |
|---|---|---|---|---|---|---|
| HDD (2026) | Seagate IronWolf 20TB | ~$600 | $0.030 | $1,500 | $3,000 | $6,000 |
| NAS SSD | 4TB SATA SSD | ~$300 | $0.075 | $3,750 | $7,500 | $15,000 |
| LTO-9 Tape | 18TB native cartridge | ~$105 | $0.005 | $300 | $580 | $1,160 |
The drive costs above are media only. They don't account for enclosures, power supplies, RAID overhead, or the electricity to run them. That changes the picture further.
100 TB of HDDs in a NAS pull 60–80W around the clock. 100 TB of tape cartridges on a shelf pull nothing. At typical electricity rates, the savings cover the cost of an LTO drive in roughly 18 months.
LTO media carries a 30-year archival rating under normal storage conditions. The mean time between failures for a spinning hard drive makes them genuinely risky past year five.
At $0.005/GB versus $0.030/GB for HDDs, tape is about six times cheaper per gigabyte at current prices — a gap that widens as your archive grows.
The upfront cost of an LTO drive (typically $2,500–$4,000 for a desktop LTO-9 unit) is the main barrier. That's real money. But spread across 100 TB of storage it works out to roughly $0.025–$0.04 per GB of drive amortization on top of your $0.005 media cost — and you only buy the drive once. The cartridges keep getting cheaper as you scale.
LTO-9 cartridges are rated at 18TB native and up to 45TB with 2.5:1 hardware compression. The table above uses native figures. For already-compressed media (video, photos, audio) you'll be closer to native capacity. For uncompressed data like RAW files or source footage, you may see meaningful gains.
The Husk Catalog: Always-On Search, Offline Media
This is the feature that separates HuskHoard from every other tape workflow — including LTFS — and it's worth spending some time on.
The Husk Catalog is a persistent metadata index that lives on your host machine. Every time you write data to tape, HuskHoard records the full file metadata — name, path, size, checksum, timestamps, and the UUID of the volume it lives on — into the catalog. The catalog is entirely separate from the physical media.
What this means in practice: you can eject all 50 of your tape cartridges, put them in a box, walk away, and still search your entire archive instantly.
You get the full path, the file size, when it was written, and which physical cartridge it's on. This is information that LTFS simply cannot give you for ejected media because LTFS has no off-tape index. Its metadata exists only on the tape itself.
The Husk Catalog also enables something even more useful for anyone running scripts or scheduled jobs: you can query the catalog programmatically to audit your archive, generate reports, or check that specific files exist without touching a single drive.
Transparent Access: No Restore Scripts, Ever
Here's where things get genuinely different from anything else in the tape world. With a standard setup — LTFS or otherwise — if you want to read a file that's on an ejected tape, you have a job ahead of you. Mount the tape, find the file, copy it somewhere on disk, wait, then open it. It's a restore operation, not a read operation.
HuskHoard uses Linux's fanotify interface to intercept filesystem access calls at the kernel level. When any application — your video editor, your backup software, a Python script — tries to open a file that's currently on an offline tape, HuskHoard catches that request before it fails, pauses it, and prompts you:
You slide the cartridge in. HuskHoard sees it, loads the file into the read buffer, and the original application continues as if nothing happened. The requesting process has no idea a tape was involved. From its perspective, the file was just slow to open.
This is the "active archive" model. Your tape library isn't a cold backup you restore from — it's a storage tier that participates in your normal workflow. Infrequently accessed files live on tape. Frequently accessed files can be cached on disk. The catalog keeps track of where everything is. The demand-load mechanism makes retrieval seamless.
Enterprise storage systems have offered something similar for decades under the name Hierarchical Storage Management (HSM). HSM is complex, expensive, and requires dedicated software and often dedicated hardware. HuskHoard brings the same core idea to a single Rust binary that runs on a standard Linux machine with a USB or SAS tape drive.
Ransomware Immunity and the Air Gap
The security case for tape is simple and physical: a cartridge that isn't in a drive cannot be accessed by anything on your network. There is no firmware to exploit, no network stack to probe, no SMB share to lock down. The data is literally unreachable.
Ransomware attacks work by finding writable storage and encrypting everything they can reach. An attached HDD — even a "cold" one you only connect for backups — is writable storage the moment you plug it in. A tape sitting in a fireproof box on a shelf is not.
When a tape is ejected, it is completely and literally isolated from your system. No network path, no USB path, no bus connection exists. There is nothing for malware to reach.
WORM (Write Once, Read Many) tape cartridges exist for exactly this use case. Once written, the data physically cannot be overwritten or deleted — even by the drive itself. HuskHoard supports WORM volumes.
Putting a cartridge in your car and driving it to a second location is a complete, legitimate offsite backup strategy. Cloud costs scale with data. A tape cartridge costs $105 whether it holds 1 GB or 18 TB.
HuskHoard's approach to the air gap goes a step further than passive ejection. Because the Husk Catalog lives on your host machine rather than on the tape, your index is always intact even if a cartridge is lost or damaged. You know exactly what was on it, which means you can make an informed decision about what needs to be re-created or restored from an offsite copy.
The 3-2-1 Rule, Revisited
The standard backup guidance is three copies, on two different media types, with one offsite. Tape makes the "different media type" and "offsite" requirements trivially easy to satisfy — and the HuskHoard catalog means you can verify your coverage without physically handling every cartridge.
Who Should Be Looking at This?
The honest answer is anyone managing more than a few terabytes of data they genuinely care about keeping. But here are the three groups where the case is most compelling.
Home Lab and Self-Hosted Enthusiasts
If you're already running a NAS with 20–50 TB of media, you're paying hard drive prices that are increasingly hard to justify for data you access once a month. Moving your cold tier to tape — even a single LTO-9 drive and a handful of cartridges — cuts your long-term storage cost dramatically and gives you a real air-gapped backup for the first time. HuskHoard is AGPL v3, open source, and runs on any Linux machine. There's no licensing cost to get started.
Small Business and IT Administrators
For businesses keeping records for compliance, tape is often already part of the conversation. But the operational burden of traditional tape workflows has pushed many small IT teams toward cloud storage that costs far more at scale. HuskHoard's demand-load model means tape doesn't have to be a separate restore workflow — it can be a live, searchable storage tier integrated into existing tooling. If your applications can talk to a filesystem, they can talk to HuskHoard.
Media and Production Professionals
Video production generates enormous amounts of data with highly variable access patterns. A project you're cutting today needs fast SSD access. A project you finished six months ago might be touched once or twice a year. Tape is the obvious answer for the latter category — the problem has always been retrieval friction. The HuskHoard demand-load model means a colorist or editor can request a file from a completed project, and the archive handles the rest. No IT ticket, no restore job, no waiting for a sys admin.
Getting Started
HuskHoard is a single Rust binary. It requires a Linux host, a compatible LTO drive (USB or SAS), and the mt-st package for tape control. The hardware investment is real — a desktop LTO-9 drive runs $2,500–$4,000 — but the software itself is free and open source.
The GitHub repository has full documentation, including a hardware compatibility list, a guide to setting up demand-load with fanotify, and example configurations for common use cases.
If you're on the fence about the upfront drive cost, the used market for LTO-8 drives is active and the cartridges are interoperable with LTO-9 drives for reading. Starting with used LTO-8 hardware while you evaluate the workflow is a reasonable path — HuskHoard supports both generations.
Tape isn't new. LTO isn't new. But a tape workflow that actually fits the way individuals and small teams work — searchable, transparent, air-gapped, and affordable — is. That's what HuskHoard is trying to be.
If you're building something similar or have questions about the architecture, come find us on GitHub. The project is young and the issues are open.