Who supports it, which drives have it, and what has to happen before you see the benefit
In our last article, we argued that Open RAO—Recommended Access Order—is one of the most consequential new capabilities in the LTO ecosystem, with IBM citing read-performance improvements of up to roughly 80–86% on the kind of scattered, many-file restores where tape has historically struggled. We closed with a promise: come back with the implementation details from our manufacturer partners. This is that article.
We left three questions on the table. Is Open RAO really an LTO-10 feature, or is it available on LTO-9 too? What host software actually has to support it? And is it implemented the same way on a half-height drive as on a full-height one—and how does it relate to IBM’s enterprise TS-series drives? Here’s what we’ve learned, with help from Phil Wandrei, Product Marketing at Spectra Logic.
First correction: it’s LTO-9, not just LTO-10.
Our original headline tied Open RAO to LTO-10, and the feature is genuinely a headline capability of that generation. But to be precise: Open RAO entered the LTO specification with LTO-9 and continues in LTO-10. As Spectra Logic’s Phil Wandrei put it plainly, “Open Recommended Access Ordering (Open RAO) is part of the LTO specification beginning with LTO-9 and continuing with LTO-10. As such, it is implemented at the tape-drive level and exposed through standard SCSI command behavior.” So if you’re running LTO-9 today, the capability is already sitting in your drives—the question is whether anything above the drive is asking for it.
The “Open” in Open RAO has a backstory.
Recommended Access Order didn’t start life as an open standard. The concept first appeared in IBM’s enterprise tape drives—the TS-series (3592) family—where it was offered as a vendor capability for customers retrieving many objects from a cartridge. Wandrei points to that lineage directly, noting that today’s feature follows “earlier Recommended Access Ordering implementations in enterprise tape drives.” What changed with LTO-9 is that the LTO consortium brought the same idea into the open, multi-vendor LTO standard. That’s the whole weight of the word “Open”: any LTO-9 or LTO-10 compliant drive, regardless of which of the LTO manufacturers built it, exposes the capability the same standardized way, and any software vendor can implement against it without a proprietary license or a single-vendor dependency.
This is also why the question “Is it on LTO or only on the TS drives?” has a satisfying answer: the idea is shared heritage, but Open RAO specifically is the LTO-standard implementation, present from LTO-9 forward across the LTO ecosystem.
Under the hood: the drive already knows where everything is.
To understand why the drive—and not the host—is the right place for this intelligence, it helps to know something most users never think about. Every LTO cartridge carries a substantial amount of format metadata alongside the data itself. The roughly 20% of tape capacity devoted to C1/C2 error correction gets the headlines, but the format also records where datasets physically live on the tape. As one veteran LTO drive architect noted, that metadata “helps with algorithms such as read access order, RAO… locating.” In other words, the drive maintains a physical map of the medium as a side effect of how LTO already works.
Open RAO turns that map into a routing service. The mechanism is a pair of standard SCSI commands—in IBM’s terminology, Generate Recommended Access Order and Receive Recommended Access Order. The host application hands the drive a list of the objects it wants to read, described by their position on the tape. The drive computes the sequence that minimizes physical tape motion—the back-and-forth shoe-shining that kills restore performance—and hands back a recommended order. The application then reads in that order. The files arrive sorted for the medium’s benefit, not the application’s request order, and the human or system on the other end simply gets data sooner.
The elegance is that all of the spatial reasoning lives in firmware, on the one device that actually knows the tape’s geometry. No middleware, no library database, no external index has to be consulted.
The catch: the application has to ask.
Here is the single most important practical point in this whole article, and it’s the one most likely to disappoint someone who buys LTO-10 expecting magic. Open RAO is not automatic. The drive will happily compute an optimal order, but only if a backup, archive, or data-management application builds the list of objects and issues the SCSI commands to request it. If your software never asks for a Recommended Access Order, the drive never reorders anything, and you get conventional sequential behavior.
Wandrei is blunt about where that leaves the customer. The practical benefit of Open RAO, he says, “depends on whether the backup, archive, or data management application is designed to generate optimized read request lists and issue the appropriate commands to the drive.” And he is specific about where it pays off: “The most significant benefit is typically seen in archive and data-management use cases involving the retrieval of many small or non-contiguous files from a single cartridge, rather than in traditional sequential backup workflows.” So the real adoption gate for Open RAO isn’t the drive or the cartridge—it’s your independent software vendor’s roadmap. When you evaluate LTO-9 or LTO-10, the question to put to your backup/archive vendor is specific: Do you generate Recommended Access Order requests, and in which product version?
The library is (deliberately) transparent.
If you run a tape library, you might assume there’s a setting somewhere to “turn on” Open RAO. There generally isn’t, and that’s by design. Because the capability lives in the drive and is driven over standard SCSI by the application, the library sits outside the loop. “From a Spectra Logic perspective, Open RAO is supported transparently as part of our support for current-generation LTO drives,” Wandrei explains. “Because it is a standards-based LTO feature implemented inside the drive, Spectra does not explicitly enable, disable, or instrument Open RAO at the library level, nor do we have visibility into whether a particular application is invoking it.” The upside for buyers is real: you don’t need to upgrade the library, replace the robotics firmware, or add a management appliance to benefit. If the drives are LTO-9/10 and the application supports RAO, it simply works through the library.
Before the standard, vendors solved this themselves.
Open RAO didn’t arrive in a vacuum, and the history is worth knowing because it tells you how mature the underlying problem is. Before the LTO standard offered access ordering, Spectra Logic built its own—an early sign of the company’s focus on solving archive-retrieval challenges for its customers, years ahead of any industry standard. As Wandrei recounts, Spectra developed a technology it called TAOS—Time-Based Access Order System—to attack exactly this small-file-recall problem “before similar functionality was available in the LTO standard.” TAOS supported LTO-7 and LTO-8 drives in Spectra’s TFinity and T950 libraries and was designed to optimize read-request ordering to reduce seek time and tape movement in archive-style workloads. When Open RAO arrived natively in LTO-9, Wandrei notes, “there was no longer a market need for a Spectra-specific implementation, and TAOS was retired as customers transitioned to newer LTO generations that natively support Open RAO.” That arc—proprietary solution, then standardization, then graceful retirement—is the cleanest possible evidence that access ordering is a genuine workload need and not a spec-sheet novelty.
Half-height vs full-height, and which workloads actually benefit.
On form factor: Open RAO is a drive-level, standards-defined capability exposed over SCSI, so it is specified the same way regardless of whether the drive is half-height or full-height—both are LTO-9/10 compliant drives presenting the same RAO commands. What can differ between form factors and configurations is how much you gain, because the payoff is entirely a function of how much physical tape motion there was to eliminate.
And that’s the nuance the spec-sheet “up to 80%” can obscure. As Wandrei’s comments make clear, the benefit is largest in archive and data-management use cases that retrieve many small or non-contiguous files from a single cartridge—precisely the deduplicated, incremental, many-object restores where conventional tape shoe-shines itself to death. A simple, sequential restore of one large contiguous file has almost nothing to reorder, so it sees little benefit. If you’re sizing the value of Open RAO for your environment, look honestly at your restore profile: lots of small scattered objects means lots of upside; big sequential streams mean less.
What to do with this.
For organizations weighing LTO-9 or LTO-10, Open RAO is best understood as latent value you’ve likely already bought into at the hardware layer—and the work left to unlock it is almost entirely a software-and-workload question. Confirm the generation: LTO-9 and LTO-10 drives have the capability today. Interrogate your ISV: ask whether your backup/archive/HSM software issues Recommended Access Order requests, and in which release—this is the real gate. Don’t wait on the library: no library upgrade or special configuration is required; the feature is transparent through the robotics. And match it to your workload: the gains track restore complexity—many small, scattered files win big; long sequential reads barely notice.
Open RAO remains, to borrow our own earlier phrase, a benefit flying under the radar—but it’s flying because the conversation has been stuck on whether the drive has it. The drive has it. The more useful question, and the one we’d encourage every LTO buyer to start asking loudly, is whether their software is finally ready to use it. Wandrei frames the bigger picture this way: “Open RAO reflects continued innovation by the LTO ecosystem to improve read efficiency for modern archive workloads, and Spectra systems are designed to fully support those standard capabilities as they evolve.”
The LTO Show thanks Phil Wandrei and Spectra Logic for contributing to this article.
Pete Paisley is the host of the LTO Show, the premier podcast for leaders in the LTO tape storage hardware community. Please reach out with story ideas or comments—we’ll respond to each directly. pete@ltoshow.com
Copyright 2026 The LTO Show and Pete Paisley. Linear Tape-Open, LTO, the LTO logo, Ultrium and the Ultrium logo are registered trademarks of Hewlett Packard Enterprise, IBM and Quantum in the US and other countries. All product and company names are trademarks or registered trademarks of their respective holders. Use of them does not imply any affiliation with or endorsement by them.