The Art of Reverse Engineering Yesteryear for Aerospace and Defense Data Storage Drives

February 1, 2025

As their decades-old electromechanical drives start to fail, the continued performance of many computer-based systems is put at risk. Form-fit-function replacements made using solid-state technology can save the day, but making them is not without challenges.

Within the aerospace and defense industry there are many computer-based systems in regular use that were designed and built decades ago. These legacy systems are heavily relied upon and include equipment installed on aircraft and in land vehicles as well as a wealth of ground-based equipment such as training simulators and stations used for configuring/testing line replaceable units (LRUs).

These systems, some of which were designed in the 1970s, include data storage technology once considered state-of-the-art; technology such as early generation SCSI, IDE (PATA) hard disk and tape drives, ESDI, Shugart, and IDC floppy disk drives. With their moving parts, these decadesold drives are failing.

Figure 1. A hard disk drive from the mid-1990s. (Image: Solid State Disks Limited)

Understandably, the drives (figure 1 shows an example) that were designed into these systems are long obsolete. However, recognizing that the electromechanical drives were a potential weakness in the host’s overall reliability, some end users and support companies did have the foresight to order spares while they were still available. When retrieved from their long storage though, these spares fail more often than not. In the case of hard disk drives (HDDs), many models have heads that rest on the surface of the disks when they are not spinning. Over time the heads can stick, preventing the disks from spinning or they are ripped from their suspension mechanisms if the disks do spin.

Finding a replacement drive that works, even a reconditioned unit, is very difficult because in most cases it must be the exact same model; and even then, success is not guaranteed. This is because of how many OEMs (of the drives and the host systems) collaborated to ensure exclusive compatibility. For instance, SCSI was an extremely popular means of connecting computer peripherals in the 1980s and through to the early 2000s. It was standardized in 1986 as the SCSI parallel interface (SPI), 8-bit wide, single-ended bus. However, OEMs did not always implement the full standard, retaining just the SCSI command protocol or the SCSI architectural model, for example.

Accordingly, in the ‘70s, ‘80s, and ‘90s many SCSI drives were designed with a specific host in mind, and the OEMs often agreed that their respective products would behave in a mutually exclusive manner, under particular circumstances. For example, they would often agree on a unique mode page setting, or mode page in the memory map and/or exclusive handshaking protocols when data passes between host and drive.

Such pairing means a donor drive would need to come from an identical host. But even then, the replacement may not work. This is because the OEMs might have also agreed that the host be capable of detecting when a new drive has been installed. The means of detection might include looking for pre-arranged data to mark its authenticity. By the same token, the host may expect to see any replacement drive prepopulated with data from the old drive, and the drive manufacturer would have offered a data duplication service in the day.

All of the above issues make finding a replacement for the original drive virtually impossible. Moreover, modifying the host to accept a different drive is impractical because it would cost a great deal, and the host would probably need to be offline for a long time. More importantly, modification will be prohibited in most cases because the host’s functional behavior will have been certified.

With the host expected to provide several more years of service, the only practical solution is to replace the failing electromechanical drive (ideally before it fails) with one that emulates the exact behavior of the original. This is rather specialist work and only a few companies offer reverse engineering services for obsolete computer drives. For those that do, their starting point is to hunt down the details of the drive’s mechanical and electrical interfaces. Finding some generic information online is usually possible, but finding details of anything the host and drive OEMs might have agreed upon to ensure exclusive pairing is much harder, seeing as one or even both companies may no longer be in business.

In the absence of sufficient information an original drive will need to be interrogated. The best scenario here is that the end user or whoever is responsible for maintenance may have a drive that is faulty (worn mechanical parts) but which still performs its communications handshaking. Worst case, analysis equipment such as a logic analyzer must be taken to the host and placed between it and the drive so that signal timings and communication protocols can be captured. Great care needs to be taken when handling the original drive though as it could be more than 30 years old and will be very fragile.

Once it is understood exactly how the host and drive communicate, that behavior can be replicated using modern technology such as low-power microcontrollers, field programmable gate arrays (FPGAs) and solid-state memory. For instance, most legacy drives are based on logical blocks, where the exact encoding of the data onto the disk is handled internally by the storage device. For some classes of device, such as ESDI or floppy, the encoding must be implemented in the new storage device’s firmware. This is a complex operation that can only be achieved by reverse engineering the particular implementation, including a detailed low-level examination of the format written to the media, which sometimes varies across the surface.

As mentioned, the host might be expecting to see certain data present to indicate a new drive has been fitted, or it might expect to see the same data as before. The first scenario is a simple formatting exercise. Data duplication though requires access to the drive in the host. Again, great care must be taken.

A NATO E-3A Airborne Warning and Control System (AWACS) aircraft sits on the tarmac at Siauliai Air Base in Lithuania. NATO is in the process of replacing the removable media assemblies in its AWACS aircraft fleet that use Seagate HDDs that are more than 20 years old. (Image: NATO)

A solid-state replacement can be created for virtually any SCSI-based drive that was in use in the 1980s to early 2000s, irrespective of how the OEMs of the host and drive might have collaborated to pair their products. By giving the new drive the same physical connector, interface protocol, memory maps, formatting (including pre-loaded data, if applicable) as the original drive, the host does not need to be modified. It will not be aware that it is writing to and reading from latest technology, high-speed, solid-state memory as opposed to a magnetic disk or tape. Indeed, the replacement drive’s software often needs to include delays to slow down data retrieval. For instance, a tape drive might need to rewind the tape. Solid-state technology can provide the data almost instantly, but the host’s software might not be ready for it.

As for the storage media, industrial grade Compact Flash (CF) and CFast cards are used in most cases. If the new drive is to replace a hard disk drive, the CF card can be hidden behind a plate so that it cannot be removed. Alternatively, it can be ejectable if the original drive had used removable media such as a floppy or optical disk or magnetic tape.

Understandably, reliability is greatly improved as there are no moving parts in the replacement drive. It draws less power than the original electromechanical drive and is quieter too. Also, some end users decide to take advantage of the opportunity to add functionality that the original drive never had.

Figure 2. NATO is replacing the electromechanical removable media assemblies of its AWACS aircraft with solid-state-based units that improve reliability and provide additional features. (Image: Solid State Disks Limited)

For example, the NATO Airborne Early Warning & Control Force (NAEW&CF), based in Geilenkirchen, Germany, is currently in the throes of replacing AWACS aircraft removable media assemblies (RMAs) that use Seagate HDDs that are more than 20 years old.

The replacements (see figure 2) replicate the exact behavior of the original HDDs, so the AWACS aircraft require no modifications. However, NAEW&CF’s new drives have the ability to eject the memory cards, whereas before the original HDD-based units had to be removed in their entirety, risking damage to the connectors on the rear. The replacement drives also have an Emergency Erase button on the front panel. This can be used to quickly destroy all data on the memory card if the crew feel there is a risk of the aircraft falling into enemy hands.

In summary, the failure of electromechanical drives has been placing, and continues to place, the reliability of many computer-based systems in use in the aerospace and defense industries at risk. Finding original replacements is virtually impossible and in most cases the host cannot be modified to accept a modern COTS drive. The only recourse is to use a solid-state-based drive that replicates the exact same behavior as the original drive.

This article was written by Brian McSloy, Chief Technology Officer, Solid State Disks Limited (Reading, UK). For more information, visit here .