As for full disk encryption, you can select where to store the key, which may be on the NAS itself (rendering FDE more or less useless) or on a USB key or similar.
As a KMIP server you use:
- Another Synology NAS with DSM >= 7.2
- A KMIP compatible key server
Except for the demo implementation that Synology uses (PyKMIP), all the KMIP compatible servers I've found have licenses in the tens of thousands a year. So if anybody has any suggestions to substitute PyKMIP...--
0: https://kb.synology.com/en-global/DSM/tutorial/Which_models_support_encrypted_volumesThe DSM itself lives in an unencrypted partition or volume. Applications with data in encrypted volumes will be inaccessible until the volumes are unlocked.
As usual, there is an easy workaround. You can run a KMIP server in a docker container and set up an external keystore. Once synology allows you to proceed with volume encryption, you can discard the KMIP server if you want and use the recovery keys.
Edit: what they deploy on their NAS is an old version of a testing implementation of the KMIP protocol. PyKMIP: https://github.com/OpenKMIP/PyKMIP
I went down the rabbit hole and implemented the KMIP client and server, that pass the tests from OASIS.
Sidenote: please, somebody nuke the OASIS from orbit. To be sure.
Not to defend Synology, but popping a drive out of the NAS so that it won't be noticed (or noticed much later) is a much easier way to steal data than carrying off the whole NAS. I assume they're guarding against the kind of scenario where an employee steals steals drives rather than ski-masked thieves breaching the office and making off with the NAS.
The primary value of disk/volume encryption is actually for scenarios like end-of-life replacement, RMA, failure and disposal - even if someone later reconstructs the disk sectors, the bits remain unreadable. This is one layer of defense in depth, not a substitute for physical security.
Synology also supports KMIP, which I see addressing two situations:
1. Data center key governance and media mobility - Multiple hosts (including spares) can use KMIP for centralized key management, improving the mobility of drives within the data center and reducing the operational cost of moving drives between machines. When decommissioning hardware, keys can be revoked directly in KMIP with an audit trail.
2. Edge/branch sites with weaker physical controls - By using KMIP, keys are kept in the more secure data center rather than on the edge device itself. The edge hardware stores no keys, so if an entire machine is stolen, it cannot be unlocked, preserving confidentiality.
My DS918+ has multichannel SMB and possibly also parallel NFS. It only works if you have multiple NICs connected.
Other than that, i completely agree. Their tech stack is horribly outdated, and while i understand their reasoning for not upgrading, there's a limit to how long you can do that. Their reasoning is that they know the software that's currently running, warts and all, and can better guarantee stability across millions of devices with fewer moving parts.
I got an issue where mind would randomly start writing disk like crazy and maxing cpu usage, to the point I was bothered by the noise. I’d stop all containers, leave it as close to idle as I could manage, still spiking.
There was no way I could learn what was causing it.
I would like to assume it was a disk maintenance process or something, but for all I know it could be mining bitcoin and I’d be none the wiser. It went on for some weeks then stopped.
May or may not be what you encountered, but had a customer caught by this and found out the hard way you can't stop it. My issue is not the processing, it's the throttling, it's so crazy how the entire NAS gets taken down for like ten minutes (and that was on a racked xeon model), no samba no nfs no nothing answering anymore.
And yes, the lack of trotting is an issue, since you can’t even reach an administration panel. When it’s bad even ssh struggles.
Mine is in the basement for this reason. When it’s still and quiet after midnight I can still hear it grinding away. God I hate the sound.
FWIW the new Ugreen NAS run Debian. I don't know a ton about it, but it's be great if they could stay a little more up to date. This Synology story with ancient forks & weird encryption sounds truly bogus.
I will say that the Ugreen NAS seems to offer more performance for less watts, so it's definitely something I will keep an eye on in the future if it pops up on Ebay.
> This Synology story with ancient forks & weird encryption sounds truly bogus.
It's not. My Synology is running Linux kernel v4, and I opted to use their "SHR" RAID configuration and can confirm that it's some weird BTRFS variant that is likely deadlocked due to the kernel.
The encrypted volumes I've made also look very much like the EcryptFS files I've been seeing on other setups.
I'm currently in the process of mainlining it to kernel v6 to reap the better power and idle / hibernation rewards, as well as just using a standard Ext4 FS with updates
SHR is mostly MD-RAID and LVM, and works with ext4 too.
It's confusing me after the preceding displeasure wrt Synology
They already had one Synology device, they don't have any IT employees on site, and I'd need to take a flight to go to their offices, so I thought that using another Synology device would be better for maintenance. They (and I) were also worried about the noise: it's an small office, and they needed at least 8*3.5" drives, and most of the decent solutions I found for 8 or more drives were big and noisy. The Jonsbo N5 appeared a bit later, that looks like a good candidate today.
Now I found that all their applications are half done, they don't upgrade or fix them regularly, security-wise is a mess, and everything on the backend is super old...
Over time their advantage has eroded as upstream has caught up, to the point that it looks ridiculously out of date today.
But don't you love it when companies invent their own security instead of using battle-tested open-source systems?
I've a Ryzen Embedded system with lost of RAM as my NAS box and a small Intel N-series based system as my Plex server that pulls media off the NAS box.
(You might want to upgrade your transcoding box to a newer generation processor that supports, say, AV1 encoding.)
And FWIW my Ryzen Embedded system isn't especially low-power by design, it was just the most accessible way of getting ECC memory for me.
Can do the same with various GPUs, but Quick Sync tends to be the lowest-power and most well-supported at the software level.
This breaks both the 'store key locally' and the KMIP setup.
And for their file-based encryption you cannot change the password. You need to create a new folder with a new password and copy all files over.
- one device died, was EOL at that point, and newer ones no longer can read the disks - stupid limits for array size. Depending on your setup adding disks can mean "copy everything off, delete arrays, and then create new ones". Also, want one 200TB array with your disks? Depending on model size you'll have to do multiple arrays instead with a bit to way lower capacity - syncing a share to another instance is broken, with pretty much no useful debug information. Already the setup is stupid (doesn't let you select which array it goes on the target machine), and then seems to change access permissions of the sync user on the target box (i.e., you can do one sync, after that you'll need to reset the access permissions). I wanted to avoid doing my own sync script, but seems I'll have to do that in the end - stupid disk compatibility warnings (which currently you can disable when you have SSH access) - wireguard only via third party addons. It's 2025. I didn't even check before if those things can do wireguard - it didn't occur to me that a device sold nowadays might not be able to do that.
While debugging I also noticed that pretty much every software component is from the stone age.
On products you can buy TODAY, you find:
They claim it's OK because they've backported all security fixes to their versions. I don't believe them. The (theoretical) huge effort needed for doing that would allow them to grow a way better product.And it's not only about security, but about features (well, some are security features too). We're missing new kernel features (network hardware offload, security, wireguard...), filesystem (btrfs features, performance and error patches...), file servers (new features and compatibility, as Parallel NFS or Multichannel CIFS/SMB), and so on...
I think they got stuck on 4.4 because of their btrfs fork, and now they're too deep on their own hole.
Also, their backend is a mess. A bunch of different apps developed on different ways that mostly don't talk to each other. They sometimes overlap with each other and have very essential features that don't work and don't plan to fix. Meanwhile, they're busy releasing AI stuff features for the "Office" app.
Edit note: For myself and some business stuff, I have a bunch of TrueNAS deployments, from a small Jonsbo box for my home, to a +16 disk rack server. This was for a client that wanted to migrate from another Synology they had on loan, and I didn't want to push a server on them, as they're a bit far away from me, and I wanted it to be serviceable by anyone. I regret it.