If you want to backup your data on Linux, there's no shortage of reliable options that offer some pretty good functionality. We have Déjà Dup that handles the basics well if you just want a simple GNOME app that protects your files without much fuss.

Timeshift takes a different angle; it snapshots your system so you can roll back after a bad update, though it's not really designed for personal data backups.

For users who want more control, Borg and Restic have been the standards for years. Both are encrypted, deduplicated, and trusted by a vast community of Linux users.

BorgBase, on the other hand, is a managed repository hosting service for Borg and Restic backups that has been around for almost a decade. Vykar is their newest project—an open source encrypted backup tool that's worth a look if you're in the market for something new.

Vykar: Overview ⭐

Vykar is an open source, encrypted, deduplicated backup tool written in Rust, developed by the BorgBase team, and released under the GPL-3.0 license. It draws inspiration from the likes of BorgBackup and Borgmatic but uses its own repository format, making it incompatible with existing Borg or Restic repositories.

The whole thing is configured through a single YAML file where you define your repositories, source directories, encryption settings, and retention policy.

Key features include:

• Scheduling via vykar daemon.
• Deduplication is doable via FastCDC.
• Compression support with LZ4 or Zstandard.
• Desktop GUI (vykar-gui) with system tray support.
• WebDAV server for browsing and restoring snapshots.
• Encryption with AES-256-GCM or ChaCha20-Poly1305 (auto-selected), and Argon2id key derivation.
• Concurrent multi-client backups, where multiple machines can write to the same backup repository at the same time.

The desktop GUI deserves a particular mention. Borg relies on third-party tools like Vorta for a desktop interface, and Restic has Backrest, which is also a community project. Vykar ships its own GUI as a first-party component.

It reads directly from vykar.yaml, runs backups on demand, and can sit in the system tray running scheduled backups in the background.

As for the performance, the project's website has put up a comparison of Vykar against Borg, Restic, Rustic, and Kopia, tested against a 49 GiB dataset of 367,000 files.

On backup duration, Vykar finished in 61 seconds, Rustic took 313, Borg 268, Restic 138, and Kopia 85. Restore times follow a similar pattern where Vykar does it in 69 seconds, versus 82 for Rustic, 225 for Borg, 130 for Restic, and 132 for Kopia.

CPU efficiency shows the clearest gap. Vykar used 234 CPU seconds for backup, compared to Borg's 250, Restic's 696, Rustic's 728, and Kopia's 428.

Memory usage is where the picture gets more interesting. Borg uses just 236 MB during backup versus Vykar's 623 MB. Restic is also lighter at 327 MB. So Vykar is trading some RAM for its speed advantage. This is something to factor in on memory-constrained systems.

Repository sizes across all five tools land between 19.7 and 19.9 GB under equivalent Zstd compression settings, so deduplication efficiency is roughly comparable across the board.

Get Vykar 📥

Before you install, know that Vykar supports four storage backends: the local filesystem, S3-compatible object storage (any provider works), SFTP, and a dedicated REST server.

The installation itself is a one-liner:

curl -fsSL https://vykar.borgbase.com/install.sh | sh

Pre-built binaries for Linux (x86_64 and aarch64, both glibc and musl), macOS (Apple Silicon), and Windows are also available on the GitHub releases page.

From there, you can refer to the quickstart guide for going through the creation of a config file and initializing the backup process.

The demand for data centers worldwide has shot up in recent years, with the blind rush for AI being a big reason behind it. I, personally, don't think we need AI shoved into every corner of our lives, but certain regimes and organizations seem to think otherwise, and they keep pushing hard for it.

So, if hearing all that makes you want to just roll up your sleeves and build one yourself, then this indie game might scratch that itch.

Data Center: A Work in Progress

Data Center is an upcoming simulation game from solo developer Václav Novák, who goes by Waseku on Steam. He is based in Brno, Czech Republic, and has been building games since 2020 using the Unity engine.

The main idea here is that you build and manage a data center from scratch. That involves buying racks, servers, and switches; physically placing and stacking your hardware; and then running cables between ports to tie it all together.

It lets the player monitor network traffic via colored, sphere-shaped balls that move around in the cables, and any idle network cable connections are easy to spot. You earn money by serving customers according to their requirements and capacity expectations.

When you process data successfully, you get money, gain experience to unlock better hardware, and build your reputation to draw in larger clients.

Worth Playing?

Possibly. I played the free demo on a Nobara Linux system, and the overall experience was okayish.

First, I configured the graphical settings, disabling NVIDIA DLSS and cranking most settings to their highest possible states. When I went into the game, the game was capped at 50 FPS.

Turns out VSync was the culprit (unsurprisingly), so I disabled it and set a 100 FPS limit.

But the performance woes didn't end there, as throughout gameplay, there were visible drops in the FPS that happened intermittently; the performance graph on the top-right (for FPS) shows it well. 👇

Even forcing the use of Proton Experimental didn’t seem to help, so I rolled with it. Plus, some reviews of the demo on Steam mention this issue, so it was probably that.

The game doesn’t have a proper new player tutorial just yet; it shows some vague objectives, and I had to find my way through the game. The closest thing to an onboarding guide is the Tutorial menu found by pressing the Esc key.

I could read up on how to handle customers, how to configure a server, what kinds of racks and units were available, and a lot more beyond that.

Initially, I was lost and placed patch panels, switches, and servers randomly, without any sense of how these are usually arranged. After a quick web search and closely following what the minuscule object tracker was saying, I was on the correct path.

First, I went to the computer at the desk and headed into the shop. Here, I could unlock some starter server hardware for 0 XP and continue building up from there.

As you can see, I purchased a few patch panels and racks to start populating my data center. The provided utility cart for transferring items from the shipping bay to the server room is, well, jank.

It either keeps any placed items intact or it tosses them off as if it has a mind of its own. Even moving it around is a chore, as it has a habit of clipping below the floor for a quick trip to the underworld.

Placing new racks is another story. You just pick up a rectangular cardboard box and press E to install it over a designated spot on the floor. And to place network switches, patch panels, or servers, a handy yellow-colored outline is shown in the available space on a rack.

Wiring all of that up is effortless too. After purchasing 100 meters (🦅 328 ft) of CAT6E cable, I was able to connect the server hardware, with some slip-ups that I fixed later during my play.

Onboarding new customers is as simple as connecting a wire from the network switch to the outbound connection panel thingy; I am unsure what that's called. Here, you select a customer, review what they want, and then a small guide pops up on how to get them hooked up.

I then turned on the power for the servers and set the IP addresses from the individual control panel to get the connection up. I had to add a few more servers and turn on the network switches before I could serve the customer, "Bermuda Triangle Backup" at 40000 IOPS speed.

If you noticed the abysmal cable management job I did in the video above, then fret not; I have redeemed myself by fixing it (in the video below). But before that, you have to know that there are handy cable management loops across the racks and the roof of the server room that make your job easier.

You can also increase the space of your server room by purchasing additional space for 1,000 in-game currency per block. If you see a pile of strewn-about servers, ignore it.

Making use of the aforementioned mounts, I neatly rerouted the Ethernet cabling to clean up the spiderweb-like arrangement I had in place. As you saw, it is fairly easy to do, provided you have some patience and a little obsession with keeping things in proper alignment.

Overall, the game needs a lot of polish before launch. A proper tutorial would be a good start because right now new players are pretty much left to figure things out on their own.

On my wishlist for later upgrades would be multi-storey data centers, a utility cart that actually works without going into a void, and a computer portal that isn't so basic and awkward to use.

How to Play?

A free demo with some restrictions has been available on Steam for some time now. The game is planned to launch on March 31, 2026, though without any pricing information yet.

There is also no mention of an early access release, so it looks like Václav is aiming for a stable 1.0 release.

matrixOS brings OSTree atomic upgrades to Gentoo with a simple motto: "emerge once, deploy everywhere."

If you're juggling multiple tasks and want something simpler than Jira or ClickUp, this is worth trying.