1. 4

I just spun up a remote desktop for a particular application that required a desktop application to stay open, so that I could run it 24 hours and not have to worry about my laptop sleeping, etc.

Inspired me to revisit the concept of remote desktop as primary work space for all those daily web based applications that I have open on my desktop: gmail, twitter, fb (sorry), terminals, etc.

Looking for feedback from lobste.rs if anyone does this, experiences, upsides, downsides.

Security is my main concern and I’m primarily trying to get a handle on for example how gmail left open on a remote desktop on a server that I manage (would be locked down, but I’m not a security expert) compares to gmail left open on my laptop at home.




  2. 2

    I’ve heard about people losing their work because they lose access to the VPS for whatever reason (lost key, deleted the wrong image, etc). To prevent this, you could treat your workspace as a cache of what you’re working on.

    Additionally, NixOS makes it really easy to declaratively reconfigure your environment, all you need is to customize your /etc/nixos/configuration.nix. Personally, I have my beefy work laptop setup as a server sitting under my desk, and I ssh into it to run certain tests that my personal laptop can’t handle (e.g. large datasets for machine learning). Using NixOS, I can wipe the entire hard drive on my work laptop, and set it back up to working condition (i.e., I can ssh into it) in a few minutes. Or I could replicate it on a remote VPS or whatever. I don’t know of any other OS that makes it so easy to do that.

    1. 2

      This is not all that different than developing in a local VM per project.

      I worked in an environment where I could only use a very bad Windows box where I instituted a similar process. Instead of doing a full remote session on the box I would just Putty into the machine and do all of my development in tmux and vim, it worked great.

      As to security, you can run your X session through ssh and it’ll be as secure as your private key (don’t use passwords!). The only other concern is the general security of your cloud provider but there isn’t much you can do there. You probably aren’t a big enough target for someone to be trying to hit you with cross vm attacks anyways, so no need to get too paranoid :)

      1. 1

        I think I’d still do ‘heavy’ things (e.g. IntelliJ) locally, just use the remote desktop for lighter, but ‘always on’ things.

        Yeah, I guess I’m more worried about my ability to lock down a Linux installation than I am about the provider.

        Security of the connection should be ok - connecting by ssh seems to be doable.

      2. 2

        You could use some kind of VNC that listens on local host and establish a connection via an ssh tunnel.

        1. 1

          When I’m home connecting into work (which I do about 5 hrs each week) I’ve got the following layers to get through:

          Linux on Laptop -> Qemu -> Windows 7 VM -> VPN with custom USB Token -> RDP -> Windows 7 on Work desktop -> VirtualBox -> Ubuntu VM

          1. 2

            I bet that’s a joy to use.