How to connect via remote desktop from Windows 10 to freebsd

Zirias said:
Well, my experience is different. Of course, on the LAN, there's no issue. But WAN? For a long time, I had one of these 16MBit DSL lines (asymmetric, with upstream only a fraction of it, probably 1MBit or something like this). X11 wasn't usable. RDP was.
Click to expand...
Yeah X11 isn't appropriate for WAN. It is far too slow, in particular the latency for each request.

RDP is the fastest I have used so far. Which is annoying considering the network is one of UNIX's strengths ;)

For the multi-terminal switching, there is Xpra but it seems unportable and/or a pain to set up.
 
Oh I forgot to add.

If you are stuck with X11 forwarding through SSH over the internet, you can eliminate a number of these round trips by running something like Xephyr and simply forwarding that *one* application. This may sound counter-intuitive in terms of performance but if you consider the latency of these round trips being the issue, you want those to be between software on the same machine (the remote server). Even better, they can use a UNIX domain socket.

Then the image as a whole will be sent to the connecting client rather than an individual trip for many different UI components.
 
Zirias said:
RDP isn't the "holy grail", but it's far more capable than X11. Don't get me wrong, the "network transparency" design of X windows definitely was a good idea. But the implementation is way behind of what you'd expect nowadays. From that point of view, I can even understand how a project like "Wayland" came to be.
Click to expand...
I do vaguely recall X11 allowing concurrent remote sessions (if you set it up properly), unlike RDP. As for Wayland, it's supposed to have addressed at least some of the weaknesses of X11's design.
 
astyle said:
As for Wayland, it's supposed to have addressed at least some of the weaknesses of X11's design.
Click to expand...
It by design doesn't specifically address these weaknesses. Instead it simply classes them as "out of scope" and leaves it to the individual Wayland Compositors to implement an ad-hoc solution. One of these common solutions is to use X11 on-top of Wayland which actually solves very little.

It is a bit like saying the UI system on Windows 95 was more secure than X11 because it doesn't listen on a socket. Technically true but very old fashioned and limited.
 
Hm, kpedersen, I see nothing wrong with "outscoping" in general, especially having the SRP in mind. Of course this doesn't mean that the problem at hand can just be ignored, it has to be solved somewhere else. So, of course running X11 on top of wayland is bollocks ;)

But in general: Why should network access (remote GUI) be part of the core GUI infrastructure? Probably makes more sense to separate it indeed.
 
Zirias said:
Hm, kpedersen, I see nothing wrong with "outscoping" in general, especially having the SRP in mind. Of course this doesn't mean that the problem at hand can just be ignored, it has to be solved somewhere else.
Click to expand...
In principle, yes. However it means that the Gnome 3 developers are going to have to write a secure remote desktop system... They can't even write a useful desktop environment so how can they ever hope to implement / maintain the equivalent of half of Xorg?

And then KDE, Xfce, (presumably dwm, Fvwm), etc, would all have to implement their own. Its going to be chaos. Or it simply won't happen and we will keep on with a (then) slightly broken and unsupported Xorg layer.

Can we just rename "Xorg" to "WaylandRemoted" and see if we can trick the hipsters so we can get on with our lives? ;)
 
What you describe would be a result of having the right idea, but doing it wrong. You could probably just as well have reusable components implementing RDP, VNC, whatever and have them used by any desktop environment. That would clearly make more sense.

I never tried Wayland so far, I can't say how good or bad it is. I just think the initial idea wasn't that bad after all ;)
 
Zirias said:
You could probably just as well have reusable components implementing RDP, VNC, whatever and have them used by any desktop environment. That would clearly make more sense.
Click to expand...
I suppose it starts to become how far they have to go before they realize that they could just as easily used XWayland and Xlib. However most compositors will probably just fall back on shoving libvncserver in there. I suppose that won't be too difficult. Not quite the same thing as RDP/XDMCP multi-users though.

Zirias said:
I never tried Wayland so far, I can't say how good or bad it is. I just think the initial idea wasn't that bad after all ;)
Click to expand...
I honestly like the idea of simplification. I just am not entirely sure that this simplification will exist once actually put into practice for typical window managers. Once we have a few more than just Sway/wlroots (which is *not* a simple codebase), we might start seeing some evidence one way or the other.
 
astyle said:
Gonna be awhile before we can get rid of x11-servers/xwayland and x11-servers/xwayland-devel. I'd love to see that happen, though. In the meantime, Wayland on FreeBSD is making steady progress.
Click to expand...
And this makes perfect sense as an adapter to simplify migration. But if the end result is that everyone uses this adapter just because noone implements the required features natively on Wayland, the whole thing would be a failed project, that delivered complication instead of simplification.
 
I use it on my CentOS box. However, x2go only has the client for FreeBSD, not the server part. It really is fast though. I've used it from my workstation at work to the CentOS box at home and it's almost like being at the CentOS box's keyboard.
 
From my understanding on Wayland, they ripped out the remote connection so Wayland by it's self doesn't do any of that stuff and depends on VNC or something else to provide the remote desktop functionality.
 
You must log in or register to reply here.
Back
Top