Latency problem

[Nooshin Eghbal]

Hello,

I was reading Development: Latency page and wondering what is the current strategy of TigerVNC out of the ideas explained in this page?

Thanks,

Nooshin

[Pierre Ossman]

On 03/10/2022 22:46, Nooshin Eghbal wrote:
> Hello,
>
> I was reading Development: Latency

> <https://github.com/TigerVNC/tigervnc/wiki/Development%3A-Latency> page and

> wondering what is the current strategy of TigerVNC out of the ideas
> explained in this page?
>

The last one, full congestion control based on TCP Vegas. There have
been numerous tweaks to the code since that document was written as well.

Regards
--
Pierre Ossman Software Development
Cendio AB https://cendio.com
Teknikringen 8 https://twitter.com/ThinLinc
583 30 Linköping https://facebook.com/ThinLinc
Phone: +46-13-214600

A: Because it messes up the order in which people normally read text.
Q: Why is top-posting such a bad thing?

[Nooshin Eghbal]

Thanks for your response.

If TigerVNC is running over TCP, why does it need to have its own congestion control?
And why you have not used BBR as a newer non-loss-based CCA?

Is there any updated documentation?

> --
> You received this message because you are subscribed to a topic in the Google Groups "TigerVNC Developer Discussion" group.
> To unsubscribe from this topic, visit https://groups.google.com/d/topic/tigervnc-devel/HXAeas5WMz0/unsubscribe.
> To unsubscribe from this group and all its topics, send an email to tigervnc-deve...@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/tigervnc-devel/4af058b2-2cb6-7d5e-09f7-6f4860d2637c%40cendio.com.
>

[Pierre Ossman]

On 05/10/2022 11:35, Nooshin Eghbal wrote:
> Thanks for your response.
>
> If TigerVNC is running over TCP, why does it need to have its own
> congestion control?

It's detailed in the wiki page you linked. But in short, relying on just
TCP's own congestion control gives a very high latency, as we often get
excessive buffering.

> And why you have not used BBR as a newer non-loss-based CCA?
>

Because TigerVNC's code is much older than BBR. And the existing code
has worked well enough that there hasn't been any incentive to look at a
different algorithm.

If you want to have a look and see if things can be improved, then you
are very much welcome to do so.

> Is there any updated documentation?
>

I'm afraid not. You'll have to check the code:

https://github.com/TigerVNC/tigervnc/blob/master/common/rfb/Congestion.cxx

[Nooshin Eghbal]

On Wednesday, October 5, 2022, Pierre Ossman <oss...@cendio.com> wrote:

> On 05/10/2022 11:35, Nooshin Eghbal wrote:
>>
>> Thanks for your response.
>>
>> If TigerVNC is running over TCP, why does it need to have its own
>> congestion control?
>
> It's detailed in the wiki page you linked. But in short, relying on just TCP's own congestion control gives a very high latency, as we often get excessive buffering.
>
>> And why you have not used BBR as a newer non-loss-based CCA?
>>
>
> Because TigerVNC's code is much older than BBR. And the existing code has worked well enough that there hasn't been any incentive to look at a different algorithm.

Is there any set of results to show how TigerVNC can improve the performance over high RTT networks compared to any other VNC implementation?

[Pierre Ossman]

On 05/10/2022 12:56, Nooshin Eghbal wrote:
>
> Is there any set of results to show how TigerVNC can improve the
> performance over high RTT networks compared to any other VNC implementation?
>

No, I'm afraid we haven't kept data sets from development and testing.
The difference should be apparent, though, as frame rate will decrease
linearly with RTT using other¹ VNC implementations. TigerVNC, on the
other hand, should be able to maintain full bandwidth usage.

¹ with the exception of TurboVNC, which implements the same system as
TigerVNC.