[gentoo-user] X-Forwarding over wireless

[gentoo-user] X-Forwarding over wireless

[Grant]

Is anyone using X-Forwarding over a local wireless connection?  I'm
forwarding a couple of light apps and they work fine with -Y but -X is
unusable.  I've been trying to use vmware workstation with an XP guest
OS over wireless, but it's no good.  Would I have better luck with
vnc?

- Grant
-- 
gentoo-user@gentoo.org mailing list

Re: [gentoo-user] X-Forwarding over wireless

[Jakob]

Did you try to export the display manually?
I also got some problems using -X but it works perfect by using
"export DISPLAY=<your ip>:0"
maybe that helps.
-- 
gentoo-user@gentoo.org mailing list

Re: [gentoo-user] X-Forwarding over wireless

[Grant]

> > Is anyone using X-Forwarding over a local wireless connection?  I'm
> > forwarding a couple of light apps and they work fine with -Y but -X is
> > unusable.
>
> Not sure why it would work with -Y but not -X, but in any case if it
> works with -Y why not just use that?
>
> But yes, I do use X11 forwarding over wireless (actually wireless and
> through a VPN) and it's been reliable... well rather than reliable I
> should say as reliable as any other wireless connection.

I wonder if I need more power on the machine running the apps instead
of more bandwidth.  It has 512MB and I do need to upgrade that, but it
feels like a bandwidth problem when I'm running vmware via
X-Forwarding.

- Grant
-- 
gentoo-user@gentoo.org mailing list

Re: [gentoo-user] X-Forwarding over wireless

[Albert Hopkins]

On Thu, 2007-02-22 at 05:06 -0800, Grant wrote:
> I wonder if I need more power on the machine running the apps instead
> of more bandwidth.  It has 512MB and I do need to upgrade that, but it
> feels like a bandwidth problem when I'm running vmware via
> X-Forwarding.
> 

Could be bandwidth.  Wifi, as you know is relatively slow.  Even if you
have 802.11G which is "rated" for 54Mbps you never actually get 54MBbps.
If you run an access point in hybrid b/g mode, that slows it down even
more.  If you have WEP/WPA encryption that slows it down even more.  If
you tunnel through ssh encryption that slows it down even more.  Add
that to the fact that wireless connections are notoriously unreliable
and your typical X11 app is *very* chatty then I would not expect
gigabit-ethernet performance.


-- 
gentoo-user@gentoo.org mailing list

[gentoo-user] Re: X-Forwarding over wireless

[Grant Edwards]

On 2007-02-22, Albert Hopkins <marduk@gentoo.org> wrote:
> On Thu, 2007-02-22 at 05:06 -0800, Grant wrote:
>> I wonder if I need more power on the machine running the apps instead
>> of more bandwidth.  It has 512MB and I do need to upgrade that, but it
>> feels like a bandwidth problem when I'm running vmware via
>> X-Forwarding.
>> 
>
> Could be bandwidth.

More likely it's latency.  Most "modern" X apps seem to require
a lot of round-trips between client and server.  The latency of
a Wifi link is probably 10-100X that of a wired Ethernet link,
even if the bandwidth is the same:  A 54M Wifi link may
actually have more bandwidth than a 10M wired Ethernet link,
but the lower latancy of the wired link will result in
better performance for some classes of applications.

-- 
Grant Edwards                   grante             Yow!  I'm continually
                                  at               AMAZED at th'breathtaking
                               visi.com            effects of WIND EROSION!!

-- 
gentoo-user@gentoo.org mailing list

Re: [gentoo-user] Re: X-Forwarding over wireless

[Grant]

> >> I wonder if I need more power on the machine running the apps instead
> >> of more bandwidth.  It has 512MB and I do need to upgrade that, but it
> >> feels like a bandwidth problem when I'm running vmware via
> >> X-Forwarding.
> >>
> >
> > Could be bandwidth.
>
> More likely it's latency.  Most "modern" X apps seem to require
> a lot of round-trips between client and server.  The latency of
> a Wifi link is probably 10-100X that of a wired Ethernet link,
> even if the bandwidth is the same:  A 54M Wifi link may
> actually have more bandwidth than a 10M wired Ethernet link,
> but the lower latancy of the wired link will result in
> better performance for some classes of applications.

Do you think vnc or nx would be a significant improvement over x-forwarding?

- Grant
-- 
gentoo-user@gentoo.org mailing list

Re: [gentoo-user] Re: X-Forwarding over wireless

[Uwe Thiem]

On 22 February 2007 16:55, Grant Edwards wrote:

> More likely it's latency.  Most "modern" X apps seem to require
> a lot of round-trips between client and server.  The latency of
> a Wifi link is probably 10-100X that of a wired Ethernet link,
> even if the bandwidth is the same:  

Where do you get that number from?

I can not imagine any reason why wifi should have alatency one or two levels 
of magnitude higher than wires.

Uwe

-- 
A fast and easy generator of fractals for KDE:
http://www.SysEx.com.na/iwy-1.0.tar.bz2
Proof of concept of a TSP solver for KDE:
http://www.SysEx.com.na/epat-0.1.tar.bz2
-- 
gentoo-user@gentoo.org mailing list

Re: [gentoo-user] Re: X-Forwarding over wireless

[Uwe Thiem]

On 22 February 2007 17:05, Grant wrote:
> > >> I wonder if I need more power on the machine running the apps instead
> > >> of more bandwidth.  It has 512MB and I do need to upgrade that, but it
> > >> feels like a bandwidth problem when I'm running vmware via
> > >> X-Forwarding.
> > >
> > > Could be bandwidth.
> >
> > More likely it's latency.  Most "modern" X apps seem to require
> > a lot of round-trips between client and server.  The latency of
> > a Wifi link is probably 10-100X that of a wired Ethernet link,
> > even if the bandwidth is the same:  A 54M Wifi link may
> > actually have more bandwidth than a 10M wired Ethernet link,
> > but the lower latancy of the wired link will result in
> > better performance for some classes of applications.
>
> Do you think vnc or nx would be a significant improvement over
> x-forwarding?

nx definitely would. I saw a whole KDE session over an ISDN link - and it felt 
like local.

Uwe

-- 
A fast and easy generator of fractals for KDE:
http://www.SysEx.com.na/iwy-1.0.tar.bz2
Proof of concept of a TSP solver for KDE:
http://www.SysEx.com.na/epat-0.1.tar.bz2
-- 
gentoo-user@gentoo.org mailing list

[gentoo-user] Re: X-Forwarding over wireless

[Grant Edwards]

On 2007-02-22, Grant <emailgrant@gmail.com> wrote:
>> >> I wonder if I need more power on the machine running the apps instead
>> >> of more bandwidth.  It has 512MB and I do need to upgrade that, but it
>> >> feels like a bandwidth problem when I'm running vmware via
>> >> X-Forwarding.
>> >>
>> >
>> > Could be bandwidth.
>>
>> More likely it's latency.  Most "modern" X apps seem to require
>> a lot of round-trips between client and server.  The latency of
>> a Wifi link is probably 10-100X that of a wired Ethernet link,
>> even if the bandwidth is the same:  A 54M Wifi link may
>> actually have more bandwidth than a 10M wired Ethernet link,
>> but the lower latancy of the wired link will result in
>> better performance for some classes of applications.
>
> Do you think vnc or nx would be a significant improvement over
> x-forwarding?

I've never directly compared the them, but I've seen posts by
others saying they've have had better luck with vnc or nx on
high latency links.

-- 
Grant Edwards                   grante             Yow!  I am having FUN... I
                                  at               wonder if it's NET FUN or
                               visi.com            GROSS FUN?

-- 
gentoo-user@gentoo.org mailing list

[gentoo-user] Re: X-Forwarding over wireless

[Grant Edwards]

On 2007-02-22, Uwe Thiem <uwix@iway.na> wrote:
> On 22 February 2007 16:55, Grant Edwards wrote:
>
>> More likely it's latency.  Most "modern" X apps seem to require
>> a lot of round-trips between client and server.  The latency of
>> a Wifi link is probably 10-100X that of a wired Ethernet link,
>> even if the bandwidth is the same:  
>
> Where do you get that number from?

My Wifi network often has latencies of 50-100ms, while typical
wired latencies are 1-5ms.  I assumed that's typical.  It could
be there's something screwy in my WAP -- it does lock up not
infrequently.

> I can not imagine any reason why wifi should have alatency one
> or two levels of magnitude higher than wires.

-- 
Grant Edwards                   grante             Yow!  I've got to get
                                  at               these SNACK CAKES to NEWARK
                               visi.com            by DAWN!!

-- 
gentoo-user@gentoo.org mailing list

Re: [gentoo-user] Re: X-Forwarding over wireless

[Hans-Werner Hilse]

Hi,

On Thu, 22 Feb 2007 15:49:42 +0000 (UTC)
Grant Edwards <grante@visi.com> wrote:

> On 2007-02-22, Uwe Thiem <uwix@iway.na> wrote:
> > On 22 February 2007 16:55, Grant Edwards wrote:
> >
> >> More likely it's latency.  Most "modern" X apps seem to require
> >> a lot of round-trips between client and server.  The latency of
> >> a Wifi link is probably 10-100X that of a wired Ethernet link,
> >> even if the bandwidth is the same:  
> >
> > Where do you get that number from?
> 
> My Wifi network often has latencies of 50-100ms, while typical
> wired latencies are 1-5ms.  I assumed that's typical.  It could
> be there's something screwy in my WAP -- it does lock up not
> infrequently.

I think that's your WAP. On my link, the latency is and stays at about
2.8 msec (11MBit 802.11b link). If you have a userland daemon involved,
you might get better results w/ a high HZ value.

-hwh

hw@butch ~ $ ping lsys
PING lsys (192.168.2.1) 56(84) bytes of data.
64 bytes from lsys (192.168.2.1): icmp_seq=1 ttl=64 time=2.59 ms
[...]
64 bytes from lsys (192.168.2.1): icmp_seq=20 ttl=64 time=2.63 ms

--- lsys ping statistics ---
20 packets transmitted, 20 received, 0% packet loss, time 18997ms
rtt min/avg/max/mdev = 2.511/2.756/3.162/0.208 ms

-- 
gentoo-user@gentoo.org mailing list