hidden terminal problem in 802.11

[Abhishek_Anand]

Hi,

I have been trying simulate the hidden terminal problem in 802.11 but
with no luck.

I have set EnergyDetectionThreshold such that my transmission range is
100m. One access point is placed at -90m and other at 95m. Both send
data to a client node at 0m simultaneously using udp sockets. Now
there should be packet loss at client node for the data from the ap
at -90m due to traffic from -95m but this is not happening. All the
packets from the node at -90m are received and all the packets from
the node at 95m are dropped. RTS/CTS mechanism has been disabled.

So, how to demonstrate the hidden terminal scenario in 802.11?

--
You received this message because you are subscribed to the Google Groups "ns-3-users" group.
To post to this group, send email to ns-3-...@googlegroups.com.
To unsubscribe from this group, send email to ns-3-users+...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/ns-3-users?hl=en.

[Quincy]

Simulation results sounds plausible. It may already be showing the
hidden terminal problem:

The hidden terminal problem is mainly talking about the fact that two
nodes transmit at the same time because they cannot hear each other,
causing interference at some other point. It does not dictate what the
outcome of the interference may be.

In your case, it could be interpreted that the packet from -90m and
the packet from 95m both arrived at the Rx node. The observation that
the packet from 95m are dropped when the Tx range is 100m may indicate
that both packets are interfering with each other (hidden terminal
effect - otherwise, both packets will be received). The packet from
95m is interfering with the packet from -90m, but the packet from 90m
can still be decoded correctly because the interference noise caused
by the packet from 95m is too low.

There is also another problem in your experiment setup - the packet
from 90m away will always arrives at the Rx node earlier than the
packet from 95m away, therefore, due to the lack of frame/preamble
capture support in NS3, you'll never be able to receive the packet
from 95m away regardless of the transmission power, but depending on
the Tx power of the packet from 95m away, you may or may not lose the
packet from 90m.

Quincy

[Quincy]

The thing I'd like to emphasise - when there is the Hidden Terminal
Problem, it does NOT mean that both packets are lost. Assuming a
single antenna, single radio system, you may (or may not) correctly
receive a transmission from one of the concurrent transmitters. See
the other discussions on SINR, Tx gains, etc.

Quincy

[pankaj]

It appears that the whole point to simulate hidden terminal is to
carry
out an experiment that can suppress an intended(-90m here)
communication due
to the presence of other interfering traffic ( from 95m here).

On May 3, 9:46 am, Quincy <quincy....@gmail.com> wrote:
> The thing I'd like to emphasise - when there is the Hidden Terminal
> Problem, it does NOT mean that both packets are lost. Assuming a
> single antenna, single radio system, you may (or may not) correctly
> receive a transmission from one of the concurrent transmitters. See
> the other discussions on SINR, Tx gains, etc.
>

I don`t expect that both packets are lost (definitely, PER = 1)
but I think that there should be way to suppress the intended
communication
with a sufficient PER to get a packet drop out of 10 or 100 or a
reasonable
number of packets sent (from -90m here)

> Quincy
>
> On May 3, 2:42 pm, Quincy <quincy....@gmail.com> wrote:
>
>
>
> > Simulation results sounds plausible. It may already be showing the
> > hidden terminal problem:
>
> > The hidden terminal problem is mainly talking about the fact that two
> > nodes transmit at the same time because they cannot hear each other,
> > causing interference at some other point. It does not dictate what the
> > outcome of the interference may be.
>

If the outcome is not dictated, does this means that intended
communication
can simply be unaffected? Please clarify.

Should the outcome not conform about the suppression of intended
communication?

> > In your case, it could be interpreted that the packet from -90m and
> > the packet from 95m both arrived at the Rx node. The observation that
> > the packet from 95m are dropped when the Tx range is 100m may indicate
> > that both packets are interfering with each other (hidden terminal
> > effect - otherwise, both packets will be received).

.
Both packets will be received but can this be counted as an increase
in throughput? Since the 95m packet was not intended for this
receiver
situated at 0m (Rx0) . It could be sending the data to other receiver
and Rx0 happens to be in the range. Hence no suppression of intended
communication is observed(i.e no -90m packet drops.)

> > The packet from
> > 95m is interfering with the packet from -90m, but the packet from 90m
> > can still be decoded correctly because the interference noise caused
> > by the packet from 95m is too low.
>

Is the distance of 95m not sufficiently close to cause sufficient
interference
noise? As it turns out that while carrying out the same simulation , I
observed
the PER to be very small ( 10e-12), which definitely is not
sufficient.

How can I sufficiently increase the PER to lets say 0.2 so that 2
packets are
dropped out of 10 packets sent from -90m while using the same hidden
terminal
configuration?

> > There is also another problem in your experiment setup - the packet
> > from 90m away will always arrives at the Rx node earlier than the
> > packet from 95m away, therefore, due to the lack of frame/preamble

> >capturesupport in NS3, you'll never be able to receive the packet

> > from 95m away regardless of the transmissionpower,

As it appears that there is no problem with 95m packet being dropped.
The
whole point is to carry out an experiment that can suppress an
intended(-90m)
communication due to the presence of other interfering traffic (95m)

> > but depending on
> > the Txpowerof the packet from 95m away, you may or may not lose the
> > packet from 90m.
>
Please clarify that how much larger do you expect Txpower that may
cause
a packet drop from -90m

[Jaume]

I've just started to work with ns-3 and I am particularly interested
in the hidden node problem. Does anyone have some example code to
start working with?

Thanks in advance,
Jaume

On May 2, 11:15 pm, Abhishek_Anand <anandshocking...@gmail.com> wrote:
> Hi,
>

> I have been trying simulate thehiddenterminalproblem in 802.11 but

> with no luck.
>
> I have set EnergyDetectionThreshold such that my transmission range is
> 100m. One access point is placed at -90m and other at 95m. Both send
> data to a client node at 0m simultaneously using udp sockets. Now
> there should be packet loss at client node  for the data from the ap
> at -90m due to traffic from -95m but this is not happening. All the
> packets from the node at -90m are received and all the packets from
> the node at 95m are dropped. RTS/CTS mechanism has been disabled.
>

> So, how to demonstrate thehiddenterminalscenario in 802.11?

[Pavel Boyko]

Take a look at examples/wireless/wifi-hidden-terminal.cc

Pavel

[Jaume]

Thanks Pavel!

[Jaume]

Hi Pavel, the example you provided has been very useful to me. I
enabled tracing and I realized that there were many icmp port
unreachable packets that were interfering with the measures. I
installed an udp server to get rid of those messages

Thanks again,
Jaume