Applicable Versions | NetSim Academic | NetSim Standard | NetSim Pro |
Applicable Releases | v13.2 |
We explain how to model and analyze interference in WiFi networks via a simple 2AP-2STA scenario.
The 2AP-2STA scenario
There are two APs and two STAs placed in a line. The APs are at the ends and the STAs in between. The AP-STA distance is D while the AP-AP distance is d; STA-STA distance is therefore D-2d.
Figure 1: A simple 2-cell scenario comprising 2 APs and 2 STAs. The AP1 and STA1 (and AP2 and STA2) send data to one another over the wireless link L1 (and L2 resp.). The AP-STA separation is d while the AP-AP separation is D.
We change the value of D while keeping d fixed at 14m and show two cases. Case #1 is where the AP-STA pairs have sufficient spatial separation to be independent and case #2 where STA1 and STA2 are in carrier sense and interference range.
Carrier Sensing (CS)
In the 802.11 CS mechanism, a wireless station withholds its transmission when it senses an ongoing transmission on the medium. It is a sender-side phenomenon, unlike interference which is a receive-side phenomenon. The CS threshold is defined as the min receive sensitivity at the control (lowest) rate which is -82 dBm (users can change the CS threshold in NetSim's WiFi C-code).
In this example, when STA1 transmits data to AP1, if the power at STA2 (due to STA1) is greater than -82dBm then STA2 will detect the medium as busy and will not transmit to AP2.
MCS Table for 802.11 g
Index | Min Rx Power (dBm) | Modulation | Code Rate | Bit Rate |
1 | -82 | BPSK | 1/2 | 6 Mbps |
2 | -81 | BPSK | 3/4 | 9 Mbps |
3 | -79 | QPSK | 1/2 | 12 Mbps |
4 | -77 | QPSK | 3/4 | 18 Mbps |
5 | -74 | 16 QAM | 1/2 | 24 Mbps |
6 | -70 | 16 QAM | 3/4 | 36 Mbps |
7 | -66 | 64 QAM | 2/3 | 48 Mbps |
8 | -65 | 64 QAM | 3/4 | 54 Mbps |
Table 1: 802.11g bit rates for different modulation schemes, and the minimum received signal power and SINR required for achieving each bit rate.
Interference: SINR-based decoding
Even though the received signal strength (RSS) is above the value shown in Table 1, the corresponding PHY rate may not be decodable at a receiver if the received signal power is not sufficiently large as compared to the sum of the receiver noise, and the cochannel interference from simultaneous transmissions, i.e., unless the SINR is large enough.
Figure 2 shows the SINR vs. PER curves used in NetSim for the different MCS supported in 802.11g. This example in this document focuses on the 54 Mbps PHY rate. Interference causes packet failure with probability 1 when SINR < 20 dB while packets succeed with probability 1 even with interference when SINR > 21.5 dB.
Let us say AP1 is transmitting to STA1 while STA2 is simultaneously transmitting to AP2. Interference can occur at STA1 i.e., STA2 transmission (to AP2) degrades the SINR at STA1.
Recall again that interference is a receiver-side phenomenon while CS blocking is a sender-side phenomenon.
Figure 2: NetSim’s 802.11g PER SINR curves for various PHY rates (MCSs).
Radio Propagation
The radio propagation model is the log-distance model whereby, if a transmitter transmits at power to a receiver at distance D (meters) (where it is assumed that D > 1), then the received power is given by:
The 40.09 dB loss is at the distance of 1 meter, and this value holds for the 2.4GHz band. We use a pathloss exponent of 3.5 in this example.
Case 1: No interference
Power at STA1 due to STA2
Since power at S2 due to S1, is less than -82 dBm the two STAs are not in the carrier sense range. Next, we compute the received power at STA1 from AP1.
At this received power, we see from the tables that the PHY rate is 54 Mbps. Assuming interference is much higher than noise, the SINR at S1 would be
From the SINR tables, we see that the SINR threshold for successfully decoding at a PHY rate of 54 Mbps is 21.5dB. The SINR at S1 is above this threshold. We thus see that each AP-STA is independent.
Case 2: STA1 and STA2 are in the interference range
Now, power at STA1 due to STA2
Since power at S2 due to S1 is greater -82 dBm than the two STAs is in the carrier sense range. At this received power, we see from the tables that the PHY rate is 54 Mbps. Assuming interference is much higher than noise, the SINR at S1 (A1 is the transmitter and S2 is the interferer) is
From the SINR tables, we see that the SINR threshold for decoding packets with probability 0 for a PHY rate of 54 Mbps is 19.5 dB. The SINR at S1 is below this threshold. We thus see the realization of a scenario where the STAs in the interference range.
NetSim Scenario
NetSim Scenario Properties:
MAC layer parameters | |
Standard | IEEE802.11g |
Operating Frequency | 2.4GHz |
Rate (MCS) selection | False. Auto rate fall back |
RTS Threshold | 3000bytes for Basic access cases 1000bytes for RTS/CTS cases |
Frequency Band | 2.4GHz |
Bandwidth | 20MHz |
Transmitter Power | 100mW for both APs and both STAs |
Medium Access Protocol | DCF |
Propagation Model (Wireless Link) | |
Channel Model | Path Loss Only (No shadowing or fading) |
Path Loss Model | Log Distance |
Pathloss Exponent () | 3.5 |
Traffic Model (Server1 to STA1, STA1 to server1, Server2 to STA2, STA2 to server 2) | |
Packet Size | 1460 B |
Packet Distribution | Constant |
Interpacket arrival time | 50 µs (leads to full buffer or infinite backlog) |
Other parameters | |
Transport Protocol | UDP |
Simulation Time | 10 seconds |
The experiment config files are attached.
Results (Basic Access)
Case | Description | Throughput (Mbps) | |||
A1-S1 | S1-A1 | A2-S2 | S2-A2 | ||
1 | Independent | 14.40 | 14.50 | 14.42 | 14.48 |
2 | S1S2 - Interference | 0.10 | 16.73 | 0.09 | 16.75 |
Case #1: There are no hidden node collisions, only simultaneous attempt collisions. This is essentially a 2 node Bianchi model
Case #2: In this case, the transmissions from AP1 and AP2 fail due to hidden node collisions (due to interference) from STA1 or STA2, respectively. This reduces the attempt rates of these nodes, resulting in lower attempt rates, thus giving more airtime to STA1 and STA2.
References
1. Performance analysis of the IEEE 802.11 distributed coordination function, G. Binachi. IEEE Journal on Selected Areas in Communications (Volume: 18, Issue: 3, March 2000)
Useful Links
1. Detailed white paper on modeling carrier sense and interference in NetSim WiFi: https://tetcos.com/wifi-white-papers.html