The MD4350 3-Radio Mesh Module was tested at the USAF Force Protection Lab and with approved network performance measuring equipment. The MD4350 3-Radio contains two 802.11a radios for backhaul and one 802.11b/g service radio. The following tests were conducted. [Other tests related to wireless network security are also available on request] 
 
     
Figure 1: (Left) Test Set up of 3-Radio modules          Figure 2: (Right) Connectivity diagram of MD4350 3-Radio nodes.

Test 1: Bandwidth preservation over multiple hops.  After much debate, it was conceded that the performance of single radio mesh networks degrades between 1/n (in the best case) and (1/2)ⁿ (typical case) with each hop away from the Ethernet feed.  MeshDynamics claimed that since we use a third generation backhaul it has minimal degradation with each hop. To validate these claims, each Mesh node was isolated in such a way that it will only "see" the mesh node as indicated above. Throughput at each hop and the bridge delay per hop was measured. Throughput was measured by using IXIA Test Software. The introduced delay was measured by using pings.  The results are depicted in Table 1. Note: These are TCP/IP numbers, not raw data rate numbers.
 

No of Hops from Root	Throughput (Mbps)	Bridge Delay (ms)
1	22-23 Mbps TCP/IP  (54 Mbps Raw)	Less than 1 ms
2	22-23 Mbps TCP/IP (54 Mbps Raw)	Less than 2 ms
3	22-23 Mbps TCP/IP (54 Mbps Raw)	Less than 3 ms

Table 1: Backhaul Throughput performance on 3 radio Structured Mesh Module (Model 4350)
 

Figure 3: TCP/IP throughput measured at 1, 2 and 3 hops (3-Radio Structured Mesh).
    

Figure 4: Bridge Latency measured at 3 hops is less than 3 ms. (3-Radio Structured Mesh).

Test 2: Bridge Latency over multiple hops  Latency sensitive transmissions (e.g. VOIP and Video streams over mesh) , are affected by the delay in wireless bridges. Recall that in the 3-radio Structured Mesh modules, there are dedicated service and backhaul radios. All data passes from the service radios to the backhaul radios via a high speed low latency wireless switch/router implemented in each Structured Mesh module. Over multiple hops, the latency introduced by the wireless bridge could be significant. This adversely affect both VOIP & Video.  As shown in Figure 4, the bridge latency introduced - at the last hop - was noted to be less than 3 ms. Further testing indicated a bridge delay of ~0.5-0.7 ms per hop for the 3-Radio Mesh Engine employed in these tests.
  

Figure 5: Mesh Topology was reconfigured in less than 2 seconds

Test 3: Rapid Self-Healing of the Mesh. A critical characteristic of high performance mesh networks is their ability to self configure rapidly in the event of node failure. One of the nodes was shut off - the system configured itself in less than 2 seconds. Note: The delay is based on a sampling interval to decide if a node is indeed "down". It is adjustable. The tests were conducted later  with a lower sampling interval. The self healing time was recorded at 0.7 seconds.  
 

 
Figure 6:(left) Test Setup to see Mesh topology changing to support latency/throughput requirements.
Figure  7: (right) How Mesh topology changed to support latency/throughput requirements.

Test 4: Dynamic control of mesh topology. The control layer is designed to automatically change the mesh topology when the signal strength changes between mesh nodes and connectivity performance is adversely affected. To demonstrate the adaptive control over mesh topology a signal attenuator gradually changed the signal strength, thereby forcing the adaptive control layer to take action. The switch from the dotted connection to the solid line connection ( Figure 7) occurred within one second of reaching the thresholds where bandwidth connectivity was adversely affected.
  


Figure 8: Automatic Switching channels to mitigate interference
  
Test 5: Dynamic Interference Avoidance on a segment basis. One advantage of the multiple radio system is that each node operates on a different channel and interference is thus contained - since interference effects are restricted to one segment of the mesh. This is a distinct advantage of the multiple radio system over single radio approaches. This test demonstrated how the system reacts to interference on one channel- the mesh node switches to another backhaul parent, operating on another (interference free) channel, to manage connectivity performance in a proactive manner. An interferer was injected after node 2 in a 4 hop 2 radio Structured mesh network. The signal became so low that node 1 switched to node 3. The time was measured between the time instance the SIR reduced and the channel switching time from node 1 to node 3. A typical screen capture is shown in Fig. 8. The test was repeated three times to determine if the behavior was repeatable and consistent.  
  
Summary:  MD4350 two radio backhaul  nodes provide distinctive advantages over traditional single radio approaches:

1) Minimal degradation of bandwidth or latency over multiple hops.
2) Ability to modify mesh topology to meet dynamic performance requirements of latency/throughput
3) Ability to switch to other channels to mitigate interference effects in one part of the network. 

Web Page MD4000 Products Page. (Overview of our product, Network Management System, Configuration options)
Web Page MD4000 Model Configuration. (Overview of our model configurations and their applications)
Web Page Frequently Asked Questions. (On Antenna selection, GPS, Mobility extensions, Hazard proof enclosures)