| Flow Type | Priority | Throughput at Tx( Mbps) | Throughput at Rx (Mbps) |
| AF class4 | 2 | 1.09 | 0.906 |
| AF class3 | 3 | 1.09 | 0.861 |
| AF class2 | 4 | 1.09 | 0.806 |
| AF class1 | 5 | 1.09 | 0.776 |
Note that lesser the numerical value of priority, more the priority for the
class, i.e AF class4 has a higher priority than AF class1. We can observe a
clear difference in the throughput among the different AF classes.
Also to see the coexistance of AF, EF and BE, a few tests were performed and their results are summarised below. The BW and priority allocations for the different classes are as follows,
| Flow Type | Priority | BW allocation (Mbps) |
| EF PHB | 1 | 2.0 |
| AF PHB | 2 | 8.0 |
| AF class1 | 5 | 1.5 |
| AF class2 | 4 | 1.5 |
| AF class3 | 3 | 2.0 |
| AF class4 | 2 | 2.5 |
| BE | 7 | 0.5 |
Data for the various classes were pumped in through ttcp, ttcp was
modified to set the TOS byte in the packets. It can be found that the traffic
flow through the appropriate classes, this can be observed with the show
in tc.
The results:
| Flow Type | Throughput at Tx (Mbps) | Throughput at Rx (Mbps) |
| EF flow1 | 1.014 | 1.011 |
| EF flow2 | 1.014 | 0.989 |
| AF class4 | 2.420 | 2.378 |
| BE | 0.979 | 0.468 |
Observe that the two EF flows together get almost close to what was allocated
for EF and BE seem to be affected in the presence of other flows.
Another set of results
| Flow Type | Throughput at Tx (Mbps) | Throughput at Rx (Mbps) |
| EF flow1 | 1.014 | 1.014 |
| EF flow2 | 1.014 | 0.994 |
| AF class4 | 2.420 | 2.379 |
| AF class2 | 1.405 | 1.368 |
| BE | 0.979 | 0.489 |