Table 1
Numeric Results Presented Graphically in Figures 7–9*
| % Bandwidth | 0 | 69 | 88 | 94 |
| Average run times (sec) | ||||
| FIFO | 28.1 | 77.3 | 132.4 | 305.4 |
| PRIORQ | 34.4 | 80.7 | 125.0 | 158.4 |
| WRED | 35.6 | 76.2 | 103.5 | 141.0 |
| WFQ | 32.9 | 75.1 | 117.3 | 149.3 |
| Standard deviation of the run times (sec) | ||||
| FIFO | 0.6 | 8.6 | 8.8 | 88.5 |
| PRIORQ | 1.0 | 3.8 | 6.2 | 8.8 |
| WRED | 1.4 | 5.1 | 2.7 | 5.7 |
| WFQ | 1.2 | 1.8 | 5.5 | 7.0 |
| Run times for WFQ with shaping (sec) | ||||
| WFQ no shape | 32.9 | 75.1 | 117.3 | 149.3 |
| WFQ shape 4000 | 33.4 | 63.9 | 73.9 | 82.4 |
| WFQ shape 2000 | 32.7 | 40.8 | 42.6 | 43.5 |
| WFQ shape 200 | 32.9 | 33.1 | 33.2 | 33.1 |
-
↵* Each reported time represents the average of 10 runs. For each set of runs, the table indicates the % of the bandwidth consumed by competing network traffic as described in the text.
