LEFT | RIGHT |
(no file at all) | |
1 .. include:: replace.txt | 1 .. include:: replace.txt |
2 .. highlight:: cpp | 2 .. highlight:: cpp |
3 | 3 |
4 RED queue disc | 4 RED queue disc |
5 --------------------- | 5 --------------------- |
6 | 6 |
7 Model Description | 7 Model Description |
8 ***************** | 8 ***************** |
9 | 9 |
10 Random Early Detection (RED) is a queue discipline that aims to provide | 10 Random Early Detection (RED) is a queue discipline that aims to provide |
(...skipping 129 matching lines...) Expand 10 before | Expand all | Expand 10 after Loading... |
140 | 140 |
141 ARED queue examples can be found at: | 141 ARED queue examples can be found at: |
142 ``src/traffic-control/examples/adaptive-red-tests.cc`` and· | 142 ``src/traffic-control/examples/adaptive-red-tests.cc`` and· |
143 ``src/traffic-control/examples/red-vs-ared.cc`` | 143 ``src/traffic-control/examples/red-vs-ared.cc`` |
144 | 144 |
145 Validation | 145 Validation |
146 ********** | 146 ********** |
147 | 147 |
148 The RED model has been validated and the report is currently stored | 148 The RED model has been validated and the report is currently stored |
149 at: https://github.com/downloads/talau/ns-3-tcp-red/report-red-ns3.pdf· | 149 at: https://github.com/downloads/talau/ns-3-tcp-red/report-red-ns3.pdf· |
| 150 |
| 151 Validation of ECN operation |
| 152 =========================== |
| 153 RED Tests |
| 154 ========= |
| 155 |
| 156 Test 1 (without ECN) |
| 157 RED stats from Node 2 queue |
| 158 142 drops due prob mark |
| 159 0 marks due prob mark |
| 160 224 drops due hard mark |
| 161 0 drops due hard mark |
| 162 0 drops due queue full |
| 163 |
| 164 |
| 165 Test 2 (same as Test 1 with ECN) |
| 166 RED stats from Node 2 queue· |
| 167 65 drops due prob mark |
| 168 71 marks due prob mark |
| 169 293 drops due hard mark |
| 170 0 drops due hard mark |
| 171 0 drops due queue full |
| 172 |
| 173 When ECN is enabled, there are marks instead of drops due to probability. |
| 174 However, in test 2, 65 packet gets dropped despite ECN being enabled because the
y |
| 175 are non-ECT packets. This can be verified by connecting to drop trace sources |
| 176 and observing the IP header of dropped packets. |
| 177 |
| 178 ARED Tests |
| 179 ========== |
| 180 |
| 181 Test 1 (without ECN) |
| 182 ARED stats from Node 2 queue |
| 183 3 drops due to prob mark |
| 184 0 marks due to prob mark |
| 185 0 drops due to hard mark |
| 186 0 marks due to hard mark |
| 187 251 drops due to queue full |
| 188 ·· |
| 189 Test 3 (same as Test 1 with ECN) |
| 190 ARED stats from Node 2 queue |
| 191 1 drops due to prob mark |
| 192 2 marks due to prob mark |
| 193 0 drops due to hard mark |
| 194 0 marks due to hard mark |
| 195 220 drops due to queue full |
| 196 |
| 197 When ECN is enabled, there are marks instead of drops due to probability. |
| 198 However, in test 3, 1 packet gets dropped despite ECN being enabled because it |
| 199 is a non-ECT packet. This can be verified by connecting to drop trace sources |
| 200 and observing the IP header of dropped packets. |
| 201 |
| 202 Test 2 (without ECN) |
| 203 ARED stats from Node 2 queue |
| 204 29 drops due to prob mark |
| 205 0 marks due to prob mark |
| 206 0 drops due to hard mark |
| 207 0 marks due to hard mark |
| 208 133 drops due to queue full |
| 209 |
| 210 Test 4 (same as Test 2 with ECN) |
| 211 ARED stats from Node 2 queue |
| 212 2 drops due to prob mark |
| 213 25 marks due to prob mark |
| 214 0 drops due to hard mark |
| 215 0 marks due to hard mark |
| 216 139 drops due to queue full |
| 217 |
| 218 When ECN is enabled, there are marks instead of drops due to probability. |
| 219 However, in test 4, 2 packet get dropped despite ECN being enabled because they |
| 220 are non-ECT packets. This can be verified by connecting to drop trace sources |
| 221 and observing the IP header of dropped packets. |
| 222 |
| 223 Test 10 (without ECN) |
| 224 ARED stats from Node 2 queue |
| 225 121 drops due to prob mark |
| 226 0 marks due to prob mark |
| 227 129 drops due to hard mark |
| 228 0 marks due to hard mark |
| 229 0 drops due to queue full |
| 230 |
| 231 |
| 232 Test 11 (same as Test 10 with ECN) |
| 233 ARED stats from Node 2 queue |
| 234 0 drops due to prob mark |
| 235 71 marks due to prob mark |
| 236 200 drops due to hard mark |
| 237 0 marks due to hard mark |
| 238 0 drops due to queue full |
| 239 |
| 240 When ECN is enabled, there are marks instead of drops due to probability. There· |
| 241 are drops due to hard mark rather than marks because UseHardDrop parameter is· |
| 242 turned on. If the UseHardDrop parameter is turned off, following stats can be· |
| 243 seen from node 2 for Test 11. |
| 244 |
| 245 0 drops due to prob mark |
| 246 20 marks due to prob mark |
| 247 0 drops due to hard mark |
| 248 660 marks due to hard mark |
| 249 0 drops due to queue full |
| 250 |
| 251 |
| 252 |
LEFT | RIGHT |