Re: hub-spoke topology and response time

Hm... why would broadcast traffic consume more CPU cycles locally? From
legacy networking (90s) we might remember that with really old NICs
having next to zero brains each broadcast packet received would trigger
a CPU interrupt and request attention. So it might be CPU-related on
remote side... At least I see some peaks in CPU IO interrupts on remote
router - see below, (approximately) at the same time with peaks in
reply time on pinging one.

And BTW today I had a different pattern on pinging side:

R1#ping ip 2.2.2.255 re 300

Type escape sequence to abort.
Sending 300, 100-byte ICMP Echos to 2.2.2.255, timeout is 2 seconds:

Reply to request 0 from 2.2.2.2, 132 ms
Reply to request 1 from 2.2.2.2, 32 ms
Reply to request 2 from 2.2.2.2, 33 ms
Reply to request 3 from 2.2.2.2, 36 ms
Reply to request 4 from 2.2.2.2, 36 ms
Reply to request 5 from 2.2.2.2, 137 ms
Reply to request 6 from 2.2.2.2, 40 ms
Reply to request 7 from 2.2.2.2, 40 ms
Reply to request 8 from 2.2.2.2, 40 ms
Reply to request 9 from 2.2.2.2, 41 ms
Reply to request 10 from 2.2.2.2, 144 ms
Reply to request 11 from 2.2.2.2, 44 ms
Reply to request 12 from 2.2.2.2, 44 ms
Reply to request 13 from 2.2.2.2, 48 ms
Reply to request 14 from 2.2.2.2, 48 ms
Reply to request 15 from 2.2.2.2, 148 ms
Reply to request 16 from 2.2.2.2, 49 ms
Reply to request 17 from 2.2.2.2, 52 ms
Reply to request 18 from 2.2.2.2, 52 ms
Reply to request 19 from 2.2.2.2, 52 ms
Reply to request 20 from 2.2.2.2, 156 ms
Reply to request 21 from 2.2.2.2, 56 ms
Reply to request 22 from 2.2.2.2, 56 ms
Reply to request 23 from 2.2.2.2, 57 ms
Reply to request 24 from 2.2.2.2, 60 ms
Reply to request 25 from 2.2.2.2, 160 ms
Reply to request 26 from 2.2.2.2, 60 ms
Reply to request 27 from 2.2.2.2, 64 ms
Reply to request 28 from 2.2.2.2, 64 ms
Reply to request 29 from 2.2.2.2, 64 ms
Reply to request 30 from 2.2.2.2, 165 ms
Reply to request 31 from 2.2.2.2, 68 ms
Reply to request 32 from 2.2.2.2, 68 ms
Reply to request 33 from 2.2.2.2, 68 ms
Reply to request 34 from 2.2.2.2, 72 ms
Reply to request 35 from 2.2.2.2, 172 ms
Reply to request 36 from 2.2.2.2, 72 ms
Reply to request 37 from 2.2.2.2, 72 ms
Reply to request 38 from 2.2.2.2, 76 ms
Reply to request 39 from 2.2.2.2, 76 ms
Reply to request 40 from 2.2.2.2, 176 ms
Reply to request 41 from 2.2.2.2, 77 ms
Reply to request 42 from 2.2.2.2, 80 ms
Reply to request 43 from 2.2.2.2, 80 ms
Reply to request 44 from 2.2.2.2, 80 ms
Reply to request 45 from 2.2.2.2, 184 ms
Reply to request 46 from 2.2.2.2, 84 ms
Reply to request 47 from 2.2.2.2, 84 ms
Reply to request 48 from 2.2.2.2, 93 ms
Reply to request 49 from 2.2.2.2, 88 ms
Reply to request 50 from 2.2.2.2, 188 ms
Reply to request 51 from 2.2.2.2, 88 ms
Reply to request 52 from 2.2.2.2, 92 ms
....
Reply to request 210 from 2.2.2.2, 345 ms
Reply to request 211 from 2.2.2.2, 48 ms
Reply to request 212 from 2.2.2.2, 48 ms
Reply to request 213 from 2.2.2.2, 48 ms
Reply to request 214 from 2.2.2.2, 48 ms
Reply to request 215 from 2.2.2.2, 352 ms
Reply to request 216 from 2.2.2.2, 52 ms
Reply to request 217 from 2.2.2.2, 52 ms
Reply to request 218 from 2.2.2.2, 52 ms
Reply to request 219 from 2.2.2.2, 52 ms
Reply to request 220 from 2.2.2.2, 353 ms
Reply to request 221 from 2.2.2.2, 52 ms
Reply to request 222 from 2.2.2.2, 56 ms
Reply to request 223 from 2.2.2.2, 56 ms
Reply to request 224 from 2.2.2.2, 56 ms
Reply to request 225 from 2.2.2.2, 357 ms
Reply to request 226 from 2.2.2.2, 60 ms
Reply to request 227 from 2.2.2.2, 60 ms
Reply to request 228 from 2.2.2.2, 60 ms
Reply to request 229 from 2.2.2.2, 64 ms
Reply to request 230 from 2.2.2.2, 364 ms
Reply to request 231 from 2.2.2.2, 64 ms
Reply to request 232 from 2.2.2.2, 65 ms
Reply to request 233 from 2.2.2.2, 68 ms
Reply to request 234 from 2.2.2.2, 68 ms
Reply to request 235 from 2.2.2.2, 369 ms
Reply to request 236 from 2.2.2.2, 176 ms
Reply to request 237 from 2.2.2.2, 72 ms
Reply to request 238 from 2.2.2.2, 72 ms
Reply to request 239 from 2.2.2.2, 73 ms
Reply to request 240 from 2.2.2.2, 376 ms
Reply to request 241 from 2.2.2.2, 76 ms
Reply to request 242 from 2.2.2.2, 165 ms
Reply to request 243 from 2.2.2.2, 80 ms
Reply to request 244 from 2.2.2.2, 80 ms
Reply to request 245 from 2.2.2.2, 381 ms
Reply to request 246 from 2.2.2.2, 81 ms
Reply to request 247 from 2.2.2.2, 84 ms
Reply to request 248 from 2.2.2.2, 84 ms
Reply to request 249 from 2.2.2.2, 84 ms
Reply to request 250 from 2.2.2.2, 388 ms
Reply to request 251 from 2.2.2.2, 88 ms
Reply to request 252 from 2.2.2.2, 88 ms
Reply to request 253 from 2.2.2.2, 89 ms
Reply to request 254 from 2.2.2.2, 92 ms
Reply to request 255 from 2.2.2.2, 393 ms
Reply to request 256 from 2.2.2.2, 92 ms
Reply to request 257 from 2.2.2.2, 96 ms
Reply to request 258 from 2.2.2.2, 96 ms
Reply to request 259 from 2.2.2.2, 96 ms
Reply to request 260 from 2.2.2.2, 397 ms
Reply to request 261 from 2.2.2.2, 100 ms
Reply to request 262 from 2.2.2.2, 100 ms
Reply to request 263 from 2.2.2.2, 100 ms
Reply to request 264 from 2.2.2.2, 101 ms
Reply to request 265 from 2.2.2.2, 404 ms
Reply to request 266 from 2.2.2.2, 4 ms
Reply to request 267 from 2.2.2.2, 4 ms
Reply to request 268 from 2.2.2.2, 8 ms
Reply to request 269 from 2.2.2.2, 8 ms
Reply to request 270 from 2.2.2.2, 409 ms
Reply to request 271 from 2.2.2.2, 8 ms
Reply to request 272 from 2.2.2.2, 12 ms
Reply to request 273 from 2.2.2.2, 12 ms
Reply to request 274 from 2.2.2.2, 12 ms
Reply to request 275 from 2.2.2.2, 416 ms
Reply to request 276 from 2.2.2.2, 16 ms
Reply to request 277 from 2.2.2.2, 16 ms
Reply to request 278 from 2.2.2.2, 16 ms
Reply to request 279 from 2.2.2.2, 20 ms

Reply to request 280 from 2.2.2.2, 421 ms
Reply to request 281 from 2.2.2.2, 20 ms
Reply to request 282 from 2.2.2.2, 24 ms
Reply to request 283 from 2.2.2.2, 24 ms
Reply to request 284 from 2.2.2.2, 24 ms
Reply to request 285 from 2.2.2.2, 24 ms
Reply to request 286 from 2.2.2.2, 28 ms

=== Remote side

R2#
R2#sh proc cpu | in five
CPU utilization for five seconds: 0%/0%; one minute: 1%; five minutes: 1%
R2#sh proc cpu | in five
CPU utilization for five seconds: 5%/0%; one minute: 1%; five minutes: 1%
R2#sh proc cpu | in five
CPU utilization for five seconds: 5%/0%; one minute: 1%; five minutes: 1%
R2#sh proc cpu | in five
CPU utilization for five seconds: 4%/0%; one minute: 2%; five minutes: 1%
R2#sh proc cpu | in five
CPU utilization for five seconds: 4%/0%; one minute: 2%; five minutes: 1%
R2#sh proc cpu | in five
CPU utilization for five seconds: 0%/0%; one minute: 0%; five minutes: 0%
R2#sh proc cpu | in five
CPU utilization for five seconds: 3%/0%; one minute: 0%; five minutes: 0%
R2#sh proc cpu | in five
CPU utilization for five seconds: 3%/0%; one minute: 0%; five minutes: 0%
R2#sh proc cpu | in five
CPU utilization for five seconds: 3%/0%; one minute: 0%; five minutes: 0%
R2#sh proc cpu | in five
CPU utilization for five seconds: 7%/0%; one minute: 1%; five minutes: 1%
R2#sh proc cpu | in five
CPU utilization for five seconds: 7%/0%; one minute: 1%; five minutes: 1%
R2#sh proc cpu | in five
CPU utilization for five seconds: 7%/0%; one minute: 1%; five minutes: 1%
R2#sh proc cpu | in five
CPU utilization for five seconds: 7%/0%; one minute: 1%; five minutes: 1%
R2#sh proc cpu | in five
CPU utilization for five seconds: 7%/0%; one minute: 1%; five minutes: 1%
R2#sh proc cpu | in five
CPU utilization for five seconds: 4%/0%; one minute: 2%; five minutes: 1%
R2#sh proc cpu | in five
CPU utilization for five seconds: 4%/0%; one minute: 2%; five minutes: 1%
R2#sh proc cpu | in five
CPU utilization for five seconds: 4%/0%; one minute: 2%; five minutes: 1%
R2#sh proc cpu | in five
CPU utilization for five seconds: 7%/0%; one minute: 2%; five minutes: 1%
R2#sh proc cpu | in five
CPU utilization for five seconds: 7%/0%; one minute: 2%; five minutes: 1%
R2#sh proc cpu | in five
CPU utilization for five seconds: 7%/0%; one minute: 2%; five minutes: 1%
R2#sh proc cpu | in five
CPU utilization for five seconds: 7%/0%; one minute: 2%; five minutes: 1%
R2#sh proc cpu | in five
CPU utilization for five seconds: 7%/0%; one minute: 2%; five minutes: 1%
R2#

GAURAV MADAN wrote:
> Well I have a feeling that it depends on CPU process invloved .
> Broadcast will consume more CPU cycles ..
> My observations are :

Blogs and organic groups at http://www.ccie.net
Received on Thu Jul 23 2009 - 21:22:17 ART