I have a Media Temple server running http://www.popsapp.com which I am having latency issues with.
If I run
ab -n 100 -c 10 http://www.popsapp.com/
from my local machine I get very bad stats e.g.:
Connection Times (ms)
min mean[+/-sd] median max
Connect: 179 3375 2185.4 2837 12525
Processing: 0 505 693.3 229 4564
Waiting: 0 50 115.4 0 415
Total: 964 3880 2094.5 3159 12608
Whereas if I run it from a rackspace server I have I get this:
Connection Times (ms)
min mean[+/-sd] median max
Connect: 75 76 3.3 75 84
Processing: 235 339 81.4 315 579
Waiting: 159 249 61.7 234 411
Total: 311 415 82.0 390 663
To me this looks like intermediate network issues, but I wouldn't have thought it could be this bad!
Any ideas how I can improve it?
Here's the trace route
traceroute to www.popsapp.com (216.70.105.183), 64 hops max, 52 byte packets
1 192.168.2.1 (192.168.2.1) 3.738 ms 0.953 ms 1.418 ms
2 host-92-22-112-1.as13285.net (92.22.112.1) 27.409 ms 97.093 ms 78.858 ms
3 host-78-151-225-141.static.as13285.net (78.151.225.141) 61.830 ms 170.484 ms 113.288 ms
4 host-78-151-225-80.static.as13285.net (78.151.225.80) 101.513 ms
host-78-151-225-22.static.as13285.net (78.151.225.22) 64.718 ms 47.309 ms
5 xe-11-1-0-rt001.sov.as13285.net (62.24.240.14) 98.381 ms 114.424 ms
xe-11-1-0-rt001.the.as13285.net (62.24.240.6) 96.592 ms
6 host-78-144-1-59.as13285.net (78.144.1.59) 36.799 ms
host-78-144-1-63.as13285.net (78.144.1.63) 178.426 ms
host-78-144-1-61.as13285.net (78.144.1.61) 85.516 ms
7 xe-10-0-0-scr010.thn.as13285.net (78.144.0.224) 88.158 ms
host-78-144-0-207.as13285.net (78.144.0.207) 35.132 ms
host-78-144-0-153.as13285.net (78.144.0.153) 121.464 ms
8 limelight-pp-thn.as13285.net (78.144.3.6) 46.987 ms
limelight-pp-sov.as13285.net (78.144.5.18) 108.025 ms 40.169 ms
9 tge11-1.fr4.lga.llnw.net (69.28.172.149) 109.603 ms
ve6.fr4.lon.llnw.net (68.142.88.221) 121.681 ms 38.609 ms
10 tge11-1.fr4.lga.llnw.net (69.28.172.149) 111.981 ms 113.744 ms 111.711 ms
11 tge8-2.fr4.iad.llnw.net (69.28.189.34) 117.102 ms
ve5.fr4.iad.llnw.net (69.28.171.214) 184.372 ms 146.178 ms
12 cr02-1-1.iad1.net2ez.com (65.97.48.254) 182.880 ms
net2ez.tge2-2.fr4.iad.llnw.net (69.28.156.170) 150.489 ms 121.862 ms
13 65.97.50.26 (65.97.50.26) 184.620 ms
cr02-1-1.iad1.net2ez.com (65.97.48.254) 156.136 ms 131.963 ms
14 65.97.50.26 (65.97.50.26) 124.899 ms 126.537 ms 123.322 ms
15 e1.4.as02.iad01.mtsvc.net (70.32.64.246) 134.647 ms 186.307 ms 211.059 ms
16 popsapp.com (216.70.105.183) 118.876 ms 113.189 ms
vzx258.mediatemple.net (216.70.104.17) 131.012 ms
Looks to me like there is significant delay across the limelight network. This would explain why the traceroute via my rackspace server doesn't suffer from the same delay as they will be using their own trunk.