High percentage of lost packets - TCP, ICMP - mtr - Complain to ISP?
Problem
I'm having high packet loss, according to mtr
, when sending packets over the Internet. Should I complain to my ISP?
Story
I am reading the OReilly Linux Networking Cookbook
and the chapter Using traceroute, tcptraceroute, and mtr to
Pinpoint Network Problems
drew my attention. Pinging a host like Google over the Internet from my ISP gives me record delays of 1200ms and higher (not only since today; since long time), so I thought I'd do no worse analyzing the way of the packets with mtr
.
Mtr is a network diagnostic tool that combines ping and traceroute into one program.
The excerpt and, at the same time, the reason for this question thread is:
If any of these consistently get hung up at the same router, or if mtr consistently shows greater than 5 percent packet losses and long transit times on the same router, then it’s safe to say that particular router has a problem. If it’s a router that you con- trol, then for gosh sakes fix it. If it isn’t, use dig or whois to find out who it belongs to, and nicely report the trouble to them.
Issue
See the mtr --report www.google.com
output yourself: (In total 12 tests, 1 test every 5 minutes; this is the report which represents the reliable 'average')
HOST: km Loss% Snt Last Avg Best Wrst StDev
1. 192.168.0.1 0.0% 10 1.2 3.7 1.2 6.3 1.8
2. 10.150.144.145 10.0% 10 89.1 77.3 58.7 90.4 11.1
3. 172.16.251.1 50.0% 10 52.2 62.1 52.2 70.3 8.8
4. 172.16.250.54 60.0% 10 74.9 87.5 74.9 100.4 12.1
5. 172.16.250.251 40.0% 10 68.6 75.4 52.4 113.8 24.2
6. 200.85.47.2 10.0% 10 109.6 110.6 80.6 146.2 21.1
7. 201.217.4.113 0.0% 10 103.6 87.3 64.4 103.7 12.2
8. 201.217.0.9 0.0% 10 229.0 102.6 46.7 229.0 48.1
9. 201.217.0.3 0.0% 10 78.8 88.1 53.9 128.8 23.8
10. So2-3-2-0-grtbueba2.red.tele 0.0% 10 134.1 129.2 71.3 176.6 29.2
11. Xe4-1-3-0-grtmiabr7.red.tele 0.0% 10 257.3 255.1 221.0 291.6 21.1
12. Xe2-0-2-0-grtmiana3.red.tele 0.0% 10 290.4 267.0 213.2 319.1 31.0
13. Xe2-0-2-0-grtmiana3.red.tele 0.0% 10 300.0 250.8 217.3 312.7 34.6
14. GOOGLE-xe-5-0-0-0-grtmiana3. 10.0% 10 249.8 256.9 206.7 324.0 34.6
15. 209.85.254.252 0.0% 10 254.3 253.8 217.1 283.1 23.4
16. 209.85.254.252 10.0% 10 301.2 280.6 252.1 319.7 21.6
17. 72.14.236.200 10.0% 10 273.4 278.4 238.4 311.0 25.0
18. 216.239.49.145 20.0% 10 291.0 276.3 240.4 293.5 19.1
19. 72.14.232.25 10.0% 10 297.9 286.3 242.4 337.1 30.0
20. yo-in-f105.1e100.net 70.0% 10 300.7 304.7 280.3 333.0 26.6
You see immediately that hosts 3-5 are experiencing a very high packet loss far over 5%. Doing a whois database query shows me that those are name-servers (please correct me if I'm wrong).
Questions
- What should I tell to my ISP? How to describe the problem..?
- What kind of research can I do in addition to facilitate troubleshooting? *1
- Any suggestions?
*1Those guys from technical supports aren't always understanding or I can't express my problem clearly enough (Sometimes they're just idiots without doubt)
Many routers are typically programmed to give lower priority to ICMP packets so they aren't "wasting" processing power over "real" traffic. Just because you see a hop with high loss doesn't mean it's slowing down "real" traffic; it may only be throwing away ICMP. That's not necessarily good because it might mean the router is too busy, but it's not guaranteed.
The router may also be programmed to limit the number of responses it sends to ICMP packets in an effort to mitigate DoS attacks.
It can be that the error is inside your network.
Which one is your internet router/gateway ?
Chances are that
3. 172.16.251.1 50.0% 10 52.2 62.1 52.2 70.3 8.8
4. 172.16.250.54 60.0% 10 74.9 87.5 74.9 100.4 12.1
5. 172.16.250.251 40.0% 10 68.6 75.4 52.4 113.8 24.2
are inside your own network.