Random High Packet Loss

[jeredish]

So I am a Comcast customer in the Greater Seattle, WA area. Recent after a huge windstorm my service has been dropping packets like they are going out of style. I have called them multiple times and had a service technician out to my place...

First and foremost they are all rejects in my opinion. So I have started my own goal to solve my problems.

I first downloaded the PingPlotter Pro program and have been pinging a google server for 4 days now to get data. This is with completely different configurations like not router, through router but wireless off, through router with wireless on, a friends modem to prove my modem did not get hurt during the windstorm power suges. Now the difference is to prove to them that this problem is not with my setup. The think that my router is to close to my modem, which it might be, but it is not enough to cause such random packet losses as the ones I am seeing.

So with the program I am seeing roughly 5%-10% packet loss every 3-5 minutes, 10-30% every 3-8 minutes and 50%+ 5-10 minutes...

http://i130.photobucket.com/albums/p246/jeredish/PacketLose.jpg

The max size of the packets lost is 50% on the lower graph, they are shown in red. The latency is shown in black and the max is 500ms, as you can see it is stable. The upper graph is just the last 10 samples. In total I have 170000 data points, I am showing you the last 6 hours as it is the same for the points prior.

It appears that the test with different network configurations has shown that my router might be increasing the jitter value, if that is possible? But that is the only difference I see in the data, and the increase is not much a few ms if that.

So the error I feel is in one of the hops at there node or switch. How to I convince them of this?

This is cutting into my downloading, gaming, VoIP (skype), etc. and to be honest is frustrating. I also used the same program on my friends computer (10 miles away) and he is getting about a 5-10% packet lose but nothing more.

My modem is a SB5120, my router is a DGL-4300.

Any insight to how I can convince the cable company to fix their problem or how to convince them I am worth troubleshooting their system if noone else is noticing the problem (which might be a case if they are just emailing and web surfing becuase the modem nevers looses sync and the high packet loss is just a spike.)

Any help or insight would be great Thanks

[trogers]

Any insight to how I can convince the cable company to fix their problem or how to convince them I am worth troubleshooting their system if noone else is noticing the problem (which might be a case if they are just emailing and web surfing becuase the modem nevers looses sync and the high packet loss is just a spike.)

Any help or insight would be great Thanks

Your ISP tech may be right. Hop 1 and 2 will be your end, being the router and modem. Your ISP's intermediate routers are at hop 3 and 4. Beyond hop 4 is not your ISP's network.

Hop 5 to 7 belongs to AT&T.

The ping plotter shows some problem at hop 2 but none at hop 3 to 4.

[jeredish]

How are you seeing a problem with Hop 2? From what I posted?

I know that Hops 3 and 4 are my ISP and I know that the next 3 are AT&T.

Plus me changing out my modem and seeing the same results should prove it is not my modem correct? I could see from the Box to my house or the Box to the Street also being messed up aswell with the windstorm we had.

From the Ping plotter I see most packet lose stem from Hops 3 or what I see usually is a 100% packet lose at Hop 4 then it propagates to the other Hops.

Or I see Hop 2 and Hop 3 with the exact same statistics starting the lose. Now I am unclear if the second Hop 2 you see is really the modem, because it has the exact same IP and statistics as Hop 3. I have been plotting the same route on another computer and Hop 2 is just dashed out a Hop with no statistics which I would take to be the modem.

I think the PingPlotter screwed up the route during the save and it has never changed it back, not sure but the second Hope should show no data.

http://i130.photobucket.com/albums/p246/jeredish/RouteTrace.jpg

Here is what a new trace looks like....not sure why the saved one I have will not place Hop 2 accordingly, but the data in what I have shown is good.

In know way is the IP of my modem that or does it have the description of the one I posted.

[trogers]

Just go a normal tracert to www.yahoo.com and post.

[joezpi]

Hi, I am sorry but I do not know how to post my own "post" in the forum or how to compare speed test results with others. Maybe you or someone can help me. I used the speed test and these are my results. Are they good? and can I speed them up? Sorry for being an idiot, I am new at posting etc...Thanks Alot!! Member: Joezpi

These are my last results:

15744 kbps down (~15.74 Mbps, 1922 KB/s) ↓
350 kbps up (~0.35 Mbps, 43 KB/s) ↑
Details:

10240 KB downloaded in 5.328 seconds
5120 KB uploaded in 119.922 seconds
Speed @ 510% of the average for ma.comcast.net
297 times faster than 56k dialup
Tested on: 2007.01.22 23:58 EST
Tested from: gamegeneration.net
Test ID: MDRZEC4OKFP
Browser/OS: Mozilla/4.0 (compatible; MSIE 7.0; Windows NT 5.1; .NET CLR 1.1.4322; .NET CLR 2.0.50727)
IP Address: 24.63.237.176
Provider: ma.comcast.net
Location: Belmont, MA, US

[jeredish]

How is a yahoo trace route better then a google, or different?

[trogers]

How is a yahoo trace route better then a google, or different?

Another member just posted a tracert to Google and hit high pings at similar AT&T intermediate routers, and ended up blaming his ISP.

[jeredish]

Ohhh

But that is a different problem, I am not looking at high pings, I am looking at complete and random packet lose. So the trace route ping data is really irrelevent.

My ping is good and it is really good to the google server. The only reason I am using that one for data is that is the one that Comcast was using when they came the first time and blamed the problem on the possibility of a very very old TV in the system that is causing interference. So there suggestion was to watch for and record either timeouts of games or dropped VoIP calls. As you can see anytime that the packet loss is about about 30% I am either timing out of a game, dropped a Skype Call, losing web, etc. But as I say again if I am not timed out I am running a great ping.

So I should if your theory is correct not have the problem when say I am gaming because it is highly unlikely it is the exact same hops for all the different servers i play on.

Anywho the yahoo trace route is

1, -------------- ,192.168.0.1
2,,-
3,GE-1-1-ur01.bellevue.wa.seattle.comcast.net,68.86.98.17
4,te-9-3-ar01.burien.wa.seattle.comcast.net,68.86.96.66
5, -------------- ,12.117.243.13
6, -------------- ,12.127.6.50
7, -------------- ,12.127.6.57
8,so-3-2-0.gar1.Seattle1.Level3.net,4.68.127.109
9,ae-1-53.mp1.Seattle1.Level3.net,4.68.105.65
10,as-0-0.bbr2.SanJose1.Level3.net,64.159.0.218
11,ae-13-51.car3.SanJose1.Level3.net,4.68.123.13
12, -------------- ,4.71.112.14
13,g-1-0-0-p171.msr2.sp1.yahoo.com,216.115.107.87
14,UNKNOWN-209-131-32-23.yahoo.com,209.131.32.23
15,f1.www.vip.sp1.yahoo.com,209.131.36.158

[trogers]

You can use the nitro test to check for excessive packet loss at your end. If there isn't any at your end, but singal routes to various other sites do show, then locate the intermediate routes that causes them:

http://nitro.ucsc.edu/

Results are under the 'Statistics' button.

[jeredish]

Here are the results

EB100 Enabled Statistics:
Checking for Middleboxes . . . . . . . . . . . . . . . . . . Done
checking for firewalls . . . . . . . . . . . . . . . . . . . Done
running 10s outbound test (client-to-server [C2S]) . . . . . 715.0kb/s
running 10s inbound test (server-to-client [S2C]) . . . . . . 9.94Mb/s

------ Client System Details ------
OS data: Name = Windows XP, Architecture = x86, Version = 5.1
Java data: Vendor = Sun Microsystems Inc., Version = 1.5.0_10

------ Web100 Detailed Analysis ------
Cable modem/DSL/T1 link found.
Link set to Full Duplex mode
Information: throughput is limited by other network traffic.
Good network cable(s) found
Normal duplex operation found.

Web100 reports the Round trip time = 40.08 msec; the Packet size = 1460 Bytes; and
There were 59 packets retransmitted, 165 duplicate acks received, and 193 SACK blocks received
The connection was idle 0 seconds (0%) of the time
C2S throughput test: Packet queuing detected: 0.94%
This connection is network limited 99.83% of the time.

Web100 reports TCP negotiated the optional Performance Settings to:
RFC 2018 Selective Acknowledgment: ON
RFC 896 Nagle Algorithm: ON
RFC 3168 Explicit Congestion Notification: OFF
RFC 1323 Time Stamping: OFF
RFC 1323 Window Scaling: ON

Server 'nitro.ucsc.edu' is not behind a firewall. [Connection to the ephemeral port was successful]
Client is probably behind a firewall. [Connection to the ephemeral port failed]
Packet size is preserved End-to-End
Server IP addresses are preserved End-to-End

[trogers]

Here are the results

EB100 Enabled Statistics:
Checking for Middleboxes . . . . . . . . . . . . . . . . . . Done
checking for firewalls . . . . . . . . . . . . . . . . . . . Done
running 10s outbound test (client-to-server [C2S]) . . . . . 715.0kb/s
running 10s inbound test (server-to-client [S2C]) . . . . . . 9.94Mb/s

------ Client System Details ------
OS data: Name = Windows XP, Architecture = x86, Version = 5.1
Java data: Vendor = Sun Microsystems Inc., Version = 1.5.0_10

------ Web100 Detailed Analysis ------
Cable modem/DSL/T1 link found.
Link set to Full Duplex mode
Information: throughput is limited by other network traffic.
Good network cable(s) found
Normal duplex operation found.

Web100 reports the Round trip time = 40.08 msec; the Packet size = 1460 Bytes; and
There were 59 packets retransmitted, 165 duplicate acks received, and 193 SACK blocks received
The connection was idle 0 seconds (0%) of the time
C2S throughput test: Packet queuing detected: 0.94%
This connection is network limited 99.83% of the time.

Web100 reports TCP negotiated the optional Performance Settings to:
RFC 2018 Selective Acknowledgment: ON
RFC 896 Nagle Algorithm: ON
RFC 3168 Explicit Congestion Notification: OFF
RFC 1323 Time Stamping: OFF
RFC 1323 Window Scaling: ON

Server 'nitro.ucsc.edu' is not behind a firewall. [Connection to the ephemeral port was successful]
Client is probably behind a firewall. [Connection to the ephemeral port failed]
Packet size is preserved End-to-End
Server IP addresses are preserved End-to-End

Test does not show any packet losses at your end. Your download speed is good at almost 10 mbps.

Test shows quite a bit of packet retransmissions, meaning packet losses along the signal route.

The next thing you will need to do is to find out at which intermediate routers the losses are occurring. They may be those of your ISP, which you can make a complain. Or they may be those of some others.

Note though that there are 3 members, including yourself who had posted problems in signal quality in the last few hours and the common denominator are intermediate routers belonging to AT&T.

[jeredish]

Note though that there are 3 members, including yourself who had posted problems in signal quality in the last few hours and the common denominator are intermediate routers belonging to AT&T.

Yhea I understand people have been posting signal problems from the AT&T routers, but I have been experiencing this for two weeks now. Again ever since the windstorm we had here in WA. Now also I am see the dropped packets starting at Hop3 and Hop4 and then propegating through.

This was also noted in a call with my ISP where he pinged my computer at the right time to hit on of the random high packet lose moments and asked "did you just diconnect your modem?", which I had not. Onw would think that if the singal loss was at the AT&T routers I would see the packet loss start there and propegate through from there to the google server, but I am not, nor have I ever in the past 4 days of data collection now.

I have a so called "technician" coming out today, but I will not let him know that there might e a AT&T issue, which I dont think there is to the point I am having, but also because I dont want to hear an excuss about not being able to fix my problem.

Thanks for for the help.