01-13-2014 06:53:56 AM
In the last week or so I have been experiencing significantly slower Internet speeds from my AT&T UVerse Max Plus service (which should deliver approximately 18 Mbps of bandwidth downstream). I can correct the problem by turning off the gateway, waiting for several seconds, then turning the gateway back on. Once the gateway has restored Internet service I retest using www.speedtest.net. The speed measured is in the neighborhood of 18 Mbps and everything seems ok.
Then, after one or two hours, I will re-test from www.speedtest.com and again see much slower meaurements. All three categories reported by www.speedtest.net will be degraded often with much slower measurements (ping response time, download speed, upload speed).
The cycle repeats. I reset the gateway, once again the speed is ok (download speed once again is in the neighorbhood of 18 Mbps). Then some time later I will retest and once again seed a significant drop in performance. Note that testing is done on a computer connected to the gateway over Ethernet (thus I have eliminated the local wireless factor).
I called up AT&T U-Verse technical support and they dispatched a technician. He did basic line testing inside the house and also immediately outside the house. The numbers he measured were ok. The technician did mention to check for cloud synchronization services such as dropbox.com, but I do not run background services/processes for dropbox, skydrive or similar.
What would be the next step in trouble shooting? I saw one posting that mentions UV Realtime. Are there other suggestions I can follow to trouble shoot?
01-13-2014 07:00:02 AM
One thing I would be concerned with if you know of no upload services is malware doing that kind of stuff "for you." Or you could have a neighbor borrowing your Internet connection via Wi-Fi.
UVrealtime could show you any line problems you're having, as well as show traffic passing through the RG. If there is traffic that you don't know about it should show up there.
01-13-2014 10:20:08 AM
I wonder if this behavior could be an indication of a failing gateway? I would be interested to hear if others have experienced a similar issue, had their RG device replaced by AT&T, and then confirmed that the problem was corrected.
I'll check for malware but I am doubtful because I run with clean systems. Also, I use Norton and similar programs to scan/protect against malware. (In fact, my Web browsing activity is fairly limited. Mostly I use my Internet to connect to my work systems via VPN.)
I did find this link http://www.combofix.org/malwarea-possible-causes-o
01-13-2014 07:50:13 PM
That's not a bad list.
01-14-2014 06:44:02 AM
I have new culprit that I am starting to investigate. This would be apps running on mobile devices. In particular gaming apps that frequently (or constantly?) communicate to Internet servers.
My house is made up of the Internet of Things and includes two iPad devices, one iPhone and one Android tablet. Games in particular will continue to run in the background (once launched) and utilize bandwidth sending RESTful messages (or something similar) to Internet servers. Often my childeren run these apps and later move on to something else without stoppping them.
Does it seem plausible that this would be eating into the bandwidth available for broadband? (The mobile apps connect to a linksys WiFi router which is in turn connected the U-verse gateway.) Have others experienced similar issues? Sending RESTful Web service messages from a mobile app normally does not use so much bandwidth, but possibly over time this contributes to problems with the U-verse gateway device?
01-14-2014 08:18:20 AM
Most game apps do not put a lot of load on the network; they send information that causes content to be drawn locally, though sometimes content needs to be sent too.
Using UVrealtime, you can see how much wireless traffic is going on and if significant, you could turn off devices until it goes away. If you use your own wireless access point/router you could use wireshark or some other promiscuous mode utility to see what's going on.
01-19-2014 07:17:22 PM
Exactly the same here as original poster. I can barely use this forum at the moment and I am about to fire ATT Uverse. The service guys have been out a few times, once replaced all the coax cable connectors, one found some bad wiring in the box out side. and then a line man came out and found a pinched wire in the main box down the street. They have replace the gateway twice. They leave and its working, but the problem returns.
So now I have been trouble shooting myself. I disconnected everything from the model to check for interference. Restart and it works fine for a while. TV works fine. I pay for 18Mbps and lucky to get 300 kbps and the thruput is very irractic, with 300ms ping, so I don't think its other devices stealing bandwidth.
Still researching for a solution.......
01-19-2014 07:53:37 PM
Seadoodude1, have you tried using UVrealtime to see what your line looks like?
01-19-2014 08:52:46 PM - edited 01-19-2014 09:19:57 PM
after reading this post, I downloaded it and it's running but I really don't know what to look for.
And then I got to thinking about how it works for a while after restarting the RG, then the problem returns. Could it be caused by CrashPlan, which backs-up our computers to the cloud?
I just paused the service and so far the speedtest is looking really good, 20.65Mbps, 1.43Mbps.
CrashPlan uses a Java process to upload a backup of files that have changed. It runs periodically, which would explain intermittent issues with the internet speed. But I don't understand why it's uploading would tank the Uverse ping and download speeds.
Does this sound plausable, one program uploading causing the everything else to tank?
I guess it does. I have run a few tests, running two speedtests at the same time. If both are downloading, they share bandwidth, this is understandable. However, if one is uploading, the second one tanks, long ping and really slow and irratic downloads. This I don't understand why? Like at work, we have a lot of people sharing bandwidth and I have never seen any issues like this. I think it's 25+ Mbps down and up service with Comcast, maybe cable.
So it appears that Uverse is asyncronous (or whatever makes it only work for uploads or downloads one at a time, but not both simultaneously). In other words, when you are uploading, the downloads are really really bad. Is this true, and the way it's supposed to work?
01-19-2014 09:18:30 PM
Downloading, request a packet, verify packet received, request next packet....each request is upload.
As upload is on most Uverse packages is only 1M or 1.5M...any large upload will delay the ability of download requests.
Think of trying to enter a busy road with a funeral procession going down highway...cannot access till uninterrupted traffic has passed.
To control this type of traffic need to purchase a router that has QoS capabilities, limit specific programs to 2/3 of available upload allowing other programs access. Will need to place new router behind Uverse RG in dmz mode.
Or schedule backups to occur at 3am when hopefully internet not being used.
01-19-2014 09:37:41 PM
Thanks for the advice, on the QoS, I will look into that, to reserve some of the limited upload bandwidth for download traffic.
Finally, problem identified!!!!!! Not completely solved, but solution understood.
01-20-2014 06:27:19 AM
Downstream bandwidth and upstream bandwidth are related, so upgrading your downstream does (usually) upgrade your upstream:
24 Mbps down => 3 Mbps up
18 Mbps down => 1.5 Mbps up
12 Mbps down => 1.5 Mbps up
6 Mbps down => 1.0 Mbps up
Your backup program may also have the ability to throttle its bandwidth utilization. I see on CrashPlan's web site (https://support.code42.com/CrashPlan/Latest/Config