dslreports logo
site
 
    All Forums Hot Topics Gallery
spc

spacer




how-to block ads


Search Topic:
uniqs
8
share rss forum feed


Pashune
Caps stifle innovation
Premium
join:2006-04-14
Gautier, MS
Reviews:
·Vonage
·CableOne
reply to navigator

Re: DSL Latency Issues

said by navigator:

We would lose net access randomly for random amounts of time but keep sync. I had that problem escalated to a higher tech who identified some issues and everything was good for a few weeks when I noticed whats occurring now. My ping in gameservers is usually 25-60ms, and when this is happening it goes up to 80-120ms. It never gets any worse and is still playable but I can tell it's lagging.

I thought this subforum was AT&T direct also, just locally. I will post there as I have trouble explaining this to techs on the phone.

Does the latency slowly ease up from 25-60 to 80-120 or is it an on-the-clock thing?

It again, sounds more like a headend problem or somewhere upstream the way you describe it. If your line is maintaining sync, you're getting full speeds, etc. then it's likely a problem somewhere within AT&T's network. The gateway you're connected to doesn't seem to be suffering latency issues but the router ahead of it appears to be having problems. It could be anything from a bad fiber splice to bad or overloaded equipment, etc.

AT&T Direct is a subforum where only you and technicians at AT&T can view the thread you post. They can dispatch a technician, file a trouble ticket or look into a problem more easily and quickly vs. talking to an automated robot through their 1-800 number.

When I still had AT&T the direct forum saved me a lot of hassle when I had big problems. Let us know if you get your problem solved
--
CableOne 10m/1m

navigator

join:2012-10-30
Franklinton, LA
The latency is either good or bad from what I can tell. I played a game around 12-1pm today and the latency was normal and fine. Then around 6pm I went to play a game and can tell it's bad again. I currently have 80ms ping to google. It's weird that it doesn't fluctuate much. It's a solid 80ms ping. I even went to my brothers house(same road) and tested his latency and same thing. I'm almost certain it's only occurring during peak hours (5pm-12am). Sometimes during these peak hours it will go back down for a short period. I suspect the problem is upstream and some route is getting saturated. If it's a line problem then I should be having trouble at all times. I have good latency to that gateway thing but it doesn't to the next router so couldn't the gateway thing still be the issue? Just because I have good latency on my side to it doesn't mean it's okay on the other side. I don't think we have any fiber locally. We may have fiber from the CO to the remote DSLAM I'm connected to. I posted in AT&T direct and waiting for response.

This is my sync signal:

Downstream Upstream
---------- ----------
SNR Margin: 14.20 12.00 dB
Line Attenuation: 26.50 13.00 dB
Output Power: 0.00 8.59 dBm

I also read up on a few things. Like interleaving for example. Could something like that have been turned on? It's weird how when my ping is bad it appears to be exactly double + 10ms what it should be.

Also I said in previous post my old sync speed was 8192 but I meant 8128/512 which is max sync rate.

I only realized in the last few years that there is a direct forum here and this is the first time I've needed to use it. So much relief to talk to someone that actually understands and can help. I will let you guys know if we get it working. Thanks for the responses so far.


DataRiker
Premium
join:2002-05-19
00000
I was trying to see if you're on Interleave mode.

Can you post the whole config page? It should say Mode: Fast Path or Interleave

navigator

join:2012-10-30
Franklinton, LA
Wow, it does say interleave. It used to be fast-path

ADSL Line State: Up
ADSL Startup Attempts: 4
ADSL Modulation: DMT
ADSL Data Path: Interleave
CO vendor: CTNW
Power Management State L0
Datapump Version: DSP 7.5.2.202, HAL 7.5.2.1

Downstream Upstream
---------- ----------
SNR Margin: 14.30 12.00 dB
Line Attenuation: 26.50 13.00 dB
Output Power: 0.00 8.59 dBm
Errored Seconds: 0 0
Loss of Signal: 0 0
Loss of Frame: 0 0
CRC Errors: 0 2
Data Rate: 8096 512
Interleaver Delay 4 1 (milliseconds)
Interleave Depth 32 4
INP 0.000 0.521

atucRevisionNum 1 ATU-C Vendor ID CTNW
TrainedPath Interleave TrainedMode G.DMT
Downstream Upstream
----------------------------------------
Interlv delay 32 4
LowestTone 35 6
HighestTone 245 28
TonesExcited 210 23

NetRate 16192 512
AggrRate 16256 544
TotalRate 8128 736
LineRate 9588 800
ATTNDR 10720 896

ACTATP -1 -1
ACTPSD -364 -345
HLINSC 0
Estimated loop length: 5584 feet (bridge tap)
Framing Mode: 3 - (REDUCED OVERHEAD WITH MERGED SYNC)
TPS-TC Encoding ATM

ATM port status : Up
Rx data rate (bps) : 8096
Tx data rate (bps) : 512

navigator

join:2012-10-30
Franklinton, LA
I found out at least 3 customers are set to interleave now. And my parents are still on fastpath. It was set to interleave back in august when I called in for the previous issue. Again, we would only lose net access, but keep sync. Never had packet loss or sync instability. I'm on copper from me to the remote DSLAM and fiber was laid a few years ago from DSLAM to CO. As far as I know, the line is in good shape. No local hardware has been changed recently. I am on maxsync profile but it only connects at 8096 now. I really want to identify the problem and get back on fastpath. I haven't been responded to in the direct forum yet other than giving my # and basic info on the problem.

I now actually think the problem started a long time ago. I use my connection frequently for work and play. I probably stay on it the most in my area so I notice any blimp in service. I know how to identify when the problem is with me, at&t, or the server I'm connecting to. I understand the net can get congested at times. Over the last few years, very rarely I would lose net access but not sync. It would happen at any time of the day so I know it wasn't because someone was working on it. But it was so rarely that I ignored it. Sometimes it would come right back and others stay off for secs, mins, hours. Then a few months ago back in august it slowly started doing it more and more frequently over the course of a few weeks til I finally called in. Wouldn't this mean something has slowly degraded over time? They keep correcting it in software when it's mostly likely a piece of hardware going bad or something else. And what's up with the on-the-clock latency issues? It's 11am and my latency is fine. Is there peak hours when something is turned on? I'm going to try and let a ping run at 5pm and 6pm to see if it switches over.


Pashune
Caps stifle innovation
Premium
join:2006-04-14
Gautier, MS
Reviews:
·Vonage
·CableOne
said by navigator:

Wouldn't this mean something has slowly degraded over time? They keep correcting it in software when it's mostly likely a piece of hardware going bad or something else. And what's up with the on-the-clock latency issues? It's 11am and my latency is fine. Is there peak hours when something is turned on? I'm going to try and let a ping run at 5pm and 6pm to see if it switches over.

I wouldn't doubt that. Over time as more and more users are added to the network, AT&T may offer higher speeds, etc. and as a result if a certain part of the network isn't upgraded, it will begin to get saturated and things like latency, speed will be affected.

Interleave isn't causing this added latency however. Interleave is basically a hardware noise correction method (of some sort) that delays bits of data, analyzes them for noise, and corrects them if needed. If it were affecting your connection your second hop would be affected as well as the rest. At 15 ms you're probably on the 1st noise profile but you'd only lose 4-8 ms of latency max if you were switched back to fast path.

Typically the reason you're switched to interleave is noise or your line being unable to support full sync rates due to distance, line gauge, etc. I didn't like MaxSync for this reason back when I had DSL.
--
CableOne 10m/1m

navigator

join:2012-10-30
Franklinton, LA
Very informative reply, thanks.

Multiple people were switched to interleave when I first called in about the loss of net access. If there was a noise problem wouldn't we have lose sync too? The IP address would drop but the modem would stay synced up to the DSLAM.

So I tried figuring out these on-the-clock latency changes. I know it was good all day up til 3:30pm then I checked at 4:06pm and it had turned bad. It stayed bad til 12:15am. I was checking it on the dot at 12am and it was still bad. I let it go to about 12:08am before I stopped it. And then I tried again at 12:16am and now its back good. I noticed this happen at the exact same time last night, that's why I was checking it at 12. As of now the problem seems to exist from ~4pm-12:15am. Tomorrow I will try and let a ping run every 10secs or so around those times to identify the exact moment it happens. Any idea why it would be on time like this?

So what would be the optimal setup for me? I want low latency and stable connection even at the cost of some bandwidth. Can I have my sync data rate lowered so its more stable? Or have the sync profile changed?

Direct forum is switching me back to fastpath to see if that helps. Not sure when that will take place though. I may have to go back to interleave if the old problem comes back.