Kitz ADSL Broadband Information
adsl spacer  
Support this site
Home Broadband ISPs Tech Routers Wiki Forum
 
     
   Compare ISP   Rate your ISP
 
Please login or register.

Login with username, password and session length
Advanced search  

News:

Pages: [1] 2 3 ... 10

Author Topic: Speed issues with AAISP  (Read 3975 times)

Craig

  • Just arrived
  • *
  • Posts: 13
Speed issues with AAISP
« on: June 14, 2018, 12:57:15 PM »

Hey,

I recently moved from Plusnet to AAISP, nothing on my home network has changed and I still use my Billion 8900ax-2400. But on the 12th May between 13:00 and 19:00 something happened which has caused my download speeds to decrease by 5Mbps. This isn't a lot in the grand scheme of things, but it's annoying considering the speeds I used to get.

After contacting AA they have said that this is most likely due to "congestion on the open internet" as speedtest.net and fast.com always appear to top out at 67Mbps, but their http://speedtest2.aa.net.uk/ shows ~73Mbps... I've attached line stats and a screenshot from Sam Knows, anyone have any advice on what I can do?

Code: [Select]
adsl: ADSL driver and PHY status
Status: Showtime
Last Retrain Reason: 0
Last initialization procedure status: 0
Max: Upstream rate = 24511 Kbps, Downstream rate = 79007 Kbps
Bearer: 0, Upstream rate = 20000 Kbps, Downstream rate = 79000 Kbps

Link Power State: L0
Mode: VDSL2 Annex B
VDSL2 Profile: Profile 17a
TPS-TC: PTM Mode(0x0)
Trellis: U:ON /D:ON
Line Status: No Defect
Training Status: Showtime
Down Up
SNR (dB): 6.3 9.2
Attn(dB): 11.5 0.0
Pwr(dBm): -1.2 -1.2

VDSL2 framing
Bearer 0
MSGc: 18 150
B: 239 236
M: 1 1
T: 23 5
R: 0 16
S: 0.0967 0.3771
L: 19856 5410
D: 1 1
I: 240 255
N: 240 255

Counters
Bearer 0
OHF: 198117635 54172648
OHFErr: 480 18
RS: 0 3519481844
RSCorr: 0 295
RSUnCorr: 0 0

Bearer 0
HEC: 3998 0
OCD: 130 0
LCD: 130 0
Total Cells: 3156476669 0
Data Cells: 1501971239 0
Drop Cells: 0
Bit Errors: 0 0

ES: 397 18
SES: 0 0
UAS: 28 28
AS: 331807

Bearer 0
INP: 0.00 0.00
INPRein: 0.00 0.00
delay: 0 0
PER: 1.67 6.15
OR: 114.65 202.87
AgR: 79114.95 20203.27

Bitswap: 2544/2544 55/55

Total time = 3 days 20 hours 10 min 35 sec
FEC: 0 295
CRC: 480 18
ES: 397 18
SES: 0 0
UAS: 28 28
LOS: 0 0
LOF: 0 0
LOM: 0 0
Retr: 0
FailedRetr: 0
FailedFastRetr: 0
Latest 15 minutes time = 10 min 35 sec
FEC: 0 0
CRC: 1 0
ES: 1 0
SES: 0 0
UAS: 0 0
LOS: 0 0
LOF: 0 0
LOM: 0 0
Retr: 0
FailedRetr: 0
FailedFastRetr: 0
Previous 15 minutes time = 15 min 0 sec
FEC: 0 0
CRC: 1 0
ES: 1 0
SES: 0 0
UAS: 0 0
LOS: 0 0
LOF: 0 0
LOM: 0 0
Retr: N/A
FailedRetr: N/A
FailedFastRetr: N/A
Latest 1 day time = 20 hours 10 min 35 sec
FEC: 0 52
CRC: 115 4
ES: 98 4
SES: 0 0
UAS: 0 0
LOS: 0 0
LOF: 0 0
LOM: 0 0
Retr: 0
FailedRetr: 0
FailedFastRetr: 0
Previous 1 day time = 24 hours 0 sec
FEC: 0 106
CRC: 121 4
ES: 99 4
SES: 0 0
UAS: 0 0
LOS: 0 0
LOF: 0 0
LOM: 0 0
Retr: 0
FailedRetr: 0
FailedFastRetr: 0
Since Link time = 3 days 20 hours 10 min 7 sec
FEC: 0 295
CRC: 480 18
ES: 397 18
SES: 0 0
UAS: 0 0
LOS: 0 0
LOF: 0 0
LOM: 0 0
Retr: 0
FailedRetr: 0
FailedFastRetr: 0
NTR: mipsCntAtNtr=0 ncoCntAtNtr=0
Logged

Ixel

  • Kitizen
  • ****
  • Posts: 1005
Re: Speed issues with AAISP
« Reply #1 on: June 14, 2018, 01:15:51 PM »

What does tbb.st do, same result?
Logged

Craig

  • Just arrived
  • *
  • Posts: 13
Re: Speed issues with AAISP
« Reply #2 on: June 14, 2018, 01:20:37 PM »

Just done a tbb.st


Logged

j0hn

  • Kitizen
  • ****
  • Posts: 1577
Re: Speed issues with AAISP
« Reply #3 on: June 14, 2018, 02:56:09 PM »

Well I just did http://speedtest2.aa.net.uk/ 3 times and each time I'm getting results faster than my IP Profile.
So they appear to be exaggerating results.

Do you know if you use BTw or TTB with your AAISP line?

If BTw can you visit https://windows.mouselike.org/be/index.asp?DoAction=BrasChecker
and post the results.
« Last Edit: June 14, 2018, 03:13:01 PM by j0hn »
Logged
BT FTTC 55/10 ECI now Huawei cab
Zyxel VMG1312-B10A bridge mode with 1508 MTU + Asus RT-AC68U running Asuswrt-Merlin

Craig

  • Just arrived
  • *
  • Posts: 13
Re: Speed issues with AAISP
« Reply #4 on: June 14, 2018, 03:00:41 PM »

That's interesting!

I am on the new TT backhaul
Logged

Weaver

  • Addicted Kitizen
  • *****
  • Posts: 5376
  • Retd sw dev; A&A; 3 × 7km ADSL2; IPv6; Firebrick
Re: Speed issues with AAISP
« Reply #5 on: June 14, 2018, 03:48:00 PM »

J0hn is right, the numbers are definitely exaggerated, my three ADSL2 lines’ combined rates reported by clueless (the lower figures, tx at 87% of sync rate, not the sync rates) come to less than the reported figure from that speed tester. Other speed testers exaggerate too, some far more and some go the other way possibly because that is what they really see, because of a poor network path or too much load on the server.

These servers ought to have an option to ask you what the maximum possible numbers are - they could ask for sync rate plus all your DSL parameters and calculate a fudge factor to de-rate the sync rate. That is what I do in a spreadsheet and in an iPad app that I have built which calculates upstream rate limiter settings and generates a snippet of a Firebrick’s XML config relating to PPPoE objects. That could be a sanity check / bull-tester for the speedtest app.

Also speed testers ought to measure the load that they are under and moan if it is too high so that that might affect the results and make it under-read. I also thing a speed tester should ping along the path between client and server for a fair while before the test begins and again afterwards, and maybe even briefly in between individual parts of a test sequence, so that spurious competing activity on the client’s own machine or own internet connection could be detected by disturbances to the latency. If the link is not quiet then the user ought to be told, or maybe the system could try waiting a little in case things quieten down while showing the user some sort of visual feedback about the quietness level related to alien traffic.

There are some test files for download at thinkbroadband which gives you a straightforward simple measure where you know exactly what you are getting and what you are taking about, the straight TCP payload rate. It is of course not close to you like a test hosted by your own ISP would be, so it should be liable to under-read. If you did it in the middle of the night that should help a lot, as long as there is not such a long RTT that you are not filling the pipe and it has bubbles in it.
« Last Edit: June 30, 2018, 05:13:50 PM by Weaver »
Logged

spring

  • Reg Member
  • ***
  • Posts: 322
Re: Speed issues with AAISP
« Reply #6 on: June 14, 2018, 03:55:12 PM »

In my mind it's only related to how they calculate the displayed speed.

I don't know what transfer method that site uses but speedtest.net uses web-socket so that's best in my view. I'm saying this because I often encounter the phrase "flash testers are inaccurate".
« Last Edit: June 14, 2018, 04:01:28 PM by spring »
Logged
No one knows what is the taste of the void.

Chrysalis

  • Content Team
  • Addicted Kitizen
  • *
  • Posts: 5396
Re: Speed issues with AAISP
« Reply #7 on: June 14, 2018, 04:01:23 PM »

well 1mbit of it may be explained by the sync speed, your current sync uptime is close to the time you said it dropped, and if you previously had a max 80mbit sync then thats 1mbit lost.  The other 4mbit I dont know, could possibly be a ip profile issue within the AAISP network if they use such profiles.
Logged
Sky Fiber Pro - Billion 8800NL bridge & PFSense BOX running PFSense 2.4 - ECI Cab - LINE STATISTICS CLICK HERE

Chrysalis

  • Content Team
  • Addicted Kitizen
  • *
  • Posts: 5396
Re: Speed issues with AAISP
« Reply #8 on: June 14, 2018, 04:04:07 PM »

Well I just did http://speedtest2.aa.net.uk/ 3 times and each time I'm getting results faster than my IP Profile.
So they appear to be exaggerating results.

Do you know if you use BTw or TTB with your AAISP line?

If BTw can you visit https://windows.mouselike.org/be/index.asp?DoAction=BrasChecker
and post the results.

make sure any a/v you have is disabled when running speedtests, any a/v that intercept's web traffic can skew speedtest results.
Logged
Sky Fiber Pro - Billion 8800NL bridge & PFSense BOX running PFSense 2.4 - ECI Cab - LINE STATISTICS CLICK HERE

Weaver

  • Addicted Kitizen
  • *****
  • Posts: 5376
  • Retd sw dev; A&A; 3 × 7km ADSL2; IPv6; Firebrick
Re: Speed issues with AAISP
« Reply #9 on: June 14, 2018, 04:08:06 PM »

@Chrys - What do you mean by IP profile, Chrys?

@Craig - what does the recent clueless.aa.net.uk log look like? It will show you the tx and rx rates in use there, which is the real truth, not the sync rate.

You may already know that you can change the speed to be a certain percentage of this max rate. If it is set to 95% then that is for VoIP users to make VoIP reliable so if you do not use VoIP you can bang that up to 100% in clueless.

Craig when you say your download speed dropped by 5 Mbps on that date, do you mean sync rate or measured performance? If the sync rate dropped and took the performance with it then it could be a deterioration of the line, perhaps because a new crosstalker arrived, a new neighbour’s activity, or it could be a new source of RF interference or forty things.

You can try and do the usual things in your own house to optimise performance, such as getting rid of extensions, shortening wiring, using high quality faceplate-type DSL filters, high quality modem-to-wallsocket cables and short ones too, mains interference filtering. Going further, better modem or even shortening the run of internal wiring inside the house from the outside to the master socket by getting it moved, which would cost money and possibly be very inconvenient (or quite the reverse!). But on a line as good as yours, your signal will be strong so these kind of measures are much less likely to give you anything compared to the effect on someone like me with a 4.55 mile long [!] phone line and hence incredibly weak signal that is therefore very vulnerable to local noise sources which relatively will appear really loud. But crosstalk is going to be a big problem for you as for all FTTC users who do not have vectoring (anti-crosstalk systems).

Could you remind me of the sync speeds now and then?
« Last Edit: June 14, 2018, 04:18:56 PM by Weaver »
Logged

Chrysalis

  • Content Team
  • Addicted Kitizen
  • *
  • Posts: 5396
Re: Speed issues with AAISP
« Reply #10 on: June 14, 2018, 04:11:58 PM »

IP profile is a name UK isp's use to basically determine a rate limit they set on customer's lines, apparently BTw require their customer isp's to set a ip profile on every line to match the expected throughput of the line.  Plusnet e.g. have issues making sure this is set correctly.

Whether or not AAISP have any such profile set on ttb lines no idea, but this speed setting you just mentioned is a form of rate limiting so may be the answer.
Logged
Sky Fiber Pro - Billion 8800NL bridge & PFSense BOX running PFSense 2.4 - ECI Cab - LINE STATISTICS CLICK HERE

Weaver

  • Addicted Kitizen
  • *****
  • Posts: 5376
  • Retd sw dev; A&A; 3 × 7km ADSL2; IPv6; Firebrick
Re: Speed issues with AAISP
« Reply #11 on: June 14, 2018, 04:25:15 PM »

@Chrys you were talking about what I thought you were talking about, good. I have BTW AA lines not TT ones. AA track BT with a certain delay and use a sync rate × some fixed fraction for ADSL2+/21CN I don't know about other systems. The have never got it wrong for me anyway, and when it changes you can see events logged in the 'clueless' control, logging and monitoring server, marked tx and rx rate. That system can, on request, also query the DSL for both sync rates plus other modem stats.
Logged

Craig

  • Just arrived
  • *
  • Posts: 13
Re: Speed issues with AAISP
« Reply #12 on: June 14, 2018, 04:26:54 PM »

@Weaver - Here's all the entries from the logs on clueless where my Tx rate has changed. My sync rate has remained pretty constant but the performance has dropped as can be seen from the samknows graph. My line rate is set to 100%.

Quote
10 Jun 16:42:42      Tx rate (adjusted) 77830724 to 77551930 (rx 20000000)   -Auto-
8 Jun 18:56:21      Tx rate (adjusted) 78529673 to 77830724 (rx 20000000)   -Auto-
8 Jun 18:14:57      Tx rate (adjusted) 78517893 to 78529673 (rx 20000000)   -Auto-
4 Jun 20:38:52      Tx rate (adjusted) 76959001 to 78517893 (rx 20000000)   -Auto-
14 May 18:21:28      Tx rate (adjusted) 77363449 to 76959001 (rx 20000000)   -Auto-
12 May 16:07:45      Tx rate (adjusted) 77143555 to 77363449 (rx 20000000)   -Auto-
12 May 15:42:10      Tx rate (adjusted) 76393559 to 77143555 (rx 20000000)   -Auto-
5 May 11:56:59      Tx rate (adjusted) 76424972 to 76393559 (rx 20000000)   -Auto-
3 May 18:29:19      Tx rate (adjusted) 76676280 to 76424972 (rx 20000000)   -Auto-
3 May 18:21:37      Tx rate (adjusted) 76770520 to 76676280 (rx 20000000)   -Auto-
3 May 17:59:55      Tx rate (adjusted) 77174968 to 76770520 (rx 20000000)   -Auto-
2 May 11:16:24      Tx rate (adjusted) 77614756 to 77174968 (rx 20000000)   -Auto-
30 Apr 00:25:39      Tx rate (adjusted) 76770520 to 77614756 (rx 20000000)   -Auto-
17 Apr 04:07:51      Tx rate (adjusted) 0 to 76770520 (rx 20000000)   -Auto-
Logged

Weaver

  • Addicted Kitizen
  • *****
  • Posts: 5376
  • Retd sw dev; A&A; 3 × 7km ADSL2; IPv6; Firebrick
Re: Speed issues with AAISP
« Reply #13 on: June 14, 2018, 04:44:24 PM »

Craig the thing about ISP-external speedtesters is that they are mostly pretty rubbish, many just wildly impossibly off - see my threads about them in the past, plus the fact that it is not their fault but they can't be in control of the path between your ISPS and their test server. The ones I use are Ookla, SpeedSmart, or speedof.me or the tester at thinkbroadband not forgetting the big test download files plus a stopwatch at thinkbroadband (no software to lie or go wrong). I am an AA long-time customer and amazingly I never knew about that AA tester until your post. Duh. So good tip, thanks. Some testers let you select a server. Make sure that you pick one in central london near docklands where AA’s core networking hardware is, and try several of these until you find the fastest one with the highest performance figures and the lowest ping times. These tester apps’ idea of suggesting to you that you pick one near to you is utter stupidity but that is what many of them do.

What are the ratios between sync rates and the various speed testers mentioned plus the AA one?

Btw if you do a test, always do several and then keep only the fastest most favourable result, because alien traffic can slow the results down but cannot boost them.
« Last Edit: June 14, 2018, 04:49:07 PM by Weaver »
Logged

Craig

  • Just arrived
  • *
  • Posts: 13
Re: Speed issues with AAISP
« Reply #14 on: June 14, 2018, 05:00:30 PM »

I understand that congestion, routing issues etc can cause issues and speedtests can sometimes give wildly inaccurate results - my issue isn't with the odd result being off it's the obvious drop in the samknows results. That drop happened on the 12th May, and I moved to them on the 17th April, so something must have changed somewhere?
Logged
Pages: [1] 2 3 ... 10
 

anything