NOTdarth wrote on Sep 8^th, 2014 at 6:15pm:


Oh dude... NEVER post like that after dinner. I wouldn't have been surprised if I had thrown up my shoes.

Resurrecting because I actually did a traceroute while it was going on and it's kinda epic in an epic fail way.

Microsoft Windows [Version 6.1.7600]
Copyright (c) 2009 Microsoft Corporation.  All rights reserved.

C:\Users\Laptop>tracert 74.201.106.23

Tracing route to 74.201.106.23 over a maximum of 30 hops

  1    82 ms    77 ms    98 ms  BThomehub.home []
  2  1727 ms  1471 ms  1382 ms  217.47.109.186
  3  1940 ms  2188 ms  2366 ms  217.47.109.161
  4   459 ms   556 ms   668 ms  213.1.69.230
  5  2382 ms  2600 ms  1795 ms  213.120.162.69
  6  2039 ms  2290 ms  2337 ms  31.55.165.109
  7  2555 ms  2566 ms  2440 ms  acc2-xe-4-1-2.mr.21cn-ipp.bt.net [109.159.250.240]
  8   509 ms   517 ms   516 ms  core1-te0-13-0-0.ilford.ukcore.bt.net [109.159.250.137]
  9   564 ms   587 ms   581 ms  peer2-xe-4-2-0.telehouse.ukcore.bt.net [109.159.252.49]
10   618 ms   661 ms   668 ms  t2c3-xe-2-1-2-0.uk-lon1.eu.bt.net [166.49.211.192]
11   706 ms   703 ms   716 ms  195.66.224.130
12   906 ms   933 ms   916 ms  vb1042.rar3.nyc-ny.us.xo.net [207.88.13.202]
13  1008 ms  1031 ms  1012 ms  ae0d0.mcr1.cambridge-ma.us.xo.net [216.156.0.26]

14  1049 ms  1059 ms  1055 ms  ae1d0.mcr2.cambridge-ma.us.xo.net [216.156.1.14]

15  1121 ms  1103 ms  1108 ms  209.117.103.10
16  1204 ms  1247 ms  1232 ms  border1.te8-1-bbnet2.bsn003.pnap.net [63.251.128.107]
17  1282 ms  1284 ms  1321 ms  turbine-7.border1.bsn003.pnap.net [64.95.76.202]

18  1368 ms  1363 ms  1374 ms  74.201.102.170
19     *        *        *     Request timed out.
20     *        *        *     Request timed out.
21     *        *        *     Request timed out.
22     *        *        *     Request timed out.
23     *        *        *     Request timed out.
24     *        *        *     Request timed out.
25     *        *        *     Request timed out.
26     *        *        *     Request timed out.
27     *        *        *     Request timed out.
28     *        *        *     Request timed out.
29     *        *        *     Request timed out.
30     *        *        *     Request timed out.

Trace complete.

C:\Users\Laptop>

So what's going on there, apart from BT hating me and my connection being made of string? Latency is given at about 1.7 secs in game.

Terebinthia wrote on Sep 12^th, 2014 at 1:50pm:


that really looks like a problem with your router/connection to your isp. 2nd hop hitting 1.7sec is not good.

now that I think of it, you could be using some sort of wifi connect to get 80ms 1st hop, perhaps its an RF interference issue that happens at certain times?

Is there anything I can do about the first bit? I don't use our wireless network, instead I am wired into a Homeplug ( one of those things that enable you to run the connection through the power cables. Hubby tells me the switch is a gigabit and the individual plugs are 500 mb. Sadly there is no way to run the cable directly to the router due to our house layout. I am also the last house but one on the exchange and over 6 miles from the exchange down the valley.

You may all point and laugh at my connection now

It isn't like it all the time, thankfully, it is impossible to play like this, too much rubberbanding. It happens fairly randomly, more when I play UK daytime I think.

When I play with my two UK based running partners we often experience lag at the samr time. I know one is on Plusnet and they both have better connections than I do.

Reading through all that DACS junk has only confused me further, because it seems to say it doesn't work with broadband and I definitely have that!

Neighbours have equally miserable connection speeds. I guess I could call BT to confirm it isn't switched on but I think it must not be since we can get broadband.

Right, so the DACS box is probably just a historic thing. Interesting. There is a scheme to provide fibre to the cabinet going on atm via http://connectingshropshire.co.uk/faqs/ and that is scheduled to hit my exchange sometime before December. I guess it will help with some of the hops. It can't hurt!

I am a good 6 miles from the exchange, so don't think they are going to upgrade that, based on the population density. What weirds me out about the DACS box is it is probably only 4 or 5 years old, my guess on when previous owner got their act together and wanted a second line. So WtF is it doing there? Grrrr

It will be interesting to see how they deliver the promised upgrade to 2MB that is supposed to be happening to all the really rural areas, for sure. For comparison speedtest results usually show something like 800kb and it's more like 120 kb when downloading stuff. Anything is better than nothing, I guess. You can't have more than one person doing anything though - I imagine we will need to reactivate the second line for the business just for that once it gets going.

I will do a traceroute when it is not being sucky sometime.

Well you still have an issue with the DSLAM ( Hop 2 ). 1.8 sec is way too long.

The rest of the first colums is mostly irrelevant due to that hop 2. ( and to a lesser extend due to heop 3 )
but 140ish ms to Turbine entrance router is noty a bad value for our side of the Pond. Just crossing the Pond costs 60/70ms at least.

The trace shows that the issue with the second hop is not permament ( otherwise you owuld have propagation time climbing up from 1.8 sec instead of lower times )
But there's enough weird values in it to tell that there's something screwed in that Hop2.
As an example mine shows the following :

  1     2 ms    <1 ms    <1 ms  Homerouter.home
  2    16 ms    16 ms    16 ms  80.10.123.201
  3    16 ms    16 ms    16 ms  10.123.109.202
  4    16 ms    16 ms    16 ms  ae41-0.nista202.Paris.francetelecom.net [193.252.98.177]
  5    18 ms    18 ms    19 ms  81.253.184.90
  6    17 ms    17 ms    17 ms  be3044.ccr21.par04.atlas.cogentco.com [130.117.15.93]
  7    17 ms    17 ms    17 ms  be2308.ccr41.par01.atlas.cogentco.com [154.54.73.229]
  8    24 ms    24 ms    24 ms  be2271.ccr21.lon13.atlas.cogentco.com [130.117.49.117]
  9    30 ms    32 ms    31 ms  be2283.ccr21.lpl01.atlas.cogentco.com [130.117.49.34]
 10    97 ms    98 ms    97 ms  be2386.ccr21.bos01.atlas.cogentco.com [154.54.44.161]
 11    96 ms    95 ms    96 ms  te3-4.ccr01.bos06.atlas.cogentco.com [154.54.46.134]
 12    99 ms    98 ms   102 ms  38.122.52.50
 13   114 ms   114 ms   118 ms  border1.te7-1-bbnet1.bsn003.pnap.net [63.251.128.43]
 14   113 ms   114 ms   113 ms  turbine-7.border1.bsn003.pnap.net [64.95.76.202]
 15   103 ms   103 ms   104 ms  74.201.102.170
 16     *        *     ^C
 

As you can see values just keep climbing, and they are pretty constant for a given hop.
They all fall within 5 ms of each other.

Yours are just incoherent :

13   144 ms   143 ms   680 ms  ae0d0.mcr1.cambridge-ma.us.xo.net [216.156.0.26]
14   615 ms   326 ms   128 ms  ae1d0.mcr2.cambridge-ma.us.xo.net [216.156.1.14]
 

That's half a second between first and last set... one way or another...

So, I've worked out how to access my ADSL stats:


Line state      Connected
Connection time      3 days, 18:10:37
Downstream      1,280 Kbps
Upstream      448 Kbps
ADSL settings
VPI/VCI      0/38
Type      PPPoA
Modulation      ITU-T G.992.1
Latency type      Interleaved
Noise margin (Down/Up)      6.4 dB / 11.0 dB
Line attenuation (Down/Up)      63.5 dB / 31.5 dB
Output power (Down/Up)      16.2 dBm / 12.3 dBm
Loss of Framing (Local)      0
Loss of Signal (Local)      0
Loss of Power (Local)      0
FEC Errors (Down/Up)      624473 / 480
CRC Errors (Down/Up)      7128 / N/A
HEC Errors (Down/Up)      N/A / 236
Error Seconds (Local)      5316

And read up a little bit on Line attenuation - and the kitz tool at http://www.kitz.co.uk/adsl/max_speed_calc.php gives an approximate line length of 4.6 kn when the actual distance to the exchange as double-checked is 7.9. So, that suggests the physical phone lines aren't in too bad condition, am I right?

I suppose I need to gird my loins and talk to the Hell Desk, but am not looking forward to it, as they basically will try to fob me off. Hate BT customer support

Terebinthia wrote on Sep 14^th, 2014 at 10:41am:


Yup, they don't look that bad... 8 klicks for 1Mb down is actually quite good. 63dB is in line with 1Mb bandwidth, so there's nothing strange here.
( it's not in like with 8 klicks though... 8 klicks shoud be in the 70/80 dB attenuation, so the physical line is quite nice, probably something relatively new, laid to dispose of the DACS box. ) FEC errors don't matter since your line is in interleaved setting.

I'm surprised by the PPPoA ( that's point to point protocol over ATM ) since it's a mostly deprecated system most DSLAMs nowadays uses PPPoE ( that's the same over Ethernet ), so you're probably connected to a first generation DSLAM, which might explain why it has issues with it's backend link. ( depending on the technology [ Ericsson, Alcatel, Nortel, Siemens, ... ] the uplink after the DSLAM might be limited to one to four 2Mb links... for 128 Subscribers )

Hell Desk Drone will probably be unable to help you... if they are like here... I usually end up talking to supervisors as the drones have no clue beyond following the page on their screen... At lesat supervisors listen when I tell them that DNS server at IP X.Y.Z.A stopped answering queries, or that the RADIUS server W.X.Y.Z is down... and they check the thing. ( while the drone keeps telling me to restart my box ( despite the fact that I told him/her 4 times that I do not have a box, but that I have a router with full control over it ).

At least you might be able to get them to check the DSLAM health and see if there's anything wrong there.