Kitz Forum
Broadband Related => FTTC and FTTP Issues => Topic started by: chuffer on May 25, 2019, 10:28:00 AM
-
Anybody who's on a Huawei AIO all in one, see here at the bottom https://kitz.co.uk/adsl/fttc-cabinets.htm
Are you getting the maximum possible throughput ?
I have a sync of 80Mb but am lucky to get 55/60Mb downloads.
BT/OR engineers have made several visits but of course as they only check sync speed say everything is OK and it's very difficult to get them to do an actual download test.
Help !.
[Moderator edited to change "A10" to "AIO" in the first sentence.]
-
The cabinet/DSLAM type is pretty irrelevant.
ISP's with poor support just keep sending OpenReach engineers for poor throughput but there's zero the engineer can do.
The issue will be on the ISP's part of the network, beyond the exchange, known as the backhaul.
Who is the ISP?
-
BT/OR engineers have made several visits but of course as they only check sync speed say everything is OK and it's very difficult to get them to do an actual download test.
Help !.
What happens to your connection once it leaves the exchange back towards the ISP is, unfortunately, not their problem. They can't do anything with whatever they may find even if they do this test. If the Openreach kit were broken everyone on the cabinet would have issues.
-
My ISP runs KDB tests and it shows in my profile that downloads have retransmission at high.
6 of my neighbours from a range of ISP's also have throughput issues so I doubt it is down to all the individual ISP's.
One, with Sky, has had an email saying they cannot complete the free upgrade they promised to 80 Mb due to network issues.
-
What's your IP profile?
Try entering your number here
https://windows.mouselike.org/be/index.asp?DoAction=BrasChecker
Can you run a speed test and post a link to the results
https://www.thinkbroadband.com/speedtest
Both will help pinpoint issues.
-
Just making sure.................
I guess you are asking for the BT ipprofile ( throughput and sync speed ) ?.
That is not the same profile data as shown in the KDB test.
Mouse like does not work saying BT refused the request.
That info is also available running the further diag from the btwholesale speed test site.
That also does not work with a failed to write to the server message, I can get that to run for another number.
The thinkbroadband test often shows high latency, currently around 500/800 but earlier, but not for a while, it was frequently 5000/6000.
I keep a laptop with very little other software installed, and not used for anything else, so to get consistent results, and no way to copy/paste the link.
Last test on Friday, I write them down was 54.1D 14.7U Latency 842.
If you really need a link I'll copy it down and post.
-
I keep a laptop with very little other software installed, and not used for anything else, so to get consistent results, and no way to copy/paste the link.
If your running the Think Broadband speed test, then surely you can use the same browser to login into Kitz and paste the link.
Its the graph John's interested that the speed test produces, not so much the figures.
-
Its the graph John's interested that the speed test produces, not so much the figures.
What Ronski said.
Can you post a link to the results here so we can see the graph of the test.
The numbers provided by the BT Wholesale Speed test (IP profile in further diagnostics) site are indeed what I was looking for.
You never mentioned the ISP. That would help.
-
I had a problem where packets were getting corrupted - low upstream SNRM - and when I fixed it the results from speed testers improved by about 20%, again that happened to be upstream, but I don’t see that’s important. Described in an earlier recent thread - if only I can find it.
And that was regardless of sync rate. So your own line / modems can affect real throughout aside from sync speed and isps and the internet. However, it just doesn’t happen very often, or perhaps more likely people don’t notice it because they aren’t looking into the details of what’s going on.
So I’m assuming there are no problems with line or modems. If you are using wireless then that’s a whole kettle of slippery fishies on its own. Ought to rule that out first. Speed testing on a wireful connection straight into the router is your friend.
The kind of speed tester that is very close to you, such as a BT special one - if you are on an Openreach / BTW service - or one at your own ISP, is especially valuable as then you miss out all the dangers of congestion on the wider internet. You still have the problems of overload at the test server or the pipe going into it if the speed tester is too popular.
-
Gentlemen.
Thanks for all your input here and I understand you all would like to establish a base line and I'm just trying now to save you wasting time.
So here goes................
I am on a BTW service, all speed tests are using a wired connection.
Always using the same laptop ( as above ) and six testers and they are in general agreement, some variations as of course they all do it using slightly different methods.
Taking my laptop to a friend 10 miles away, who is on a BTW circuit ( 245 Mtrs ) who syncs at 80/20, and I get a consistent general 76Mb ish download across the 6 testers.
At home netstat shows a upload error rate of 7.5% against a suggested 0.05% max.
Used a range of different Modems and routers connected directly to test socket in NTE5C.
There are only 7 connections on my pole which is wired directly to the chamber in front of my A10 cabinet.
The last OR engineer and his supervisor, using my test laptop, got the failed message at the BTW speed test further diag, he entered the number for one of the other 6 and it reported a IPprofile of 50Mb.
My neighbours are on a range of 5/6 ISP's.
The most telling things for me is
1, the recent email to my neighbour from Sky.
" we cannot complete the free upgrade we promised to 80 Mb due to network errors "
2, a Delta report on my circuit given to my ISP by OR quoting 50.4Mbps.
I am really hoping to hear from anybody on a A10 Huawei cabinet, sometimes installed as an infill, and users were originally on another cabinet further away and therefore now on a much shorter cable length, (I went from 1200Mts to 170.) and it is then connected to the BTW network in a different way from the other type cabinets.
I'm more than happy to answer any further question you might have.
Thanks
-
The DSLAM's in the All in One cabinets are the same DSLAM's used in the smaller Huawei cabinets.
There are literally 10,000's of these in the country.
There is nothing unique about the DSLAM's used in All in One cabinets that would cause issues.
You've ignored the questions that could help pinpoint your issues and instead seem focused on something that 2 experienced members already tried to tell you is NOT the issue.
Why is it so hard to post a link to a speed test?
Sky having network issues does not mean the DSLAM's are broken.
Good luck with your issue. I'll leave it to others to respond.
-
The cabinets are AIO (All In One) and not A10 (A-ten).
They contain identical Huawei SmartAX MA5616 DSLAMs (actually an MSAN configured and connected as a DSLAM) as the other smaller, Huawei equipped, cabinets.
A Huawei AIO cabinet is, invariably, installed as a "fill-in" cabinet. That is to say an extra cabinet, installed when the distance from the existing PCP (primary cross-connection point) and its "fibre twin" cabinet is too long for a cluster of end-users to receive a VDSL2 based service.
-
I wasn't sure what cab I have exactly so I just walked up the road to check.
Looks like a Huawei 288~ to my untrained eye?
Looking at the pics of those cabs, however, one thing that does spring to mind is... are the air vents greatly reduced on this AIO unit?
If so and these cabs overheat, is the processing speed of the chipset reduced to compensate? Could this explain the reduction in throughput you're seeing?
-
I assume this is my PCP to the left of it, although it looks a little more heavy duty than some I've seen?
-
I wasn't sure what cab I have exactly so I just walked up the road to check.
Looks like a Huawei 288~ to my untrained eye?
That is, indeed, a Huawei SmartAX MA5603T equipped cabinet which has been enlarged (at the right-hand end) to take all the expansion gubbins, increasing the number of circuits from 288 to 384.
-
I assume this is my PCP to the left of it, although it looks a little more heavy duty than some I've seen?
That is a very standard Primary Cross-connection Point, with a pressed-steel shell.
-
Thanks for confirming, bawss! :D
chuffer, how is your latency so high btw? What does your UK RTT look like?
Latency is a factor in TCP throughput, I believe?
TCP-Window-Size-in-bits / latency-in-seconds = bits-per-second-throughput
Have you had a squint at the wireshark output when you're running a speedtest? Does MSS look normal? TCP retx?
I'm sure you've checked all this so was more for anyone else who might search on your thread in future.
-
one thing that does spring to mind is... are the air vents greatly reduced on this AIO unit?
If so and these cabs overheat, is the processing speed of the chipset reduced to compensate? Could this explain the reduction in throughput you're seeing?
The vent is the exact same size as that of the smaller Huawei cabinets that also house the MA5616
https://kitz.co.uk/adsl/fttc-cabinets.htm#fttc_street_cabinet
There are 10,000's of these working across the network.
-
Might it be possible the 55Mb policer is applied to his line? Can Openreach test the throughput to the cab / cpe, j0hn?
As an aside, this is from the FTTP SIN:
802.1p = 0 marked traffic will have access to half of PBS & CBS.
802.1p = 1,2,3 or 4 marked traffic will have access to the full PBS & CBS.
A reduced burst-size can affect throughput on some kit.
Does this have any relevance for FTTC? People love to donkey around with the 802.1p bits on the HG612 :D
-
tumbleweed.gif
-
tumbleweed.gif
Have you been on my cat-nip or the falling-over juice? :angel:
-
Still have low throughput with high sync problems.
My ISP shows me KDB test results and some of the data does not appear to be in real time.
Sync's appears to be real time.
Profile name does not.
0.128M-74M Downstream, Retransmission Low - 0.128M-20M Upstream, Error Protection Off
* this appears to read retrans only on the downstream *
It says my cable is 162 Mtrs and my line is now banded ( Rate Capped Profile ? ) at 74, down from 80.
There is a Time Stamp and it can be up to a max of 13 days behind the date of them running the test.
ISP reports occasional copper faults but never seen by OR engineers on site.
" Current RRT Brag stability index is good last 3 days, very good the 3 days before that. Some upstream forward error correction evident,
nothing on the downstream, it is very variable however, only 8 yesterday, 128 on Friday."
* this says retrans only on the upstream *
Standard speed test sites often report high latency
Comments welcomed
-
The profile name section is exactly 13 days out of date on all GEA tests with any ISP.
We've known about this for a while.
The ISP can get a more up to date profile running the RRT(Reactive Repair Tool) that you mention.
-
Thanks jOhn.
It that the only field 13 days behind or are there others ???
-
jOhn.
Are you able to add anything to your last post re any other field in the KDB database being 13 days behind ?
The other thing I have noticed, every time an OR engineer has visited me and run tests the KDB date gets set to that date and time and remains so for the 13 days.