|
|
Hello all
can anybody analyze the following example of a peak time speed test http://www.thinkbroadband.com/speedtest/results.html...
and an out of peak time speedtest
http://www.thinkbroadband.com/speedtest/results.html...
here are some of my previous speed tests http://sdrv.ms/1emcZBz on skydrive
starting from the bottom
(MAY to NOV 2013 ) these results are what I would like to see all the time.
4th DEC to 9th DEC (on Talk Talk connection) My old broadband provider, the post office moved my connection as part of their "upgrade" process.
This is when I asked for my MAC and moved to Zen (slow peak time single stream download speeds)
9th DEC to TODAY Slow peak time single stream download speeds.
my exchange VP=green http://usertools.plus.net/exchanges/?exchange=Rhos&e... on the 5th December
several neighbors (on the same exchange) have also commented on the change of speeds in the last month.
Ian
|
|
|
generally speaking if speeds slow during peak times and as long as the line itself doesnt show other symptons (such as increased error rates, low snrm) then it points to congestion issues.
If I were to guess I would say a VP issue at your exchange.
Talktalk was probably a LLU connection so not using a BTw VP.
Really you need someone from zen to tell you more, as they will have access to exchange information.
Plusnet Fibre Unlimited BQM
|
|
|
My exchange http://www.samknows.com/broadband/exchange/SWRSO very small exchange, no LLU so was using BTw also about the time i noticed peak time slowdowns.
if a zen bod can give me any more recent info re VP status it would be much appriciated
IAn
|
|
Register (or login) on our website and you will not see this ad.
|
|
|
Wow that is a small exchange.
Probably a capacity issue ... there was a VP capacity issue in September.
http://usertools.plus.net/exchanges/?exchange=Rhos&e...
So BT probably didn't increase the capacity enough.
|
|
|
|
I didn't notice any slowdown in September......i did read another thread on this forum about small exchanges terminating (if that is the correct term) at larger local exchanges....can't find it at the moment.
Ian
|
|
|
Out of interest what does VP stand for when you mention vp capacity?
Freeserve Dial-Up --> BTopenworld --> <n>ildram -->Talk Talk LLU --> ZeN
ASUS RT-AC66U
|
|
|
You should have followed the link 
V irtual P ath.
|
|
|
|
Hi,
I spotted your post in ISP Unhappiness; rather than cross-post I thought it best to reply here.
I'd recommend running the BT speed test at the time you experience slow speeds, so we then have a record we can present to BT showing when the problems occurs - from what I can see last time you contacted us the BT speed test result was good. However some tests showing poor performance should help add weight to a fault report.
If you're still in contact with others having the same problems then suggest they do the same with their ISPs.
regards,
Phil.
|
|
|
|
Thanks Phil,
thought that would be the case, not sure if the BT speedtest shows the slow down, (at least not as much as the single thread TBB test).
I will run a few tests over the next couple of evenings.
Ian
|
|
|
Is the effect reproduced if you download a single file, i.e. are speeds slow, but if you download several files the sum of the speeds is actually good.
|
|
The author of the above post is a thinkbroadband staff member. It may not constitute an official statement on behalf of thinkbroadband.
|
|
|
The exchange is so small it may have been one of those provided as an "Exchange activate" site then taken over by BT wholesale. It may only have 34Mb backhaul over copper, but spread over a max of 78 customers that is more than a larger exchange would get. However 16 people trying to stream Iplayer or the like would give contention.
If it has no fibre at all then it will be hard to increase the backhaul. It still shows as under evaluation on the Openreach lists so depends on the BDUK money and the Welsh Assembly priority.
( Exchange Activate was were a CP paid an up front fee to supply 30 customers with service on an exchange that would not otherwise have had any service)
|
|
|
The exchange is so small it may have been one of those provided as an "Exchange activate"
No , a BT wholesale enablement in July 2007 (one of the last in Wales)
VP history shows a problem in September (I did not notice it) so backhaul may have been added.
Ian
|
|
|
if you download a single file, i.e. are speeds slow, but if you download several files the sum of the speeds is actually good.
Yes, in the linked picture (on skydrive) you can see three downloads max out my connection
http://sdrv.ms/KPvHbZ
Ian
|
|
|
and a slow TBB test http://www.thinkbroadband.com/speedtest/results.html...
BT speedtest shows 5.4 Mbps at the moment
|
|
|
The BT test is probably a multi thread test, so hard to use that to judge contention based issues.
|
|
The author of the above post is a thinkbroadband staff member. It may not constitute an official statement on behalf of thinkbroadband.
|
|
|
Still having slow single thread speed tests http://speedtest.zen.co.uk/ test code AZSV2, BTW speedtest showing > 5 Mbps as expected.
Ian
|
|
|
|
another slow test (6499X) speedtest.zen.co.uk
and BTW <4 Mbps with IP profile of 6 Mbps
Ian
|
|
|
|
Phil
Can you take a look at my latest support ticket (6294901:4724571)
Thanks
IAn
|
|
|
|
Hi Ian,
I've checked in on this and had a chat to the chap dealing with is - who's responding based on the latest information you've provided; I've also highlighted this thread as there's some additional information here.
I'm not around much today, and then on leave for the next few weeks - so won't be able to check in on this again. I'd recommend dealing with support directly.
In general it does sound like contention - but from the logged BT speed test it's not at a level below the threshold for your line speed, so that might explain why BT don't currently see this as an overly-contended VP.
regards,
Phil.
|
|
|
generally speaking if speeds slow during peak times and as long as the line itself doesnt show other symptons (such as increased error rates, low snrm) then it points to congestion issues.
If I were to guess I would say a VP issue at your exchange.
Talktalk was probably a LLU connection so not using a BTw VP.
Really you need someone from zen to tell you more, as they will have access to exchange information.
If it was VP congestion, surely no matter how many threads you create, it should still be the same overall speed? BT don�t see the per thread connection, they just see the one L2TP connection, thus each subscriber at the exchange gets an equal share of the available bandwidth.
|
|
|
No, because with multiple threads you are throwing more packets at the problem, and thus are able to grab a larger share of the limited capacity that is available.
It is how these packets are encapsulated and transported over the VP that are the issue.
|
|
The author of the above post is a thinkbroadband staff member. It may not constitute an official statement on behalf of thinkbroadband.
|
|
|
No, because with multiple threads you are throwing more packets at the problem, and thus are able to grab a larger share of the limited capacity that is available.
It is how these packets are encapsulated and transported over the VP that are the issue.
But the exchange / BT network will only ever see one thread, which is the L2TP tunnel. Thus 1 user couldn't take an unequal share of what bandwidth is available.
You could have multiple threads within the L2TP tunnel, which would help maximise what bandwidth is available if the source / ISP was congested, however it would never equal more than your equal share at the BT exchange / cloud.
|
|
|
WIth multiple threads you have more encapsulated packets appearing at the L2TP endpoint queue and thus grabbing a better share of the resource.
Latency also has a bigger impact for single thread download
|
|
The author of the above post is a thinkbroadband staff member. It may not constitute an official statement on behalf of thinkbroadband.
|
|
|
|
Hi Ian,
Replying here to follow up to your question in the fibre forum.
I've had a look at our data for your exchange, we've got a couple of customers there but our quality data is a bit inconclusive because there aren't many of them. The VP status was fine as at 13th March. We've only had quality data reporting for a couple of days so I'll have a look after the weekend and see if I can see anything then.
|
|
|
Dave
Thanks for the time and effort
IAn
|
|
|
|
We've got some decent data for Saturday evening, unfortunately it shows no packet loss. Sunday is hard to be sure as there's not much data, there is a bit of packet loss across the day and some latency increases but as there's only a couple of customers it could easily be misinterpreted. The 4pm latency increase on your TBB graph is interesting though as it does look like I can see that and there are patches of packet loss, I'll have a dig and see if I can see any further upstream, if it's at the exchange I've not really got any evidence but if there's a shared network element further up I might have customers affected.
Do you know if any of your neighbours are with Plusnet? If so and they are seeing the same as you get them to contact us and we can try and see what we can see. Even if they aren't still worth a try as we can with across industry to find the problem.
|
|
|
Dave
This problem does not seem to show up as a large increase in packet loss, although last night (Sunday) TBB speedtester single thread was only showing about 1.5 Mbps.(6x thread showing >5.5 Mbps)
BTw speedtester was reading about 2.5 Mbps (My connection speed 6880 kbps).
Do you know if any of your neighbours are with Plusnet?
I have refered two neighbours (i'm an ex customer) to plus net, i shall speak to them over the next few days and report back.
Thanks again
Ian
|
|
|
Dave
Spoke to my neighbours, one only uses the internet for email and browsing, the other family have recently had a line fault (noisy voice line) and are happy enough with thier iplayer/youtube/netflix streaming.
I did notice more packet loss yesterday evening on my bqm http://www.thinkbroadband.com/ping/share/614637e3c0f... from 8 to 10 PM.
Thanks
Ian
|