cancel
Showing results for 
Search instead for 
Did you mean: 

TBB points out Plusnet peak time speed differences.

Andrue
Aspiring Pro
Posts: 775
Thanks: 88
Fixes: 1
Registered: 12-01-2015

TBB points out Plusnet peak time speed differences.

17 REPLIES
deathtrap
Grafter
Posts: 1,063
Thanks: 4
Registered: 23-04-2013

Re: TBB points out Plusnet peak time speed differences.

Quote from: Andrue
..but maybe not what you think.
It's not what we may think , but more what we are experiencing or have experienced , they are facts, the data in that blog suggests that throughput increases during peak times, so is ludicrous and factually incorrect
Community Veteran
Posts: 4,975
Thanks: 372
Fixes: 16
Registered: 10-06-2010

Re: TBB points out Plusnet peak time speed differences.

Or it's something like more Plusnet customers with fast lines run lots of speedtests in the evenings.
Instead of running a speedtest, you could use some program that displays the speed for what you're actually doing, even Windows Task Manager can do that.
Andrue
Aspiring Pro
Posts: 775
Thanks: 88
Fixes: 1
Registered: 12-01-2015

Re: TBB points out Plusnet peak time speed differences.

Quote from: deathtrap
It's not what we may think , but more what we are experiencing or have experienced , they are facts, the data in that blog suggests that throughput increases during peak times, so is ludicrous and factually incorrect
TBB is the premier UK speed tester and they have access to large amounts of data. Far more than we do. All we know is that a handful of people (me for a while) have experienced some performance issues on and off since January.
I'm not trying to suggest that there isn't a problem. I have never even suggested that the problem has been fixed. For the past couple of months i've been lucky that's all. But these results from TBB cannot be ignored even if they don't fit our expectations. I've noticed before that you have something of a problem when it comes to discussions. You dismiss anything that doesn't match your preconceived notions and then get all shirty with those putting forward the alternate view. It's something of a character flaw that you really ought to address.
What I see in those results is odd (purely because you don't expect a speed increase in peak hours) but fits in with the general impression I've had that this problem has only been experienced by a relatively small number of PN's customers. That could also explain why it's taken PN so long to track it down. The mystery then becomes - why so few people?
But anyway - pick your toys up and put them back in your pram. I'm only drawing attention to an odd report. It doesn't really tell us anything either way. It's just an interesting discussion point.
matthews
Rising Star
Posts: 141
Thanks: 6
Fixes: 1
Registered: 13-08-2014

Re: TBB points out Plusnet peak time speed differences.

Quote from: ejs
Instead of running a speedtest, you could use some program that displays the speed for what you're actually doing, even Windows Task Manager can do that.

That's got to be a better solution all round. I dread to think how much of an ISP's bandwidth is taken up by running speedtests to try and demonstrate that something is wrong. Personally I know that I'm using about 3gb per day alone with a Samknows box for collecting stats. I think a reasonable proportion of that is from a CDN though (YouTube/iPlayer) so doesn't leave Plusnet's network. That's before any of my regular usage is taken into consideration.
The wonderful thing about all of these statistics is that they can be used to prove anything! There may be more data, it may be more people on fibre as opposed to ADSL are running them. Personally I've not noticed a slowdown at peak time or off-peak time, but I'm taking into consideration that very few of the sites that I visit will be able to give me the full 76mb download, so I very rarely see my connection get used at 100% anyway.
deathtrap
Grafter
Posts: 1,063
Thanks: 4
Registered: 23-04-2013

Re: TBB points out Plusnet peak time speed differences.

Quote from: Andrue
Quote from: deathtrap
It's not what we may think , but more what we are experiencing or have experienced , they are facts, the data in that blog suggests that throughput increases during peak times, so is ludicrous and factually incorrect
TBB is the premier UK speed tester and they have access to large amounts of data. Far more than we do. All we know is that a handful of people (me for a while) have experienced some performance issues on and off since January.

What I see in those results is odd (purely because you don't expect a speed increase in peak hours) but fits in with the general impression I've had that this problem has only been experienced by a relatively small number of PN's customers. That could also explain why it's taken PN so long to track it down. The mystery then becomes - why so few people?
But anyway - pick your toys up and put them back in your pram. I'm only drawing attention to an odd report. It doesn't really tell us anything either way. It's just an interesting discussion point.
Really well tonight i identified an issue with TBB speedtester (flash based) that they have since fixed, those like you who don't question everything and trust no one would and have automatically assumed that they where seeing peak time congestion, So you see my views are not bias i'm open to suggestions, but tbb's results  may not reflect reality , you are very quick to dismiss the scale of the number of customers who have since january2015 and before that  experienced peak time issues
which for some are still ongoing, the number of customers who have complained or vented their frustration on this forum cannot be used to realistically gauge the number of customers affected i doubt even plusnet would try that one, and that's pushing bounderies
Lots of customers have already migrated away , And regardless of where the underlying issues lie within BTW network or with Plusnet  these issues should of been identified and resolved by now facts are they have not for many,  i gave plusnet every chance  they failed
kitz
Rising Star
Posts: 817
Thanks: 46
Registered: 08-06-2007

Re: TBB points out Plusnet peak time speed differences.

When Im getting exactly the same speeds on a http download that is nothing to do with Flash or the TTB speedtester, then Im inclined to believe that the TBB speedtester was (for me at least) giving the correct results.
I never do trust one speedtester and if something looks odd then I will carry out my own investigations.   I'm very aware that multithreaded speedtesters such as Ookla and the BTw performance tester only masks congestion issues.
The reason I use NetMeter is because it is so damn good at showing actual throughput regardless of where its coming from and doesnt rely on things such as Flash.   Ive used to for years because its so good at picking things up and for elimination purposes of all sorts of things - even trojan activity.
This was my TBB speedtest

and here you can see my netmeter real time graph.   Note how it shows whats happening on the TBB speedtests and how single thread is limited.  Note how it jumps up at hhtpx6.   Then note how when carrying out http thats nothing to do with the speedtester or flash then Im still getting the pathetic 12.5Mbps average.
Netmeter is real good at showing signs of congestion or if its ISP throttling.       Use of Ellacoyas/Proceras or any ISP throttling show totally different shape when they are graphed.   I have some old Ellacoya throttling ones somewhere but when youve seen the shape it makes you'll recognise it again.   Its like very tiny/small castle tops.   Rate limiting via such things as capping or too low PN profile show as a flatline.  MTU or buffering can show as stop start wavy lines.    The spikeyness on the graph is typical of congestion.
Andrue
Aspiring Pro
Posts: 775
Thanks: 88
Fixes: 1
Registered: 12-01-2015

Re: TBB points out Plusnet peak time speed differences.

Quote from: deathtrap
those like you who don't question everything and trust no one would and have automatically assumed that they where seeing peak time congestion
Actually no. I was already wondering about the shape of the graph and that was why I visited TBB and saw that thread. You'll also note if you read my first post to that topic that I am uncertain about the cause:
From the mesage I first posted about the issue here:
"Edit:I just checked earlier in this thread and it's the same gateway I was on when I last posted about how for me at least things were looking good. I wonder if this time it is a TBB issue?
« Last Edit: 19/07/2015, 21:13
by Andrue »"
Then a bit later from my first post on the issue to TBB:
TBB: Sunday: 22:20 - "As a PlusNet customer (and one of the original victims of their current network issues) I'm not sure who to 'blame'"
And my second post of the night here I point out the odd looking graphs.
So there you go again. Making unwarranted accusations and assumptions. I was aware that something fishy was going on and was doing the research necessary to try and figure out what.
You really should try to keep an open mind.
Community Veteran
Posts: 38,432
Thanks: 1,011
Fixes: 60
Registered: 15-06-2007

Re: TBB points out Plusnet peak time speed differences.

Perhaps you should have read this http://forums.thinkbroadband.com/plusnet/f/4418931-peak-versus-off-peak-speeds.html?page=1&o=31 where, after a question, it was confirmed that they are multistrand test results
sjptd
Grafter
Posts: 467
Registered: 01-09-2014

Re: TBB points out Plusnet peak time speed differences.

My experience of graphs such as Kitz is showing were a result of internal network issues, and I know of a couple of other instances of this.
In my case it was a mismatch between the gigabit port on an HH3B and my basic gigabit switch.
computer -- switch -- HH3 gigaibit  ..... bad
computer -- switch -- HH3 100Mbps  ..... good
computer -- HH3 gigabit  ..... good
In each case it only showed up as bad on TTBx1;  TTBx6 and upload were fine in all cases.  Same on Infinity and Plusnet.
Other gigabit connections through the switch were all as they should be.
Kitz knows a lot more about these things than I do, so I doubt that is the reason in his case,
but worth other people checking if they have an issue with just TTBx1.
g1000
Grafter
Posts: 186
Thanks: 1
Registered: 08-03-2014

Re: TBB points out Plusnet peak time speed differences.

Quote from: deathtrap
the data in that blog suggests that throughput increases during peak times, so is ludicrous and factually incorrect

Yes, looks like they have just taken all the speed tests done in the evening and averaged them, and all the speed tests in the day and averaged them!
So these are different, uncontrolled groups doing the tests and being compared, no wonder the results appear essentially random. Totally meaningless.
Community Veteran
Posts: 6,824
Thanks: 1
Registered: 27-10-2012

Re: TBB points out Plusnet peak time speed differences.

Quote from: g1000
So these are different, uncontrolled groups doing the tests and being compared, no wonder the results appear essentially random. Totally meaningless.

Or the tests show that Plusnet's traffic management system works?
deathtrap
Grafter
Posts: 1,063
Thanks: 4
Registered: 23-04-2013

Re: TBB points out Plusnet peak time speed differences.

But it appears to show an increase against the known quieter periods of the day, match it maybe but to beat that something isn't right with it
Community Veteran
Posts: 6,307
Thanks: 86
Fixes: 3
Registered: 08-01-2008

Re: TBB points out Plusnet peak time speed differences.

I'm prepared to believe the TBB data, just remember that this is an average of many tests by many end users and for whatever reason it suggests that users on faster connections have more tendency to test speeds in the evening compared with those on slower lines.
It doesn't seem totally unreasonable to assume that those in regular employment ('9 to 5') are more likely to be paying for Fibre and that these people are more likely to be using the internet in the evenings, there may be other explanations but this is one example that could fit and explain what is being seen as an anomaly.
Call me 'w23'
At any given moment in the universe many things happen. Coincidence is a matter of how close these events are in space, time and relationship.
Opinions expressed in forum posts are those of the poster, others may have different views.
g1000
Grafter
Posts: 186
Thanks: 1
Registered: 08-03-2014

Re: TBB points out Plusnet peak time speed differences.

It's just random - look at the last 12 months of results, half are higher in evening, half are higher in day. This is because the groups are random and the results meaningless.
(Even if, hypothetically, the groups were controlled, Plusnet's traffic management still couldn't even begin to account for the obviously ridiculous results. It can't magically accelerate things in the evening compared to day, it doesn't even operate on upload but upload went faster in the evening when download went faster, and it is only meant to have an effect when there are capacity issues in the first place.)