ASA Ban PlusNet Advert

[#1]

Makes interesting reading ..
http://www.ispreview.co.uk/index.php/2016/03/asa-ban-plusnet-advert-misleading-totally-unlimited-bro...
Kind regards
Mike

[#2]

Quote

“there were no provider-imposed limitations to the service and to explain clearly any limitations they applied“.

So, apart from the self-contradiction in the quotation, "totally unlimited" means that the ISP cannot apply a "profile" to a line!
I'm not clear what the ASA are really objecting to; I assume it is that prospective customers should be explicitly told the obvious: that they cannot get more throughput than the line is capable of providing? I do not consider a minimum bandwidth provision, when it is needed, imposes a limitation in the way described.
I don't use P2P services, so don't know why "immediacy" is a requirement for some people.
Perhaps I'm in a minority of those who consider advertising to have a "believability" index slightly better than that of politicians (there is no "lie by omission" factor    )
Phil

[#3]

I think the main issue is that the ASA clearly don't understand how the prioritisation works either (which may confirm that it's poorly explained)

Quote

For example, we understood that if higher priority traffic was using the full capacity of the line and traffic prioritisation applied, Plusnet would allocate up to 95% of the capacity of the line to the higher priority traffic. p2p traffic’s minimum running capacity would therefore be 5% and so the p2p download would take 20 times as long as if the connection was solely being used for a p2p download.

They are completely missing the point here...

[#4]

It's madness that something like this gets wrongly picked up while "Free" continues to be misused.

[#5]

ISTM the ASA have made an A55 of themselves.

[#6]

Just shows a total lack of user standing of how things work.  Without traffic prioritisation (which is quite different from limiting) there would soon be complaints in households up and down the country when real time applications such as VoIP and video streaming become 'defective' as a reality of someone else in the household doing p2p filesharing.  It taking a little longer to transfer a file whilst someone else in the household enjoys catching up with east Enders on iplayer is not exactly the end of the world.
The ASA has shown them selves to be quite shoddy here.
I don't see any investigation to BT Retail's very misleading suggestions that having the very fastest wifi will make a big difference to your internet connection experience.  Even if wifi ran at 1Gbps, it's of no benefit if the crappy BT infrastructure will only deliver 5Mbps to the exchange.  Even the fastest of fibre connections is significantly slower than the capability of moderate wifi standards.  I consider BT Retail's advertising to be grossly misleading, does the ASA do anything? No!
However, I suspect that this complaint arose from some silly little twerp with an axe to grind against PlusNET.

[#7]

@Townman Pn could of offered the option for a customer to disable prioritisation , as not every subscriber  has several devices with several people sharing the bandwidth , So it is useless to them, in fact because it exists and can't be disabled  it actually acts to restrict the max ds throughput,

[#8]

After the debacle of the ASA allowing BT to describe FTTC as fibre broadband I was beginning to wonder if the ASA was somewhat steered by major advertisers such as BT.  Sadly I am beginning to think the ASA is just plain incompetent.

[#9]

Quote from: deathtrap

Pn could of offered the option for a customer to disable prioritisation

They do,  it's called the "Plusnet Pro Add-On"

Quote from: deathtrap

not every subscriber  has several devices with several people sharing the bandwidth

but it also improves the performance for a single user doing more than one thing at a time, for example web browsing while waiting for a download to complete.

Quote from: deathtrap

it actually acts to restrict the max ds throughput,

Many people think that, but in practice that is not how it works.  For anyone who has spent time setting up their router's "QoS" or tuning their settings to minimise "BufferBloat" soon realise that maximum THROUGHPUT occurs when sending data at a rate which is slightly slower than the maximum BANDWIDTH of the connection (in this case the BT line profile).  This is because in practice, errored or dropped packets need resending in addition to the wanted traffic, the overhead associated with resending dropped packets (because of sending data AT full bandwidth) is greater than artificially sending the data at a fractionally slower rate (when fewer packets are lost - and therefore don't need resending).

In my opinion, Plusnet's superior traffic management is one of the reasons that I have remained a customer, as my 20CN (up-to) 7.15Mbps connection, shared between four users - all with heavy usage,  needs all the help it can get to achieve MAXIMUM and FAIR throughput,  as our download bandwidth is running maxed out for at least four hours every day - with my router's traffic graph literally flat lining at the Plusnet current line speed.  
Because traffic management can only be achieved by the data sending end, and because my broadband is so overloaded, I have implemented a similar traffic management scheme with specific traffic prioritization, on the upstream path from my home router, and the resulting improvements to responsiveness, reduced buffering, and eliminating gaming lag, is very noticeable and worthwhile.  Unfortunately many Plusnet customer's use the supplied modem/router which means that they have virtually no ability to implement their own upstream traffic management, hence the many threads on this forum about gaming lag, latency issues, or video buffering.

[#10]

Quote from: AlaricAdair

After the debacle of the ASA allowing BT to describe FTTC as fibre broadband I was beginning to wonder if the ASA was somewhat steered by major advertisers such as BT.  Sadly I am beginning to think the ASA is just plain incompetent.

The problem is that they are trying to run a pseudo-legal court process - without the resources while under pressure to release decisions quickly as advertising is by it nature a short time-related immediate issue.
So if you complain, the company is asked to respond, you are allowed to respond to their comments and them to you second response.
Then they make a decision.
You are told under a NDA, it is then released to the press and embargoed until a date when it is displayed on their website.
I've won one BB related complaint and I've lost one BB related complaint - neither incidentally involved Plusnet.  Both required a lot of effort on my part to put in a detailed argued submission - rather than the usual of "I'm outraged by the advert".
There is no facility to introduce expert or independent witness comment/submission  - unless I suppose the complainers are themselves are regarded as experts in the field.
There is an appeals facility to an independent assessor.
To make it more robust is going to mean decisions take a lot longer and more importantly the elephant in the room of who would be paying for all of this

[#11]

Quote from: Nibiru

Quote from: deathtrap Pn could of offered the option for a customer to disable prioritisation They do,  it's called the "Plusnet Pro Add-On"

That doesn't disable the prioritisation, it only changes some of the prioritisation levels.

Quote from: Nibiru

Quote from: deathtrap it actually acts to restrict the max ds throughput, Many people think that, but in practice that is not how it works.

I thought it had been clearly demonstrated that that is indeed what actually happens - the Plusnet profile makes the sustained throughput slightly slower than what it would be without the Plusnet profile, and the difference is not just that the Plusnet profile is rounded down slightly from the BT IP profile. Pretty much every explanation offered by Plusnet has been demonstrated to be incorrect, and I still suspect the speeds have somehow been calculated or applied incorrectly or they've accounted for some layer of protocol that shouldn't have been.
See https://community.plus.net/forum/index.php/topic,114083.0.html and https://community.plus.net/forum/index.php/topic,126233.0.html
Perhaps you've never seen the difference for yourself because your profile has always been 7.15 Mb, or perhaps there isn't quite the same difference on 20CN. On a 21CN connection with a speed that might vary slightly from time to time, I have the silly situation where the line speed and BT IP Profile might drop slightly, but the download throughput remains the same, until the Plusnet profile goes down to "match" the BT IP Profile.

[#12]

I stand corrected regarding the Pro add-on.   :-[,  
I had forgotten that there was still a slight prioritization of time sensitive protocols, even when low priority protocols have been raised.

---

I also agree that the traffic prioritization is not perfect, due to -

• The unreliable/delayed updating of the Plusnet current line speed (after changes in the BT line profile).
  
  


• The BT line IP profile only tracks the customers modem connection rate and does NOT adaptively make any consideration for when the exchange is overloaded and has become congested, and therefore the Plusnet "Current line speed" becomes higher than the actual available throughput - and therefore traffic prioritization starts to fail.
  If the BT line IP profile did dynamically reduce to match short-term exchange congestion (and assuming that the update process was fast enough), then the congestion itself would reduce and the effect on customers actual throughput would improve, because the ISPs would slow down the rate at which they sent traffic, and the amount of dropped traffic would gracefully reduce - as opposed to the exponential meltdown which occurs with the current implementation.

[#13]

Quote from: ejs

the Plusnet profile makes the sustained throughput slightly slower than what it would be without the Plusnet profile

I agree that the Plusnet current line speed, for whatever reason, has always seemed to have been set about 100Kbps too low, according to speed test results.
I would also add a proviso, that I think it makes a difference both where and how you do the measurement, as to the way you perceive the results of the line profile / current line speed.
For example, my test setup -
  Plusnet (7.15Mbps profile) -->-- BT 20CN ADSL -->-- Modem -->-- Router -->-- PC
If I run any multithreaded speedtest on the PC and monitor the links between the modem, router, and the PC, I get the following results -

• 6.92Mbps : Modem LAN interface to router WAN interface


• 7.15Mbps : Router LAN interface to PC


• 6.86Mbps : PC speed test result

As my Plusnet "Current line speed" is 7.15Mbps, is it reasonable to expect what comes out of the router LAN interface to exactly match ?, and that the other results to be different due to protocol overheads, packet structures, and unoptimized MTU ?.
I would presume that the Plusnet end is effectively the same, but in reverse, so their corresponding LAN queue (for my connection) would be set at 7.15Mbps.
Because the LAN speed at both ends matches, it could also be assumed that the effective rate of LAN traffic at the exchange would also be the same, and matches my BT line IP profile, so achieving optimum throughput.
So in conclusion, I have Plusnet download traffic prioritization active, the traffic speed appearing from my router's LAN port matches EXACTLY both the Plusnet "Current line speed", and BT "IP profile",  I am not losing ANY performance due to the traffic management, and at the same time having my traffic prioritized for best user experience.

[#14]

I suspect the router LAN interface to the PC result is including the Ethernet layer that only exists between the PC and router.
Assuming the speedtest result is measuring the application layer throughput, I'll attempt to add on the TCP and IP layers (typically 20 byte header each) to get something comparable to the IP profile figure:
6.86*1500/1460
7.0479452054
6.86*1478/1438
7.0508205841
6.86*1430/1390
7.0574100719
(depending on the MTU)
If I attempt to also factor in a 14 byte Ethernet header:
6.86*1514/1460
7.1137260273
6.86*1492/1438
7.1176077885
6.86*1444/1390
7.1265035971
That was one of the things Plusnet came up with - the speedtester was measuring it wrongly, and if you look at the utilisation on a network interface while doing a download, it'll be higher. Of course looking there will give a higher result if it includes network layers that only exist between your computer and the router, headers which are added by your router and stripped off on arrival at your computer (or vice versa), and so aren't transmitted over the broadband connection. So you can measure the wrong thing, and see a higher result.
I think the Plusnet profile limiting your speed to less than what the line is capable of explains why attempting to optimise the MTU may make no difference (at least regarding avoiding ATM cell padding, in the downstream direction).
Apparently it knocks off 1 or 2 Mb from the highest FTTC speeds: http://forums.thinkbroadband.com/plusnet/t/4429133-pn-ip-profile-and-bt-ip-profile.html