In reply to a post by fuzed:

yeah I get that, a yes I did get a bit OTT on the tests, but they aren't going to cause any significant impact on a network.

They will if several users are doing it. Particularly if on the same BT Wholesale node/MSIL. Even the high quality AA/Zen/IDNets of this world don�t plan their capacity for that.

In my Entanet example, they literally had to put out a message on these forums and to their resellers to pass on that they certainly had a problem, (not in fact of their own making I need to point out), but the hundreds of end users doing frequent speed tests were making the problem worse, with the traffic exceeding the real stuff.

It doesn�t take many concurrent/overlapping speedtests along with a few film streams to fill the rented capacity on a shared MSIL.

Edited by RobertoS (Wed 21-Feb-18 09:14:49)

In reply to a post by Philce:

We have been told numerous times that the network is not even running at 50% capacity within the cloudscape network, what this actually means is open to interpretation. A few users doing speed tests shouldn't impact a non contended service.

Let's review:

1. As I understand it, OP is now on a TTB backhaul
2. ISP says its own network has plenty of headroom, which we must assume includes the practical, sustained bandwidth of its routers (not just the raw links)
3. Migrating backhaul shows a distinct change in behaviour, but it still exhibits congestion at times

It sounds to me that OP has been provisioned with a low TTB traffic priority -- perhaps the same low priority that TTR gives its users. If that's the case, OP and ISP should discuss whether the service has been correctly provisioned, given the class of service OP has purchased and OP's reasonable expectations.

If they decide it's been correctly provisioned, and if ISP really wants to get to the bottom of this, it could temporarily increase the TTR priority for this connection -- to the sort of level that Pulse8 and A&A use. This would allow OP to gather useful speedtest data for comparison.

What are your actual issues with everyday use of your connection. You should be able to view a couple of concurrent HD streams during the worst time and pushing 2 x 4K at the best.

Edited by broadband66 (Wed 21-Feb-18 09:09:10)

It'll be something I'll discuss with them if the connection gets any worse than it is.

The issues were before the move that the connection was so poor that streaming netflix, youtube etc was slow, none existent or very poor quality, less than SD quality at times.

due to all the issues I've had with the service I'm keeping an eye on the connection to see how and what the issues are, at the end of the day I'm paying these guys more than I would for a cheap service, which is what I was expecting, not a service that is below par.

Some may think 15mb down at peak is ok, it might be for a short period, maybe a few seconds, but not all the way through the peak times. I'm keeping an eye on the service and will calm down with the speed tests, but I just want to make sure I don't have the same issues I've had since I moved to Cloudscape in November.

Hi Richi,

Not sure where you are getting this from. None of our users are on TTB for FTTC. The OP and the others in this discussion have been moved from Vodafone to BTW backhaul.

So please do not make wild speculations regards provisioning, traffic priority etc.

We are working with these users and the responsible carriers to get this resolved.

We are already gathering useful data and this has been passed back to the carriers involved.

The issue with Philce is unique, and being addressed separately.

Hi Ben,

As you acknowledge you change over on 27/2/18

We know there is an issue with your Voda cct.

Firstly, I misread a post upthread that seemed to imply TTB, rather than BTW. Apologies for the confusion (but my advice to OP still applies).

Secondly, the ISP concerned has confirmed my opinion that I would not choose to use it, given the attitude of its formal presence on this forum. To misquote Auric Goldfinger: Once is an accident, twice is a coincidence, three times is policy.

Can you define what you actually mean by several?

e.g. would 3 people with IDNet be able to DoS capacity for other users?

It is impossible to define for any given ISP. The only facts we have are the problems Entanet encountered after the BT Wholesale mind change immediately before the implementation of Entanet's full 20-node network, and the intractable congestion problems some Plusnet users experienced due to Plusnet being on shared MSILs rather than dedicated ones.

Large providers with hundreds of thousand end users would have a large "several", as we saw from Entanet's appeals to lower the volume of speed tests. That being in the days of mere ADSL2+ capacity needs and provision.

Small providers "several" would depend on the number of users they have and the routing of customers from different parts of the country to whatever node of whichever wholesaler they are allocated to, and the type of node, shared/dedicated, and the way small provider's throughput capacity is handled by the wholesaler.

In the case of Cloudscape and it's backhaul suppliers, we have no idea what capacity on which MSILs they have contracted for. As you will know, but some readers may not, shared MSILs mean that other ISPs' traffic can also influence Cloudscape's throughput.

I was merely making the point that it is possible that numbers of end users running a high frequency of speedtests with the throughput volumes those incur under FTTC could have an impact on an MSIL or even a GEA cable-link.