A large several does not exist, it is simply two or more.

So better to say things like if 30% of a providers customers speed test all at the same time then problems can occur.

In scenarios where someone is hitting run test, just 1 second after the completion of another then I'd say that was unfair, but then you also have to also say anyone with a big Steam download or other high thread action is also unfair.

That "30%" is a completely arbitrary figure that cannot be justified. Nor does queuing theory work the way you seem to imagine, particularly given the variety of pinch-points I cited, which are only a subset of those that can exist.

Every ISP's setup is different and a generalisation in percentage terms is nonsense. For instance three or four users with 80Mbps sync on the same wholesaler on an exchange where that wholesaler only has a 1Gbps GEA Cable link, speedtesting while some others are using multiple Netfix streams each, is a totally different case from the same combination on a 10Gbps link.

As I also pointed out, if even Plusnet could run out of capacity on some shared MSILs, how much more so could Cloudscape?

I suggest continuing the discussion is pointless. Whatever you or I say doesn't alter the fact OTT speedtesting helps no-one, and certainly adds to contention somewhere. The point where it converts to congestion is far too complex for here. I don't understand why you raised it.

Cloudscape's webby claims

"All our fibre products utilise our state of the art network which provides 1:1 contention with no speed throttling or packet shaping so you get 100% of the speed 100% of the time. "

So they should be able to handle customers using a speed tester. As should any provider out there without any problems.

So if you wish to speed test your connection I say test away. However I suspect getting out of the contract and moving along to another provider (preferably not another reseller) would be a better use of time.

yeah that may be great for the connection between me and the provider, but anything after that isn't true... the sad thing is this is what I joined them for, to have a service that was not contended to the levels other providers are/were contended.

Maybe it's time for them to reword that bit on their webpage!!!!

They don't specify the boundaries of their network anyway. Their interconnects may be fine, which is the far side from you, and their internal routing may have (at the moment) capacity for all customers fully using their sync, but as per my replies to MrSaffron there is a lot that can go wrong between your router and your/anyone's traffic reaching their first in-house router.

The furthest in your direction they can reach on BT Wholesale is probably these MSILs I have mentioned, where they may simply have a slot on one or two but the MSIL(s) also handling traffic from other ISPs and simply not cope.

See this explanation.

How Vodafone traffic is handled once it leaves your exchange I have no idea, sorry.

Edited by RobertoS (Wed 21-Feb-18 21:49:48)

How are you getting on with this now?

My migration took place today, here is the first result,



Not setting that single thread speed on fire (still looks like an issue to me, @MrSaffron, you opinion?), but multi thread speed is ok at the moment, the acid test comes at 8-11 though I suppose, so we'll see!

Mine went through last week, seeing similar, they are still investigating with their carrier.
Looks like there is something else wrong somewhere.

Although last night we were back seeing this.

In reply to a post by Ben88:

How are you getting on with this now?

My migration took place today, here is the first result,



Not setting that single thread speed on fire (still looks like an issue to me, @MrSaffron, you opinion?), but multi thread speed is ok at the moment, the acid test comes at 8-11 though I suppose, so we'll see!

Edited by Philce (Tue 27-Feb-18 15:43:52)

If the shape of tests is similar all the time, points towards something limiting to around 20 Mbps per thread.

A high start with a relatively smooth drop off is often a sign of some throttling, but could be many things of course

Happy to report that the most serious congestion issues seem to be resolved for me now, here is my BQM from yesterday:

My Broadband Ping

Compared to the same day last week...

My Broadband Ping

Speedtests were generally in the 65+ region for multi threaded activities consistently last night, which is a massive improvement obviously.

I have done some tests for the single thread issue and passed them back to the team at CloudScape for further investigation, so hopefully they help them in resolving it.

I find the results vary quite wildly on DSLReports, I can run 4 single thread tests and see one at 15, one at 22, and then two 40-45. TBB tester consistently puts single thread in 20-25 region.

I migrated a few weeks back, and to be fair It's been pretty good, I've not seen any speed drops below 14MB and that was ONLY the once, the rest of the time I've had pretty good speeds 70ish mb on single thread tests, so touching wood that it stays like this, but I'm happier!

Down HTTPx1 Avg Date
34970 - 28/02/2018 20:53
69420 - 27/02/2018 20:46
69500 - 27/02/2018 20:38
68840 - 27/02/2018 19:59
68360 - 27/02/2018 19:12
69450 - 27/02/2018 19:04
69010 - 26/02/2018 20:36
69590 - 26/02/2018 20:25
66590 - 26/02/2018 11:04
69480 - 24/02/2018 17:29
69470 - 24/02/2018 17:07
69590 - 23/02/2018 23:02
68560 - 23/02/2018 22:49
66890 - 23/02/2018 20:58
68920 - 23/02/2018 20:51
69830 - 22/02/2018 21:52
63110 - 22/02/2018 21:46
58420 - 22/02/2018 20:32
58630 - 22/02/2018 20:05
64540 - 21/02/2018 22:09
58100 - 21/02/2018 22:02
58220 - 21/02/2018 21:57
68300 - 21/02/2018 20:05
69680 - 20/02/2018 23:41