Mine was seemingly only getting about 30 - 40Mbit/s just now (On an 80Mbit/s sync) but then again I've been playing with different routers over the last couple days so it's quite possible it was the router I have on the line at the moment.

Reconnecting PPP fixed it.

In reply to a post by dragon2611:

Mine was seemingly only getting about 30 - 40Mbit/s just now (On an 80Mbit/s sync) but then again I've been playing with different routers over the last couple days so it's quite possible it was the router I have on the line at the moment.

Reconnecting PPP fixed it.

Would reconnecting PPP feasibly be likely to fix a problem with the router you are using?

Edited by deleted (Wed 25-Mar-15 22:31:49)

I see no problem generally: My SamKnows box shows no variance in download speeds, upload speeds, or latency during peak hours compared with off-peak hours. The new tests for video show identical performance 24x7 for Netflix and iPlayer too.

Because that runs 24x7, it has many more results than I see from manual tests, and it is properly wired to the router.

But when I do run some manual tests, I can see some issues with the single-thread test. Sometimes. But my desktop isn't connected perfectly, so the issues could be my LAN.

Everything I read in the previous threads suggested that *something* was happening, and that it appeared to be restricted to some segments of the backhaul pipes. But the issue never seemed to hit many people at a time, so it is hard to say what is going on. It certainly wasn't enough to say that PN was systematically congested.

It wouldn't surprise me if PN's whizzo traffic-management edge routers are applying some different congestion avoidance techniques that lead to a controlled slowdown in places, rather than sudden packet loss.
http://en.wikipedia.org/wiki/Network_congestion#Prac...

Like many (most) others, I have been getting good speeds on all the tests I have made.
I tend to make speed tests during the day which seems to be less of a problem time, but Netflix/iPlayer have been working fine in the evening. Of course, that only uses a tiny proportion of the bandwidth.

On 40/20 connection, with potential on the line to go well above 70 down.

Edited by StephenTodd (Thu 26-Mar-15 09:16:27)

In reply to a post by Ignitionnet:

In reply to a post by dragon2611:

Mine was seemingly only getting about 30 - 40Mbit/s just now (On an 80Mbit/s sync) but then again I've been playing with different routers over the last couple days so it's quite possible it was the router I have on the line at the moment.

Reconnecting PPP fixed it.

Would reconnecting PPP feasibly be likely to fix a problem with the router you are using?

It probably was P/N but given i'd been playing with my network setup recently and didn't have time to test properly I don't want to say 100% that it's them as I can't be sure.

Current setup is as follows

Plusnet connects to a RB750 running routerOS, I was also playing with VRRP between it and an RB450G with the RB450G being the master, the default route on the RB450G points at the VM cable line so in order to force the traffic path to use plusnet I triggered a Mangle rule in the firewall to redirect traffic via an alternate routing table.

This means that the traffic path is RB450G > RB750 > Plusnet and also the RB750 isn't really an ideal router for an FTTC line as it's a little under powered for that (Depending on which features were being used).

I'd normally be using an EdgeRouter Lite but there seems to be a problem with the HW acceleration and the load-balancer, especially if you start using PBR to force certain traffic to a line and the device has a problem dealing with 150 + 80M and Policy routing connections in software routing mode.

Currently got a Linksys LRT224 on order, in a completely different class compared to the routers above, infact it's almost primitive in comparison but claims to do 900Mbit/s across the Nat and load balance, so it sounds like they've got the LB accelerated... Interestingly it seems to be a slower single core version of the CPU in the edgerouter (Cavium octeon)

It is quite random, one minute it is fine and the next its taking 1 to 2 minutes for a site to fully load or start to miss images and javascript etc. I've changed over to opendns servers to see if it was in anyway in part PN's dns servers but it isn't.

I will change back to the eci modem saturday to see if its anything to do with the g.imp update as hg612 modem stats is showing imp at 7 for my line.

Which gateway are you on?
http://usertools.plus.net/@gateway/

I find some behave that way and others are better.

It was either Tuesday or Wednesday I rang around 05:00 complaining about having video streaming from TWiT.tv regularly buffering when usually there was no problem, and speed down to about 2.5 or 3 Mbps instead of normally 10-11 Mbps on download.

Hope your fibre service was installed and you are seeing some benefit. My second line (with Post Office, so I have a dynamic IP too) went down on Friday and repair unlikely before Easter (at least they're replacing aluminium with copper, apparently).

Yep fibre installed and now flying with downloads around 66meg!! I'm the only one enabled at the moment on my cab until week after next so I'll enjoy the crosstalk-less connection whilst it lasts

Currently on bng02 with less issues

Seen issues on bng02,04, ptn-ag02,03