Network Problems?

[Steve]

Nothing at all. I also just tried to log in directly to the BT OpenReach router but can't connect. I've been waiting for over an hour for someone from IDNet support to ring me back. The first person said it would be within 15 minutes - an hour ago. Not impressed.

I have use the BT test server log-in previously to attempt to clear a stale session, whether it worked or it was coincidental with my next attempt at IDNet authentication I'm not sure.

Example here!
https://help.fasthosts.co.uk/app/answers/detail/a_id/1598/~/testing-your-connection-to-bt's-radius-server

That link is a bit slow in loading for me currently.

[nomad101]

IPv6 was working properly from my end after the outage, it was only the BQM that required the reconnect.

I don't know much about how these things work, but it was as though routing from me was fine, routing to me hadn't failed over properly.

Interesting.  Routing normally requires known routes in both directions, otherwise replies can't get back to you.  I suppose it depends on where the incoming traffic is coming from and whether routing tables along that specific route have stale entries.  It may work from some locations but not others.

I'm still getting my head around IPv6.  I've been working with IPv4 for 20+ years and now they go and change it all on me! 

[Bill]

I suppose it depends on where the incoming traffic is coming from and whether routing tables along that specific route have stale entries.

That gives me a thought- if it happens again I'll try disabling the IPv6 one for a few minutes, and see if it works after re-enabling it.

If I remember

[Bill]

Afterthought.

Interesting.  Routing normally requires known routes in both directions, otherwise replies can't get back to you.

Don't forget that there were multiple connections involved here, albeit all between me and the tbb server(s).

The one that stopped working had been initiated by tbb's pingbox and had been running for weeks before the outage. The other ones were initiated by me (viewing the graph, running the test), and hence were "shiny new" connections. Would that be relevant?

[psp83]

I get 10/10 for the readiness test, the connection one varies, but that could be the operating system (OS X in my case) not IDNet.

Check this link out, I quite often get it on the test-ipv6 site:

http://test-ipv6.com/faq_avoids_ipv6.html

It was working fine before the problem today, I have IPv6 address, just can't get nothing to ping it, its like IDnet system doesn't know who it belongs to.

[psp83]

ok, I just reconnected and get 9/10 on IPv6 now.

Fails on this though.

Your DNS server (possibly run by your ISP) appears to have no access to the IPv6 Internet, or is not configured to use it. This may in the future restrict your ability to reach IPv6-only sites.

[andrue]

One oddity- Everything was working fine, with the sole exception of my IPv6 BQM (other IPv6 tests passed as OK).

I had to disconnect/reconnect to get the BQM working again

Yah, I noticed that when I ran a speed test later on. I had to disconnect then reconnect my router to get IPv6 back. I've also noticed that it seems my connection is a bit better but perhaps that's because the network is unbalanced and I'm on a relatively empty node or somesuch.

I'm still getting my head around IPv6.  I've been working with IPv4 for 20+ years and now they go and change it all on me! 

Tell me about it. I'm going to try again this week to get my AAA record set up. I keep stumbling over what to use for the static IP address. Aside from being spoilt for choice (  ) I'm just not sure of the steps involved and it's hard to find concrete advice. But I've decided it's basically the same process just a lot more hex digits. 

[mervl]

No problems with IPv6 since the network recovered, though I've noticed it can be a bit hit and miss whether you connect via IPv6 or IPv4 to an enabled site. But I've not yet found any functional need for it!

I just wonder though whether some of the recent reports on here of higher latency were a sign of pending equipment failure (like the flickers which often precede a major blow-out on the electricity network), and as with humans often technology seems to get better just before it expires. Trouble I suppose is that as always you can only detect failure with certainty after it occurs, and in the modern "contracted" world the need for evidence precludes preventative action any more; and the other way is cheaper which is what matters to us all more than anything.

[brian_idnet]

Hi,

You can try testing using:

ping -6 ipv6.idnet.net

Regards,

Brian
IDNet

[sparkler]

hi does anyone know how to get bt's auto rate limiting thing disabled as since yesterdays fiasco im stuck at 444 / 9720 kbps

[psp83]

Is it me or has things slowed down again ?

[psp83]

never mind, seems to be back to normal now, just a blip!

[andrue]

Is it me or has things slowed down again ?

Seems okay for me at the moment:

http://www.thinkbroadband.com/speedtest/results.html?id=137184167721127156146

[psp83]

Seems okay for me at the moment:

http://www.thinkbroadband.com/speedtest/results.html?id=137184167721127156146

How do you get your green and yellow to match? mine is always 10 Mbps apart 

[andrue]

How do you get your green and yellow to match? mine is always 10 Mbps apart 

I'm not sure but if single threaded is below multi-threaded I think it suggests something can't keep up. Possibly a router that is overwhelmed, or maybe the clients itself. In effect it means there is a delay between the client sending a packet and the ACK from the server. Multi-threaded gets around that because while one thread is waiting for the 'yes, I got it' another can be sending its packet on the way.

What router do you have? Mine is a DLink-645 these days.

[psp83]

What router do you have? Mine is a DLink-645 these days.

Billion 7800n (gigabit ports)
Cat5e
PC NIC is gigabit as well

[andrue]

Billion 7800n (gigabit ports)
Cat5e
PC NIC is gigabit as well

Should be at least as good as mine then. Maybe having both lines together is actually bad? Suggesting that single threading maxes out my router and multi-threading gains nothing?

The leased line at work (installed last week) typically has the yellow zig-zagging either side of the green. That's a business-class Cisco router for whatever that's worth :-/

http://superuser.com/questions/261663/why-is-multi-thread-downloading-faster-than-single-thread

Which suggests that my graph is the way it should be..but I dunno

[psp83]

Should be at least as good as mine then. Maybe having both lines together is actually bad? Suggesting that single threading maxes out my router and multi-threading gains nothing?

The leased line at work (installed last week) typically has the yellow zig-zagging either side of the green. That's a business-class Cisco router for whatever that's worth :-/

http://superuser.com/questions/261663/why-is-multi-thread-downloading-faster-than-single-thread

Which suggests that my graph is the way it should be..but I dunno

I have no idea, would be good to know though, maybe support can enlighten us?

This is what I get at the mo.

[mervl]

Um, just out of interest on FTTC I get multi-thread downloads (which I believe are supposed to "max out" the connection) on the TBB Speedtester which are one-third of those on the single thread, consistently. (When actually in use the connection performs as per the single thread test according to the TBB monitor which I keep running on the PC, same with the Android test on the mobile phone using the home wifi). My second (non BT) connection also shows the yellow (multi-thread) line zig-zagging around the same more consistent level of the green (single thread) line. I'm a bit lost on that one, could it be a quirk of the routers QoS, perhaps?

[psp83]

could it be a quirk of the routers QoS, perhaps?

QoS is not enabled on mine, so I could really say, perhaps support knows?

What modems do you two have ( mervl & andrue ) ? I have an ECI modem.

[brian_idnet]

Hi,

Mine is always similar to yours Paul

http://www.thinkbroadband.com/speedtest/results.html?id=137198727219535584596

This usually runs at around 7-8Mbps difference between the 2 which I would expect there to be the difference between the threaded and multi-threaded tests.

I am currently using a Zyxel VMG1312-B Series VDSL router

Regards,

Brian
IDNet

[andrue]

I've not had much of a response from other forums but the consensus appears to be that multi-threading is a workaround for issues with the TCP stack. So it seems my connection is demonstrating the ideal. It might also be demonstrating throttling except that for my connection speed my results are very good. So if it proves anything it seems to prove that the bottleneck is my connection, as it should be.

But..I've only had two replies in two different fora so it's hardly conclusive. However my speeds do seem to be the maximum for my connection so it's either correct or else my single threaded downloads are unusually good 

[psp83]

Hi,

Mine is always similar to yours Paul

http://www.thinkbroadband.com/speedtest/results.html?id=137198727219535584596

This usually runs at around 7-8Mbps difference between the 2 which I would expect there to be the difference between the threaded and multi-threaded tests.

I am currently using a Zyxel VMG1312-B Series VDSL router

Regards,

Brian
IDNet

I suppose getting the single thread download to be the same as or as close to the multi-thread is best then.

Just means you can download quicker.

How are you finding that router? Since I can't unlock my modems without the chance of turning it into a paper weight, I've been thinking about getting a VDSL router so I can actually see the line stats just incase any future problems happen & if the modems BTo give me keep failing.

[psp83]

I've not had much of a response from other forums but the consensus appears to be that multi-threading is a workaround for issues with the TCP stack. So it seems my connection is demonstrating the ideal. It might also be demonstrating throttling except that for my connection speed my results are very good. So if it proves anything it seems to prove that the bottleneck is my connection, as it should be.

But..I've only had two replies in two different fora so it's hardly conclusive. However my speeds do seem to be the maximum for my connection so it's either correct or else my single threaded downloads are unusually good 

Have you tweaked your TCP/IP settings?

On ADSL2 I used to tweak the TCP/IP settings to get as much possible from my connection but I haven't bothered with FTTC and fresh install of Win8 Pro.

[andrue]

Have you tweaked your TCP/IP settings?

I played with them last year on and off but I don't think I saw much difference so gave up and returned them back to Windows 7 defaults. It'd be interesting to compare the values.

C:\Users\Andrue>netsh int tcp show global
Querying active state...

TCP Global Parameters
----------------------------------------------
Receive-Side Scaling State          : enabled
Chimney Offload State               : automatic
NetDMA State                        : disabled
Direct Cache Acess (DCA)            : disabled
Receive Window Auto-Tuning Level    : normal
Add-On Congestion Control Provider  : ctcp
ECN Capability                      : disabled
RFC 1323 Timestamps                 : disabled

Odd that I have NetDMA and DCA disabled but it's possible my laptop doesn't support them. Not that I'd expect them to make much difference given that it only has a 10/100 port (I only use Wifi when I have to - too many neighbours around me crippling throughput  )

MTU is 1500.