In reply to a post by lisaj:

When he left, we where synced at 42MB, now the strange thing is we always are, but hours after they leave our profile at BT always changes to 31.36MB

This part is likely to be a consequence of the way the cabinet's DLM system works.

When the BT engineer changes things, he'll get the operations centre to reset your line - which will reset DLM back to its original "go as fast as you can, and monitor for errors" setting.

DLM will monitor, and (presumably, in your case) actually detect a high error rate. It will then switch profiles to turn on interleaving and (as a consequence) the forward-error-correction process.

This, in turn, steals some of your bandwidth to transfer the extra parity information that allows the forward-error-correction process to work. It gives you a more stable connection, at the expense of raw speed. Worst case I've seen is that this overhead consumes around 23% of your original throughput.

If errors continue, then DLM can reduce your profile further, limiting the speed as well as turning interleaving on.

Ordinarily, DLM will make the first change around 48 hours after being reset. It can act faster than this (when slowing things down) if it detects too many errors. On the other hand it is *very* slow the other way around: If things are fixed, it can take from 3 weeks upwards to get the profile raised again.

I always think it is a bit of a swizz. The BT engineer can leave the property with the modem appearing to run far faster than it will be in 2 days - while the consumer can see the speed immediately, he *can't* see the error rate, so he thinks it is fixed. Meanwhile, when things go wrong in 48 hours, it's going to take days to get an engineer back again.

In reply to a post by lisaj:

Constant 19/20 milliseconds

Have you tried setting up a TBB "quality meter" from here?

This keeps a ping running all the time, which allows you to see if your line is suffering from packet loss. It can also tell you when a disconnection happens (if between 4AM and 8AM, there's a good chance that DLM changed the profile), and lets you see if latency changes (again, can be because DLM changed profile).

From my last comment, you might especially find it useful to see behaviour next time an engineer comes to fix things - and especially to see what happens in the 48 hours after they go!

For example, my first line had errors in the first 48 hours, showing around 5% packet loss. DLM intervened by adding interleaving - and I saw a disconnection, an increase in latency, and a decrease in packet loss. My line was reset, and went through an identical process the second time.

I'd post a link, but the historical data for the BQM meters seems to have disappeared, and I can't get the graphs any more.

But the tbb BQM is only designed for fixed IP addresses, not dynamic which PN ones are unless you've paid for fixed. (Maybe that's how you lost your old graphs?)

In reply to a post by WWWombat:

In reply to a post by lisaj:

Constant 19/20 milliseconds

Have you tried setting up a TBB "quality meter" from here?

This keeps a ping running all the time, which allows you to see if your line is suffering from packet loss. It can also tell you when a disconnection happens (if between 4AM and 8AM, there's a good chance that DLM changed the profile), and lets you see if latency changes (again, can be because DLM changed profile).

From my last comment, you might especially find it useful to see behaviour next time an engineer comes to fix things - and especially to see what happens in the 48 hours after they go!

For example, my first line had errors in the first 48 hours, showing around 5% packet loss. DLM intervened by adding interleaving - and I saw a disconnection, an increase in latency, and a decrease in packet loss. My line was reset, and went through an identical process the second time.

I'd post a link, but the historical data for the BQM meters seems to have disappeared, and I can't get the graphs any more.

Post edited to hide my embarrassment lol
Lisa

Edited by deleted (Thu 27-Sep-12 09:51:21)

Total red means your router is not responding to pings.

You probably need to log into the router and enable respond to Internet WAN ping/icmp

In reply to a post by MrSaffron:

Total red means your router is not responding to pings.

You probably need to log into the router and enable respond to Internet WAN ping/icmp

Whoops, just changed it in the router

Thanks Andrew

Lisa

That just means the router is giving priority to routing traffic and not responding to ICMP traffic on its local interface.

Nothing to panic about on its own, only an issue if we see pings varying further down the line.

Well, is it a thomson/technicolor device?

Ours is a technicolor router

L

Yes, the first hop enormous ping was a feature from the good old Speedtouch days. Glad to see it's still present now they're called Technicolor.

Presumably, you've tried rebooting modem and router?