100M : Plymouth PL3 broken in evenings since 29th Aug

[hd1080ts]

Since last month VM connectivity has fallen apart in the evenings.
Last night got a ping of 538ms using Speedtest.net to the VM server in Bristol.

Called VM and was told there has been an issue since 29th August and will be fixed 25th September, so seems to be linked to the 120Mbps upgrade.

CS credited a months Internet charges to my account.

So is this bad upgrading procedure for 120Mbps or did something break?

The other theory is the students are back before the connectivity upgrade.

H

[Kushan]

For it to take a month to fix, it's probably congestion of some kind. I doubt it's related to the upgrades. All I would say is keep an eye on that date, these ones have a habit of sliding for weeks.

[kwikbreaks]

The other thing I noticed about congestion (apart from the ever sliding fix time) was that the onset can be very fast indeed. One day seems to be OK, the next pretty poor, the next almost unusable. I'm not sure if this is still the case with 6 or 8 downstreams but it was certainly true a couple of years or so back when I was on 50Mbps. At that time I just downgraded to 10Mbps as a lot of the time that was all I was getting.

[hd1080ts]

Originally Posted by kwikbreaks

One day seems to be OK, the next pretty poor, the next almost unusable.

That's what we had, at first I blamed the kids then it got worse and worse and now evenings it's pretty much unusable.

[Kushan]

Unfortunately that makes sense, due to the nature of it. If the area hits 95% usage, you won't notice because you'll still get full speed when you need it, but as soon as it hits total utilisation, packets get dropped and speeds decline sharply.

[hd1080ts]

At 3pm the connection completely died.

VM CS ran some tests and they are sending an engineer out on Thursday.

Posted via tethered phone.

H

Cable Modem Status
Item Status Comments
Acquired Downstream Channel (Hz) 459000000 Hz Locked
Ranged Upstream Channel (Hz) 45800000 Hz Success
Provisioning State FAILED Offline

log
First Time Last Time Priority Error Number Description
Time Not Established Time Not Established Warning (5) 82000900 B-INIT-RNG Failure - Retries exceeded;CM-MAC=4c:60:de:18:3c:30;CMTS-MAC=00:1d:70:cc:89:63;CM-QOS=1.0;CM-VER=3.0;
Time Not Established Time Not Established Critical (3) 82000300 Ranging Request Retries exhausted;CM-MAC=4c:60:de:18:3c:30;CMTS-MAC=00:14:f1:e2:38:01;CM-QOS=1.0;CM-VER=3.0;
Time Not Established Time Not Established Critical (3) 82000200 No Ranging Response received - T3 time-out;CM-MAC=4c:60:de:18:3c:30;CMTS-MAC=00:14:f1:e2:38:01;CM-QOS=1.0;CM-VER=3.0;
Time Not Established Time Not Established Critical (3) 84000100 SYNC Timing Synchronization failure - Failed to acquire QAM/QPSK symbol timing;;CM-MAC=4c:60:de:18:3c:30;CMTS-MAC=00:00:00:00:00:00;CM-QOS=1.0;CM-VER=3.0;
Time Not Established Time Not Established Critical (3) 82000200 No Ranging Response received - T3 time-out;CM-MAC=4c:60:de:18:3c:30;CMTS-MAC=00:1d:70:cc:89:63;CM-QOS=1.0;CM-VER=3.0;
10/08/2013 14:00:05 10/08/2013 14:00:05 Critical (3) 82000400 Received Response to Broadcast Maintenance Request, But no Unicast Maintenance opportunities received - T4 time out;CM-MAC=4c:60:de:18:3c:30;CMTS-MAC=00:1d:70:cc:89:63;CM-QOS=1.1;CM-VER=3.0;

[kwikbreaks]

Originally Posted by Kushan

If the area hits 95% usage, you won't notice because you'll still get full speed when you need it, but as soon as it hits total utilisation, packets get dropped and speeds decline sharply.

Which begs the question why VM aren't pro-actively upgrading areas before they get to the point where hundreds of customers are waiting maybe months to get an acceptable connection...

---------- Post added at 17:07 ---------- Previous post was at 17:05 ----------

Originally Posted by hd1080ts

VM CS ran some tests and they are sending an engineer out on Thursday.

Let's hope it was a fault and not congestion then so when you can connect you get an acceptable connection.

[hd1080ts]

Just got an automated message from VM cancelling engineer visit saying it's a network fault and engineers are fixing.

Estimated fix 8pm this evening.

---------- Post added at 17:31 ---------- Previous post was at 17:28 ----------

Originally Posted by kwikbreaks

]Let's hope it was a fault and not congestion then so when you can connect you get an acceptable connection.

Fingers crossed, my Three 3G phone connection feels faster than what I was getting.

[Sephiroth]

I wonder whether or not they had some sort of backhaul failure. Unless I'm mistaken, PL3 has a long backhaul to Exeter (can be verified in a traceroute). If something went amiss there, the remaining capacity would be under strain. For example, they are terminating all of PLx onto, say EXE2.

If you search the VM forums, PL3 has a long history of troubles.

EDIT: Just seen the above post. More or less fits what I said - but VM will never tell us exactly what it was.

[Kushan]

Originally Posted by kwikbreaks

Which begs the question why VM aren't pro-actively upgrading areas before they get to the point where hundreds of customers are waiting maybe months to get an acceptable connection

Well, the problem goes back years and years. Back in the NTL days, the network started to come under huge strain but NTL weren't exactly making money - so how could they fix anything? Instead, implement obscene STM to keep things ticking over, outsource the call centres and hope for the best.

The problem is that the areas just got worse and worse, to the point where VM now has a horrendous reputation. Recently (last year or two) they have been making an effort to fix those broken areas but now you have a choice - do you fix the broken areas, areas that have had shoddy service for months or years, or do you proactively fix the areas that aren't quite ready to fall over but aren't far off? The cost in either case is much the same...

So I believe that's what Virgin is doing, concentrating on the worst areas first, the areas that have had issues the longest and are making their way down the priority chain until (hopefully) they start being proactive.

I know the criteria for even raising a fault used to be 99% usage for more than 10% of the time, then that dropped to (I think) 95% and I think it has dropped again. From looking at what Virgin used to be like (75% STM for 5 hours) to what they're like now (40% max for 2 hours) and given that less people seem to be stuck in congestion hell, I think they are slowly getting better. They still have a lot of work to do but stuff is being done, at least.

[kwikbreaks]

I think things probably are improving. I haven't looked through the TBB monitor thread for quite some time but the charts I have seen posted from time to time really don't look as colourful as they were a year or so back.

I've gambled on a reasonable year to come for me despite the SamKnows reports suggesting there is at least moderate congestion in my area by accepting what can only be called a bribe to take another 12 month minimum term.

[hd1080ts]

Engineers on site, estimated fix just slipped to 11:35pm.

[hd1080ts]

Update.

Total loss of service fixed, but evenings got worse, more engineer visits no fix in sight!

[Sephiroth]

Students must be back in Plymouth.

[hd1080ts]

Originally Posted by Sephiroth

Students must be back in Plymouth.

But its been bad since August, before the students came back.

Thinking of switching to BT Infinity 2.