Cutting Edge going blunt?

[Ignitionnet]

That's really rather odd.

Anyway pip how's Infinity rolling?

[Chrysalis]

ignition you think a docsis2 port at 78% utilisation can handle speed uplift or needs node split?

on my port they did a split of users and changed upstream from 10mbit to 20mbit, the 2nd 10mbit channel is gone and as was you told me merged into the 20mbit channel. So that was effectively a 4x upgrade and end result is 78% utilisation still on non uplifted speeds.

[pip08456]

Originally Posted by Ignitionnet

That's really rather odd.

Anyway pip how's Infinity rolling?

Actually I'm quite impressed with it. Connection is pretty much rock solid at 37.5 down and 8.5 up So far it's never dropped below 35Mb at peak times.

Ping, jitter and latency are much better than VM.

[Ignitionnet]

Originally Posted by Chrysalis

ignition you think a docsis2 port at 78% utilisation can handle speed uplift or needs node split?

on my port they did a split of users and changed upstream from 10mbit to 20mbit, the 2nd 10mbit channel is gone and as was you told me merged into the 20mbit channel. So that was effectively a 4x upgrade and end result is 78% utilisation still on non uplifted speeds.

78% is too high on its own IMHO Chrys.

[Chrysalis]

yeah, I raised the query after I seen my vmng300 graphs show what appears to be congestion at peak which for some reason didnt really show anywhere near as bad on the superhub, and 78% was the answer I got back, they claimed of course that was low and nothing to be worried about. 78% after a 4x is a bit of an ouch massive oversubbing before hand.

[pip08456]

[Chrysalis]

damn that virgin looks rough, massive difference to where you were back in sept 2010. arguably the uplifted speeds have wrecked your area.

[pip08456]

Originally Posted by Chrysalis

damn that virgin looks rough, massive difference to where you were back in sept 2010. arguably the uplifted speeds have wrecked your area.

I can do better than that!

[img]Download Failed (1)[/img]

[craigj2k12]

Originally Posted by pip08456

I can do better than that!

thats gotta be something LAN side, a download?

[pip08456]

Originally Posted by craigj2k11

thats gotta be something LAN side, a download?

As there was no PC connected to it at the time and the wireless is switched off I fail to see how.

[Chrysalis]

I admit pip the way it jumped up before 6am looks unusual for a congestion pattern. But is possible I guess at that moment of time VM flicked a switch somewhere that caused a sudden increase in utilisation.

[craigj2k12]

Originally Posted by Chrysalis

I admit pip the way it jumped up before 6am looks unusual for a congestion pattern. But is possible I guess at that moment of time VM flicked a switch somewhere that caused a sudden increase in utilisation.

could it possibly be someone else on the same port downloading?

[Ignitionnet]

Maybe the modulation fell back to QPSK.

---------- Post added at 13:32 ---------- Previous post was at 13:32 ----------

Originally Posted by Chrysalis

yeah, I raised the query after I seen my vmng300 graphs show what appears to be congestion at peak which for some reason didnt really show anywhere near as bad on the superhub, and 78% was the answer I got back, they claimed of course that was low and nothing to be worried about. 78% after a 4x is a bit of an ouch massive oversubbing before hand.

Sure that was an upstream figure?

[Chrysalis]

yes upstream.

---------- Post added at 14:31 ---------- Previous post was at 14:30 ----------

Originally Posted by craigj2k11

could it possibly be someone else on the same port downloading?

I wouldnt expect that kind of jump form a single person starting a download.

I think ingition's explanation seems most likely a fallback to QPSK. As that would instantly halve upstream capacity.