HUGE Latency What can be done to lower this

[jona_b]

here's pic

like thread title says what can be done to lwoer this

[cybernetic_tiger]

Firstly run a tracert to the BBC to see where the latency is, then post your results but remember to remove your IP address and reverse DNS from the trace,

[jona_b]

ok removed my ip

[cybernetic_tiger]

Thanks for that,

Are you connected to the cable modem via USB or Ethernet?
How long has this been going on for?
Do you have any other programs running, possibly provide a "netstat -a"?
Can you please ping 192.168.100.1 and post the round trip times?

Thanks,

[jona_b]

ok connecting through ethernet

btw had this problem over 11months evening pings sky rocket early hours of morning very low as 9 ping to bbc

told tecnical support no end of times they just keep saying it willbe fixed at the end of the month, they it repeats i phone at the end of the month and they say teh same be fixed at teh end of this new month

[cybernetic_tiger]

Well from all of that it looks like its likely to be an issue between your modem and the uBR. This will more than likely be caused by either upstream traffic or an SNR related issue.

Either way if it's been going on for a while I would suggest you call India because your PC looks fine and the latency is after the modem. Are you an ex-Telewest or ex-NTL customer?

[jona_b]

ex-blueyonder

are signals good?

[cybernetic_tiger]

As your ex-BY if you would like me to take a look PM me your @blueyonder.co.uk e-mail address and I will check out your modem and the HFC in your area,

EDIT: Your levels are fine however the upstream SNR may not be

[jona_b]

pm sent

[cybernetic_tiger]

pm sent

I have just confirmed your address to you via PM, please reply to confirm/deny that this is you, thanks.

[jona_b]

pm replied

[cybernetic_tiger]

I have taken a look and this is due to utilisation of the upstream channel. As these channels are now all load-balanced if I changed your upstream it would re-negotiate back within a few hours.

I have checked and these upstreams (and the downstream channel) are marked as black meaning that they require upgrading urgently; Black is as bad as it gets.

At least you know that there is nothing you can do about this, you will need to keep chipping away at tech support...

Hope this helps,


ubr02wolv#sh cable modem 0014.####.####
MAC Address IP Address I/F MAC Prim RxPwr Timing Num BPI
State Sid (dB) Offset CPE Enb
0014.####.#### 10.231.###.### C5/0/U2 online 4147 4.50 2045 1 N

ubr02wolv#sh cable modem 0014.####.#### QOS
Sfid Dir Curr Sid Sched Prio MaxSusRate MaxBrst MinRsvRate Throughput
State Type
13869 US act 4147 BE 1 384000 1522 0 5071
14758 DS act N/A BE 0 4096000 3044 0 52717

ubr02wolv#pi ip 10.231.###.### repeat 100 size 1420

Type escape sequence to abort.
Sending 100, 1420-byte ICMP Echos to 10.231.###.###, timeout is 2 seconds:
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!.!!!!!!!!!!!!!!!.!. !!!!!!!!!!!!!!!!!!!!
!!!!!!.!!!!!!!.!!!!!!!!.!!!!!!
Success rate is 94 percent (94/100), round-trip min/avg/max = 32/60/220 ms
ubr02wolv#sh contr C5/0 | inc SNR
US phy SNR_estimate for good packets - 27.2499 dB
US phy SNR_estimate for good packets - 28.6620 dB
US phy SNR_estimate for good packets - 28.3574 dB <<<<<<<<<< Your Upstream
US phy SNR_estimate for good packets - 28.2930 dB
ubr02wolv#sh int c5/0 mac-sche | inc util
Avg upstream channel utilization : 72%
Avg upstream channel utilization : 60%
Avg upstream channel utilization : 82% <<<<<<<<<< Your Upstream
Avg upstream channel utilization : 55%
ubr02wolv#

[jona_b]

so is there eny approx date for when this will be fixed

[cybernetic_tiger]

There are currently 150 cables in "Black" status and not all of these have upgrades scheduled.

I have taken a look and it looks like upgrades should be complete by 25/01/2008 on your cable however :

A) This is a long way off and priorities may change
B) I don't work in planning so could not confirm this date

You will need to chase tech support so that they chase the planners to expedite your case,

regards,

[jona_b]

ok well appreciated what u done for me thnx

[ECW_Original]

Sorry to hijack this, but I am also on UBR02, and was told a good 4-5 months ago that this is over-subscribed, could this be the case?

Also as this has been going on for what was it, 11 months? surely VM should know by now to do something about it...

[Stuart]

Sorry to hijack this, but I am also on UBR02, and was told a good 4-5 months ago that this is over-subscribed, could this be the case?

Also as this has been going on for what was it, 11 months? surely VM should know by now to do something about it...

You may both be on UBR2, but you are in different areas. Indeed, you are also on different networks (yours is ex-ntl, jona_b's is ex-telewest). There are multiple UBRs in each area, and each area follows a similar numbering convention. So, there is a UBR02 in each area.

However, it does NOT follow that issues on a UBR in one area will also occur on the same numbered UBR in another area.

[evilmonkey]

I have taken a look and this is due to utilisation of the upstream channel. As these channels are now all load-balanced if I changed your upstream it would re-negotiate back within a few hours.

I have checked and these upstreams (and the downstream channel) are marked as black meaning that they require upgrading urgently; Black is as bad as it gets.

At least you know that there is nothing you can do about this, you will need to keep chipping away at tech support...

Hope this helps,


ubr02wolv#sh cable modem 0014.####.####
MAC Address IP Address I/F MAC Prim RxPwr Timing Num BPI
State Sid (dB) Offset CPE Enb
0014.####.#### 10.231.###.### C5/0/U2 online 4147 4.50 2045 1 N

ubr02wolv#sh cable modem 0014.####.#### QOS
Sfid Dir Curr Sid Sched Prio MaxSusRate MaxBrst MinRsvRate Throughput
State Type
13869 US act 4147 BE 1 384000 1522 0 5071
14758 DS act N/A BE 0 4096000 3044 0 52717

ubr02wolv#pi ip 10.231.###.### repeat 100 size 1420

Type escape sequence to abort.
Sending 100, 1420-byte ICMP Echos to 10.231.###.###, timeout is 2 seconds:
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!.!!!!!!!!!!!!!!!.!. !!!!!!!!!!!!!!!!!!!!
!!!!!!.!!!!!!!.!!!!!!!!.!!!!!!
Success rate is 94 percent (94/100), round-trip min/avg/max = 32/60/220 ms
ubr02wolv#sh contr C5/0 | inc SNR
US phy SNR_estimate for good packets - 27.2499 dB
US phy SNR_estimate for good packets - 28.6620 dB
US phy SNR_estimate for good packets - 28.3574 dB <<<<<<<<<< Your Upstream
US phy SNR_estimate for good packets - 28.2930 dB
ubr02wolv#sh int c5/0 mac-sche | inc util
Avg upstream channel utilization : 72%
Avg upstream channel utilization : 60%
Avg upstream channel utilization : 82% <<<<<<<<<< Your Upstream
Avg upstream channel utilization : 55%
ubr02wolv#

82% is not high enough to cause them ping times. check FEC errors and CPU on the card.

[cybernetic_tiger]

82% is not high enough to cause them ping times. check FEC errors and CPU on the card.

Actually this was bursting to 99% causing the drops. FEC errors are common on the HFC network and as always in busy areas are clocking. uFEC errors are also clocking and accounting for some of the drops. CPU is fine.

Just because I didn't post everything doesn't mean I didn't check it!

[evilmonkey]

Actually this was bursting to 99% causing the drops. FEC errors are common on the HFC network and as always in busy areas are clocking. uFEC errors are also clocking and accounting for some of the drops. CPU is fine.

Just because I didn't post everything doesn't mean I didn't check it!

Its an MC28U card so you shouldnt see any Uncorrected FEC or SNR drops unless the card is not properly configured/installed. Thing being is these sorts of outages never get fixed properly due to "engineers" not knowing how to check the cards correctly.