Insane pings through UUNet

F

fragtion

Expert Member
Joined
Dec 26, 2004
Messages
2,899
Reaction score
195
Location
Cape Town
Ok guys i've been testing the UUNet adsl system, and while the pings have been incredibly unstable (local pings of up to 1500ms at times), they've been incredibly low too...

[root@fragland root]# traceroute jolt.co.uk
traceroute to jolt.co.uk (195.149.21.11), 30 hops max, 38 byte packets
1 196-47-3-1.access.uunet.co.za (196.47.3.1) 16.500 ms 9.425 ms 26.596 ms
2 196.30.9.109 (196.30.9.109) 482.831 ms 274.681 ms 79.402 ms
3 pos8-1-0.gw11.jnb6.alter.net (196.30.121.157) 129.905 ms 203.498 ms 177.437 ms
4 fe4-1-0.cr2.jnb6.alter.net (196.30.156.34) 98.233 ms 106.509 ms 91.160 ms
5 Serial5-0-1.GW3.LND10.Alter.NET (146.188.5.217) 271.696 ms 118.982 ms 129.916 ms
6 ge5-0.cr1.lnd10.gbb.uk.uu.net (158.43.150.1) 99.686 ms 277.178 ms 271.671 ms
7 pos1-0.cr1.lnd6.uk.gbb.uu.net (158.43.253.129) 268.992 ms 58.658 ms 69.768 ms
8 * * *
9 lon1-9.nildram.net (195.66.224.59) 270.671 ms 209.872 ms 84.295 ms
10 jolt-gw.nildram.net (195.149.20.126) 95.273 ms 260.401 ms 456.342 ms
11 clarity.jolt.co.uk (195.149.21.11) 298.138 ms 59.745 ms 59.915 ms


^^ is it even possible to reach 59 ms to the UK from ZA ?


[root@fragland root]# ping jolt.co.uk
PING jolt.co.uk (195.149.21.11) 56(84) bytes of data.
64 bytes from clarity.jolt.co.uk (195.149.21.11): icmp_seq=1 ttl=47 time=272 ms
64 bytes from clarity.jolt.co.uk (195.149.21.11): icmp_seq=2 ttl=47 time=100 ms
64 bytes from clarity.jolt.co.uk (195.149.21.11): icmp_seq=3 ttl=47 time=278 ms
64 bytes from clarity.jolt.co.uk (195.149.21.11): icmp_seq=4 ttl=47 time=119 ms
64 bytes from clarity.jolt.co.uk (195.149.21.11): icmp_seq=5 ttl=47 time=210 ms
64 bytes from clarity.jolt.co.uk (195.149.21.11): icmp_seq=6 ttl=47 time=271 ms
64 bytes from clarity.jolt.co.uk (195.149.21.11): icmp_seq=7 ttl=47 time=119 ms
64 bytes from clarity.jolt.co.uk (195.149.21.11): icmp_seq=8 ttl=47 time=266 ms
64 bytes from clarity.jolt.co.uk (195.149.21.11): icmp_seq=9 ttl=47 time=271 ms
64 bytes from clarity.jolt.co.uk (195.149.21.11): icmp_seq=10 ttl=47 time=105 ms
64 bytes from clarity.jolt.co.uk (195.149.21.11): icmp_seq=11 ttl=47 time=316 ms

--- jolt.co.uk ping statistics ---
11 packets transmitted, 11 received, 0% packet loss, time 10094ms
rtt min/avg/max/mdev = 100.306/212.020/316.947/79.781 ms

edit
Just did a test thru my windows box and the results seem much more 'realistic'...

Tracing route to clarity.jolt.co.uk [195.149.21.11]
over a maximum of 30 hops:

1 16 ms 15 ms 15 ms 196-47-3-1.access.uunet.co.za [196.47.3.1]
2 26 ms 29 ms 21 ms 196.30.9.109
3 19 ms 21 ms 19 ms pos8-1-0.gw11.jnb6.alter.net [196.30.121.157]
4 20 ms 16 ms 24 ms fe4-1-0.cr2.jnb6.alter.net [196.30.156.34]
5 272 ms 288 ms 274 ms Serial5-0-1.GW3.LND10.Alter.NET [146.188.5.217]

6 269 ms 269 ms 273 ms ge5-0.cr1.lnd10.gbb.uk.uu.net [158.43.150.1]
7 272 ms 268 ms 266 ms pos1-0.cr1.lnd6.uk.gbb.uu.net [158.43.253.129]
8 * * * Request timed out.
9 268 ms 272 ms 268 ms lon1-9.nildram.net [195.66.224.59]
10 266 ms 275 ms 263 ms jolt-gw.nildram.net [195.149.20.126]
11 268 ms 314 ms 319 ms clarity.jolt.co.uk [195.149.21.11]

Trace complete.
 
Last edited:
Well I've had sub 200 ms pings to the UK on ISDN. A quick word though, uunet's product is still in testing, it'll only be released at the end of this month (that's what I've been told anyway). I have seen some issues connecting to SAIX and IS though.
 
Well i get 150'ish ping to jolt from my friend's box at UCT, no idea if that goes through SAT3 or not though, but the point is 100~ms ping is definately possible, telkom's infrastructure is what's causing us the pathetic pings. (I average 350-600 ms to jolt with std. telkom adsl shaped/uncapped)...

-bash-2.05b$ traceroute jolt.co.uk
traceroute to jolt.co.uk (195.149.21.11), 64 hops max, 40 byte packets
1 nortel8610.uct.ac.za (137.158.36.1) 4.492 ms 1.598 ms 1.582 ms
2 e0-pyroteichos.freebsd.ab-e-3-30.uct.ac.za (137.158.128.5) 0.387 ms 0.446 ms 0.384 ms
3 e14-tenet.cisco.ab-f-3-40.uct.ac.za (137.158.200.6) 2.109 ms 3.671 ms 2.115 ms
4 uct.int.uni.net.za (155.232.50.246) 7.025 ms 7.001 ms 2.937 ms
5 wblv-er-2-fe-2-0-0.telkom-ipnet.co.za (196.25.134.65) 4.409 ms 5.724 ms 3.322 ms
6 lon-ip-dir-telecity-pos-6-0.telkom-ipnet.co.za (196.43.9.50) 169.439 ms 162.736 ms 167.197 ms
7 pos2-1.cr02.ldn01.pccwbtn.net (63.218.52.17) 157.703 ms 151.862 ms 152.825 ms
8 ge2-2.br02.ldn01.pccwbtn.net (63.218.12.94) 191.936 ms 206.542 ms 281.023 ms
9 lon1-9.nildram.net (195.66.224.59) 157.314 ms 152.061 ms 174.635 ms
10 jolt-gw.nildram.net (195.149.20.126) 154.264 ms 155.378 ms 153.596 ms
11 clarity.jolt.co.uk (195.149.21.11) 155.443 ms 152.265 ms 155.212 ms
-bash-2.05b$
 
I wish I had 350 ms

I currently get 450 ms off peak & 550ms during peak times this is on ADSL384 shaped.

What bugs me is I used to get 350 ms last year when I checked, anyone else seeing similar ?

My skype calls are really bad quality of late, they were better last year (skype reduces qulaity to try to accomodate the latency)
 
The following is a ping directly from uunet to jolt.co.uk , this should be the lowest ping possible for uunet.

Im wondering how you guys going through uunet getting lower ping than uunet? lol :)

PING jolt.co.uk (195.149.21.11) from 196.7.147.27 : 56(84) bytes of data.
64 bytes from clarity.jolt.co.uk (195.149.21.11): icmp_seq=0 ttl=49 time=249.671 msec
64 bytes from clarity.jolt.co.uk (195.149.21.11): icmp_seq=1 ttl=49 time=249.368 msec
64 bytes from clarity.jolt.co.uk (195.149.21.11): icmp_seq=2 ttl=49 time=248.554 msec
64 bytes from clarity.jolt.co.uk (195.149.21.11): icmp_seq=3 ttl=49 time=248.566 msec
64 bytes from clarity.jolt.co.uk (195.149.21.11): icmp_seq=4 ttl=49 time=247.899 msec

--- jolt.co.uk ping statistics ---
5 packets transmitted, 5 packets received, 0% packet loss
round-trip min/avg/max/mdev = 247.899/248.811/249.671/0.836 ms
 
Unshaped ADSL to Jolt
Pinging clarity.jolt.co.uk [195.149.21.11] with 32 bytes of data:

Reply from 195.149.21.11: bytes=32 time=253ms TTL=50
Reply from 195.149.21.11: bytes=32 time=252ms TTL=50
Reply from 195.149.21.11: bytes=32 time=253ms TTL=50
Reply from 195.149.21.11: bytes=32 time=254ms TTL=50

Ping statistics for 195.149.21.11:
Packets: Sent = 4, Received = 4, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
Minimum = 252ms, Maximum = 254ms, Average = 253ms

ISDN 64K via DataPro / IS
Pinging clarity.jolt.co.uk [195.149.21.11] with 32 bytes of data:

Reply from 195.149.21.11: bytes=32 time=180ms TTL=53
Reply from 195.149.21.11: bytes=32 time=180ms TTL=53
Reply from 195.149.21.11: bytes=32 time=180ms TTL=53
Reply from 195.149.21.11: bytes=32 time=179ms TTL=53

Ping statistics for 195.149.21.11:
Packets: Sent = 4, Received = 4, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
Minimum = 179ms, Maximum = 180ms, Average = 179ms


ADSL Shaped to Jolt
Pinging clarity.jolt.co.uk [195.149.21.11] with 32 bytes of data:

Reply from 195.149.21.11: bytes=32 time=454ms TTL=55
Reply from 195.149.21.11: bytes=32 time=453ms TTL=55
Reply from 195.149.21.11: bytes=32 time=455ms TTL=55
Reply from 195.149.21.11: bytes=32 time=454ms TTL=55

Ping statistics for 195.149.21.11:
Packets: Sent = 4, Received = 4, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
Minimum = 453ms, Maximum = 455ms, Average = 454ms

For International latency, ISDN via the correct ISP, is still the clear winner.
 
Last edited:
Sybawoods you have been moitoring this for a while

So is the trend for DSL getting worse, IMO operators like Telscum should be working on installing better routers, changing network topolgies, reducing hops, and over time the latencies should be coming down as they improve the network (they have no choice but upgrade - its growing every day).

Your traces seem to indicate a much wider gap between Shaped & Unshaped than there used to be when you started with unshaped. IMO during off peak times these should be similar unless there is a deliberate conspiracy to add latency to the Unshaped subnet.
 
jake2 said:
IMO during off peak times these should be similar unless there is a deliberate conspiracy to add latency to the Unshaped subnet.
Click to expand...

That's my view as well. There was a time (early days of ADSL) when basic shaped ADSL had latency even lower than today's unshaped. Obviously congestion is an issue, but I'm convinced that currently latency on the shaped account is being deliberately / "artificially" increased, in part to justify a much more expensive unshaped option for those with low-latency needs.

There are random times - I assume it's when the oke who is hired to sit there solely to push the "increase latency button" is on holiday or in the toilet - when basic shaped ADSL is as low or lower than unshaped. This indicates that better latency is possible evn on the existing infrastructure, they just don't want to activate it.
 
jake2 - for telkom to configure optimal latency it requires alot of work compared to just upgrading bandwidth(which doesnt happen either)

I think its safe to say, telkom dont care if we ping 500-1000ms to overseas gaming servers. As long as they have you on their network, its good enough for them.

Regarding the "push the lantency up" button, well thats how we know telkom? Offer a service that averages at 1000ms anything lower is a bonus, but its theoretically possible to be as low as 150ms. They dont want people to expect to much.
 
There are random times - I assume it's when the oke who is hired to sit there solely to push the "increase latency button" is on holiday or in the toilet - when basic shaped ADSL is as low or lower than unshaped. This indicates that better latency is possible evn on the existing infrastructure, they just don't want to activate it.
Click to expand...

I guess that's what happened here.
 
Thanks sybawoods

Shaping by its very nature, meaning a software approach as opposed to a more hardware based router IS going to add latency and that added latency is simply a software setting on that router device. You can throttle bandwidth in 2 ways, drop packets or add delay buffers & modify receive windows - the latter is more efficient use of the SAT3 pipe.

My assumption is telkom is scared of VOIP systems like www.skype.com now that this has been legalised so they have rendereed the network unusable for voice.

Also remember they are trying to make ADSL unusable for business they cannot aford to lose those R5000pm diginet customers. (& we're b**ching about R1000)

I am seeing ALL that latency on 1 hop, the telecity hop
 
You must log in or register to reply here.
Back
Top
Sign up to the MyBroadband newsletter
X