SL suffering due to USA servers
 

Some of you are probably aware that in USA some ISP's oversell the amount of bandwidth they have. Thus they get stressed down either at various times of day or all the time. Some of them will use p2p shaping or blocking to return some bandwidth. They may apply this to p2p sharing or games. It's cheaper for them to do this than to buy more bandwidth since their business customers & their residential customers who only browse is where they get their money. Of course they never tell anybody about their crises.

An example of a p2p game program that is suffering, overseas users from Australia & some from UK are getting high packet loss as a result of a poorly behaving server in USA. Let's hope the Lindens listen. ;)
The following trace routes were done by (1) myself & after contacting my ISP, (2) they did their own & also (3) arranged for other Australian & UK ISP's to also do trace-routes to Linden.

Due to forum's rules on privacy, some ip addresses have been edited out.

Last login: Sun May 6 20:05:00 on console
Welcome to Darwin!
[L's-Computer:~] L% traceroute sim400.agni.lindenlab.com
traceroute to sim400.agni.lindenlab.com (69.25.104.86), 64 hops max, 40 byte packets
1 x.x.x.x (x.x.x.x) 1.526 ms 0.540 ms 0.427 ms
2 nexthop.vic.ii.net (x.x.x.x) 13.820 ms 15.229 ms 14.456 ms
3 gi1-2.mel-pipe-bdr1.ii.net (203.215.6.22) 14.458 ms 14.251 ms 14.796 ms
4 gi0-15-0-0.syd-ult-core1.ii.net (203.215.20.12) 28.785 ms 29.447 ms 28.768 ms
5 202.139.19.33 (202.139.19.33) 29.529 ms 29.587 ms 29.359 ms
6 203.208.191.157 (203.208.191.157) 191.065 ms 188.809 ms 185.618 ms
7 ge-0-0-0-0.plapx-dr2.ix.singtel.com (203.208.149.2) 186.882 ms 192.964 ms 185.755 ms
8 ge-1-0-0-0.laxow-dr2.ix.singtel.com (203.208.149.254) 197.911 ms 197.145 ms 197.397 ms
9 * * *
10 core1.sfo002.inappnet-26.cr1.sje007.internap.net (66.79.148.134) 196.148 ms 196.549 ms 197.518 ms
11 border2.ge1-1-bbnet1.sfo002.pnap.net (63.251.63.17) 332.327 ms 397.494 ms 235.313 ms
12 * * *
13 * * *
14 * * *
15 * * *


From Optus:

Performing trace to sim400.agni.lindenlab.com...

1 FastEth8-1-0.sb2.optus.net.au (192.168.34.3) 1.832 ms 1.474 ms 1.316 ms
2 203.208.191.157 (203.208.191.157) 157.781 ms 161.735 ms 158.034 ms
3 ge-0-0-0-0.plapx-dr2.ix.singtel.com (203.208.149.2) 157.359 ms 157.927 ms 157.650 ms
4 ge-1-0-0-0.laxow-dr2.ix.singtel.com (203.208.149.254) 169.471 ms 169.462 ms 169.335 ms
5 * * *
6 core1.sfo002.inappnet-26.cr1.sje007.internap.net (66.79.148.134) 169.102 ms 169.256 ms 169.486 ms
7 border2.ge2-1-bbnet2.sfo002.pnap.net (63.251.63.82) 169.037 ms 169.069 ms 169.073 ms
8 * * *
9 * * *
10 * * *

Telstra:

traceroute to sim400.agni.lindenlab.com (69.25.104.86), 30 hops max, 40 byte packets
1 vlan250.lon-service6.Melbourne.telstra.net (203.50.2.177) 0.363 ms 0.262 ms 0.256 ms
2 TenGigE0-12-0-2.win-core1.Melbourne.telstra.net (203.50.79.129) 1.379 ms 1.182 ms 1.337 ms
3 Bundle-Pos1.ken-core4.Sydney.telstra.net (203.50.6.21) 15.546 ms 13.735 ms 14.88 ms
4 Port-Channel1.pad-gw2.Sydney.telstra.net (203.50.6.29) 13.143 ms 13.268 ms 13.121 ms
5 10GigabitEthernet2-0.sydp-core02.Sydney.reach.com (203.50.13.50) 13.218 ms 13.191 ms 13.124 ms
6 i-9-0.sydp-core01.net.reach.com (202.84.221.89) 13.373 ms 13.403 ms 13.391 ms
7 i-14-0.wil-core03.net.reach.com (202.84.143.246) 160.088 ms 160.15 ms 160.118 ms
8 i-6-7.wil03.net.reach.com (202.84.251.230) 160.3 ms 160.496 ms 160.705 ms
9 ge-6-20.car3.LosAngeles1.Level3.net (4.68.111.133) 166.286 ms 166.205 ms 166.227 ms
10 ae-32-54.ebr2.LosAngeles1.Level3.net (4.68.102.126) 164.831 ms ae-32-56.ebr2.LosAngeles1.Level3.net (4.68.102.190) 168.127 ms ae-32-54.ebr2.LosAngeles1.Level3.net (4.68.102.126) 160.679 ms
11 ae-1-100.ebr1.LosAngeles1.Level3.net (4.69.132.5) 160.871 ms 160.664 ms ae-2.ebr1.SanJose1.Level3.net (4.69.132.9) 174.163 ms
12 ae-2.ebr1.SanJose1.Level3.net (4.69.132.9) 175.675 ms 180.591 ms ae-1-100.ebr2.SanJose1.Level3.net (4.69.132.2) 168.314 ms
13 ae-1-100.ebr2.SanJose1.Level3.net (4.69.132.2) 175.798 ms 180.467 ms 180.977 ms
14 INTERNAP-NE.car2.SanFrancisco1.Level3.net (4.78.242.18) 175.416 ms ae-4-4.car2.SanFrancisco1.Level3.net (4.69.133.157) 169.585 ms 169.445 ms
15 INTERNAP-NE.car2.SanFrancisco1.Level3.net (4.78.242.18) 176.113 ms INTERNAP-NE.car2.SanFrancisco1.Level3.net (4.71.44.6) 169.337 ms border2.ge1-1-bbnet1.sfo002.pnap.net (63.251.63.17) 169.561 ms
16 border2.ge1-1-bbnet1.sfo002.pnap.net (63.251.63.17) 169.833 ms * 169.764 ms
17 * * *
18 * * *


From jellybaby.net to linden

1 finch-core-1.router.jellybaby.net (194.159.247.129) 0.734 ms 0.683 ms 0.673 ms
2 finch-service-1-13.router.demon.net (194.159.253.247) 1.672 ms 1.024 ms 1.059 ms
3 finch-backbone-11-162.router.demon.net (194.159.36.65) 3.209 ms 2.899 ms 2.920 ms
4 anchor-backbone-11.router.demon.net (194.159.7.5) 4.868 ms 4.821 ms 6.909 ms
5 anchor-access-5-g3-0-0-s297.router.demon.net (194.70.98.45) 1.795 ms 1.783 ms 2.173 ms
6 ny1-border-1-g0-3-0.router.demon.net (194.70.97.234) 78.454 ms 78.557 ms 78.492 ms
7 cr2-nyiix-peer.nym009.internap.net (198.32.160.185) 78.777 ms 78.930 ms 78.610 ms
8 cr.wdc005.inappnet-73.cr.nym009.internap.net (66.79.146.210) 83.906 ms 83.952 ms 98.947 ms
9 cr1.sje007.inappnet-36.cr1.wdc005.internap.net (66.79.146.213) 166.033 ms 606.455 ms 165.643 ms
10 cr2-cr1.sje007.internap.net (66.79.146.190) 165.644 ms 166.041 ms 165.788 ms
11 core1.sfo002.inappnet-26.cr1.sje007.internap.net (66.79.148.134) 166.258 ms 156.243 ms 155.880 ms
12 border2.ge1-1-bbnet1.sfo002.pnap.net (63.251.63.17) 155.480 ms 155.723 ms 155.399 ms


From staff machine:

traceroute to sim400.agni.lindenlab.com (x.x.104.86), 30 hops max, 38 byte packets

1 staff-fw-hsrp.wa.iinet.net.au (x.x.x.x) 0.384 ms 0.358 ms 0.300 ms
2 gi0-1-1.icp-qv1-lcore1.ii.net (203.59.24.253) 1.178 ms 1.164 ms 1.179 ms
3 gi3-3.per-qv1-bdr1.ii.net (203.215.4.55) 1.960 ms 1.742 ms 1.804 ms
4 gi0-14-0-0.syd-ult-core1.ii.net (203.215.20.2) 60.453 ms 60.050 ms gi0-14-0-1.syd-ult-core1.ii.net (203.215.20.4) 60.582 ms
5 202.139.19.33 (202.139.19.33) 60.289 ms 62.652 ms 60.136 ms
6 203.208.191.157 (203.208.191.157) 224.654 ms 216.697 ms 216.185 ms
7 ge-4-0-0-0.plapx-dr2.ix.singtel.com (203.208.149.6) 216.192 ms 216.481 ms 216.535 ms
8 ge-1-0-0-0.laxow-dr2.ix.singtel.com (203.208.149.254) 226.861 ms 230.779 ms 226.507 ms
9 * * *
10 core1.sfo002.inappnet-26.cr1.sje007.internap.net (66.79.148.134) 227.598 ms 227.241 ms 227.820 ms
11 border2.ge1-1-bbnet1.sfo002.pnap.net (63.251.63.17) 342.667 ms 339.339 ms 238.195 ms
12 * * *
13 * *



"this would most likely be the last hop or two before the actual linden server/uplink, spoken with a senior best advice would be to contact Linden with the above information and see what they say..
Its not just traffic going over the singtel link its also the level3 link and the internap link."


(Edit: 25/5/07: (1) Recent gif attachment included. Bandwidth is set to 320 to reduce packet loss. 25% loss was about 7-8 hours before the other. * Problem 2 is program related: -> (2) jpg image: -> Building is a club. Typical about how all one sees is grass, then eventually the floor. Can see the distant buildings but not the building walls. After perhaps 15-30 mins eventually the building will show. Nice huh. Oh & "IF" AV's are visible, they are usually grey. Add to that chat lag, IM lag, movement lag, difficulty loading appearance & inventory, etc. The list goes on. But for some reason, FirstLook seems to overcome most of this problem. But where is FirstLook now? Why is there no other option for us? Look around this forum & you will see many p2p programs where one version might not work for one person but does for another .. & then the next version perhaps vice versa. So ... more than one version is needed for users. Not just your only choice SL version works for you or you are stuffed philosophy.)

Would be nice to hear some SL feedback! ? ;) Despite the tone earlier, excuse me but, I am wanting to solve this for both myself & others who are in the same boat.

"FirstLook" the Saviour !!!
 

Well finally after about 6+ weeks FirstLook is available again. What a saviour. Life is back to normal in SL. For some reason packet loss is minimised. So why is FirstLook not available ALL the time, not just when the Lindens feel inspired to release it. There are a number of people in the same boat as me who cannot use the normal SL program properly but solve their problem using FirstLook. SL is the first p2p program I have come across that dogmatically forces everyone to use a single version regardless of whether it works for everyone or not. And fact is, with all p2p programs there are always a notable percentage of persons who can not use the most recent or particular version thus stay with what they have or go back a version. FirstLook should be available 100% of the time for all users if they are struggling with the main SL program. Doing so is not making the network go backward, it is catering for everyone.

Note the packet loss sample image below & compare it to the packet loss of the above attachments in the previous post. 1% compared to over 20%. Also, user differences are: 1. Rezzes buildings almost immediately. 2. Rezzes people almost immediately (though they may be grey, at least they can be seen.) 3. Less inventory issues. 4. As a result of less packet loss, no clothes or object issues when adding or removing them from the AV. 5. Objects are more easily rezzed to the ground or returned to inventory.

The 2nd attachment GiF below is taken from a club. Same club as an image above where the building can't be seen. The AV was logged directly into the club & the walls rezzed within a second or 2. Other minor details a little longer. Packet loss was a little high for about a minute .. about 14%. But over time this settled down. As can be seen by both frames of the GiF image, the average packet loss became minimal after 10-20 mins.