dslreports logo
site
 
    All Forums Hot Topics Gallery
spc

spacer




how-to block ads


Search Topic:
uniqs
1589
share rss forum feed


JigglyWiggly

join:2009-07-12
Pleasanton, CA

4 edits

So why are comcast pings so much higher than att dsl?

So I have xtreme 50
and a DSL modem from ATT.

Pfsense firewall rules so I can configure ports to go through whichever modem I want.

So for quake live, west coast:
I get 30-45ms with Comcast
18ms with ATT DSL

80ms with comcast
65ms to tex with DSL

I host a killing floor dedicated server
my friend in florida gets 104ms with comcast
72ms with att DSL

My friend from chicago gets 58ms with ATT DSL
75 with comcast

Tribes Ascend I get 16ms with ATT DSL
30ms with comcast

to east coast i get 82ms with DSL
105 with Comcast

why is comcast so bad at this? Quake Live there is a huge difference between 18ms and 30-40ms, the game's netcode is garbage. (There is a huge difference between LAN and 18ms too)

I live in Pleasanton in california
here is a random tracert to dslreports

quote:
1 48 ms 28 ms 14 ms 76.21.108.1
2 10 ms 9 ms 11 ms te-7-2-ur06.pleasanton.ca.sfba.comcast.net [68.8
5.191.69]
3 21 ms 23 ms 11 ms te-0-1-0-4-ar01.oakland.ca.sfba.comcast.net [69.
139.198.30]
4 23 ms 23 ms 67 ms he-2-15-0-0-cr01.sacramento.ca.ibone.comcast.net
[68.86.91.225]
5 77 ms 31 ms 37 ms pos-1-15-0-0-cr01.losangeles.ca.ibone.comcast.ne
t [68.86.85.86]
6 106 ms 55 ms 55 ms pos-0-7-0-0-cr01.dallas.tx.ibone.comcast.net [68
.86.88.206]
7 81 ms 95 ms 83 ms pos-0-15-0-0-cr01.atlanta.ga.ibone.comcast.net [
68.86.85.150]
8 99 ms 95 ms 95 ms he-0-3-0-0-cr01.ashburn.va.ibone.comcast.net [68
.86.89.149]
9 117 ms 99 ms 99 ms pos-0-12-0-0-cr01.newyork.ny.ibone.comcast.net [
68.86.85.30]
10 115 ms 97 ms 97 ms 173.167.58.26
11 101 ms 97 ms 97 ms 0.e1-4.tbr1.oct.nac.net [209.123.10.122]
12 113 ms 97 ms 115 ms vlan804.esd1.oct.nac.net [209.123.10.2]
13 97 ms 112 ms 143 ms www.dslreports.com [209.123.109.175]
att dsl
quote:
1 8 ms 8 ms 8 ms adsl-70-137-159-254.dsl.snfc21.sbcglobal.net [70
.137.159.254]
2 9 ms 8 ms 8 ms dist2-vlan60.snfc21.pbi.net [216.102.187.131]
3 8 ms 8 ms 8 ms 151.164.102.36
4 9 ms 9 ms 8 ms 12.122.200.13
5 11 ms 11 ms 11 ms 192.205.33.50
6 84 ms 84 ms 84 ms nyk-bb2-link.telia.net [213.155.130.130]
7 85 ms 84 ms 85 ms nyk-b3-link.telia.net [80.91.245.82]
8 85 ms 84 ms 84 ms netaccess-tic-133837-nyk-b3.c.telia.net [213.248
.99.90]
9 85 ms 86 ms 88 ms 0.e1-4.tbr1.mmu.nac.net [209.123.10.101]
10 85 ms 86 ms 85 ms 0.e1-1.tbr1.oct.nac.net [209.123.10.17]
11 85 ms 85 ms 85 ms vlan804.esd1.oct.nac.net [209.123.10.2]
12 85 ms 85 ms 85 ms www.dslreports.com [209.123.109.175]
serious this isn't a bashing comcast thread

i am asking why are the pings so much higher?

it's not my modem/signals, same problem in merced california.
If you really want them, here, but my signals are much better in merced california, 3 upstream there too.



EDIT: here are signals in merced ca


also with comcast my netgraph in quake live is not as smooth as DSL. It's smooth-ish, but with DSL, it's totally flat.

random pingtest.net


with dsl i get 1ms of jitter

i wouldn't whine as much, but DSL is dying. ATT and verizon are trying to kill it. That will piss me off. Also ATT is trying to push uverse everywhere, and that kills the pings, luckily that's not available in my area.

Also if you think I am whining for nothing, well yes, but it really does make a big difference for quake live and games. Probs happen when you are top 10 aimers in teh US. (yeah

»www.youtube.com/watch?v=Sx5qLbSwYeQ
)


whfsdude
Premium
join:2003-04-05
Washington, DC
Reviews:
·Comcast

There isn't a simple answer.

1) Peering: Comcast has a pretty restrictive settlement free peering policy. As opposed to many other networks, some which have open peering policies.
»www.peeringdb.com/private/partic···p?id=822

2) Regionalization: Cable co's are pretty heavily regionalized so paths tend to go through their regions. Also peering doesn't often involve local IXs.

3) Congested paid transit: »www.merit.edu/mail.archives/nano···911.html

4) Routing: Could be using cold-potate routing in order to maximize peering links or manage congestion. We don't know though.

5) Congestion: Usually on the CMTS side.

Fiber wise you should be seeing ~125/miles per ms. So so if you have fiber from both ends NJ to CA you're looking at about ~25ms. So ~ 50ms RTT. The traceroute doesn't say much though because we need a return path from the other side.



JigglyWiggly

join:2009-07-12
Pleasanton, CA

1 edit

so are you saying comcast blows and there is nothing anyone can do about it?
Though people with Cox seem to have even worse pings in QL.

i don't see how i could get a return path ping though. Any service that lets you do that?

here is a reverse traceroute according to
»whatismyipaddress.com/traceroute-tool

quote:
Hop Time Host IP Location
1 0.841 10.0.0.1 10.0.0.1 , ,
2 10.778 10.20.62.254 10.20.62.254 , ,
3 11.019 r2-l3tca-cr2.nextweb.net 216.237.3.33 Irvine, CA, United States
4 8.68 ge-6-15.car2.Tustin1.Level3.net 4.79.142.41 , , United States
5 10.562 vl-3102-ve-133.ebr1.Tustin1.Level3.net 4.69.160.13 , , United States
6 9.483 ae-6-6.ebr1.LosAngeles1.Level3.net 4.69.153.221 , , United States
7 12.527 ae-91-91.csw4.LosAngeles1.Level3.net 4.69.137.14 , , United States
8 7.594 ae-3-80.edge3.LosAngeles1.Level3.net 4.69.144.137 , , United States
9 28.797 COMCAST-IP.edge3.LosAngeles1.Level3.net 4.71.136.10 Long Beach, CA, United States
10 25.471 pos-2-10-0-0-cr01.sanjose.ca.ibone.comcast.net 68.86.87.189 Kansas City, MO, United States
11 27.365 pos-0-15-0-0-ar01.sfsutro.ca.sfba.comcast.net 68.86.90.154 Kansas City, MO, United States
12 24.64 te-2-1-ur05.pleasanton.ca.sfba.comcast.net 68.85.155.1 Clawson, MI, United States
13 46.723 te-18-10-cdn03.pleasanton.ca.sfba.comcast.net 68.85.191.70 Bensalem, PA, United States
21 N/A c-76-21-111-xxx.hsd1.ca.comcast.net 76.21.111.xxx , ,
me tracerting them
quote:
1 66 ms 19 ms 25 ms 76.21.108.1
2 12 ms 31 ms 8 ms te-7-2-ur06.pleasanton.ca.sfba.comcast.net [68.8
5.191.69]
3 27 ms 11 ms 23 ms te-0-1-0-4-ar01.oakland.ca.sfba.comcast.net [69.
139.198.30]
4 35 ms 12 ms 35 ms he-2-13-0-0-cr01.sacramento.ca.ibone.comcast.net
[68.86.90.141]
5 20 ms 19 ms 15 ms pos-0-4-0-0-cr01.sanjose.ca.ibone.comcast.net [6
8.86.85.50]
6 24 ms 32 ms 24 ms xe-11-0-0.edge1.SanJose1.Level3.net [4.79.43.137
]
7 26 ms 37 ms 23 ms vlan90.csw4.SanJose1.Level3.net [4.69.152.254]
8 55 ms 26 ms 24 ms ae-91-91.ebr1.SanJose1.Level3.net [4.69.153.13]

9 29 ms 25 ms 25 ms ae-2-2.ebr3.LosAngeles1.Level3.net [4.69.132.10]

10 25 ms 70 ms 34 ms ae-63-63.csw1.LosAngeles1.Level3.net [4.69.137.3
4]
11 42 ms 49 ms 30 ms ae-12-60.car2.LosAngeles1.Level3.net [4.69.144.4
]
12 24 ms 26 ms 23 ms ge2-2.bbsr1.lax.megapath.net [63.215.86.150]
13 28 ms 24 ms 26 ms gig1-1.e40.lax.bb.megapath.net [155.229.120.98]

14 38 ms * 31 ms 209-101-27-70.lax-e100.cust.gw.epoch.net [209.10
1.27.70]
15 37 ms 38 ms 51 ms whatismyipaddress.com [66.80.82.69]


whfsdude
Premium
join:2003-04-05
Washington, DC
Reviews:
·Comcast

said by JigglyWiggly:

so are you saying comcast blows and there is nothing anyone can do about it?

There is nothing you can do as a consumer besides voicing your opinion and voting with your wallet.

[rant]
This is where it would be nice to have a fiber muni/local government provider that just sold users dark pairs of fiber between your house and a central office. Then you could choice what provider they want to light the fiber.

If you wanted cheaper bandwidth (meaning you can get more of it) that's slightly higher latency, you'd choose a provider like Comcast, Cogent, HE, XO, etc...

If you wanted more expensive bandwidth that's lower latency, you'd choose a provider like Level3, NTT, NLayer, etc...
[/rant]

If you're looking for your dslreports.com example take a look at NAC's LG for the return path. »www.eng.nac.net/lookingglass/lg.html


JigglyWiggly

join:2009-07-12
Pleasanton, CA

3 edits

what about the jitter though? DSL has like 0 jitter :?
That isn't related to the routing, or I think so.

I mean I have xtreme50, shouldn't I be prioritized over everyone else? It's also 2:40am atm, and I did pingtest, I got 5ms instead, and now /gnite

Is docsis just garbage for pings/jitter? Though this thread makes me wonder why their jitter is lower »Pingtest.net results..whats yours?

I don't understand how some people have 8ms to the pingtest server. I get 8ms to my first hop on the lowest.



gar187er
I do this for a living

join:2006-06-24
Dover, DE
kudos:4
reply to JigglyWiggly

hfc design....its mainly the way docsis works, always will have higher latency with cable then with dsl...at the same time your speeds will be much higher though.

you also cant compare to other people in other systems because of variables such as congestion, cmts configurations, distances, etc.
--
I'm better than you!



tshirt
Premium,MVM
join:2004-07-11
Snohomish, WA
kudos:4
Reviews:
·Comcast
reply to JigglyWiggly

said by JigglyWiggly:

...I mean I have xtreme50, shouldn't I be prioritized over everyone else?

Absolutely not, that would not be network/content neutral, and those people whine even more than you.

As has been said above it has a lot to do with DOCSIS/cable plant design.
it's not that cable latentcy is so bad, it's just not their highest priority (which is moving more bits for more people faster)
DSL plants aren't being shutoff tomorrow, their maintaince, reengineering, repair is just getting lower.

I guess you could keep moving to stay with copper until fiber gets alittle closer, the upside is when everyone else is on LTE with 100-200+ms latentcy, your cable will be the fastest lowest jitter thing around.