I'm Interested in a T1 Line

Status
Not open for further replies.
Uberbob102000 said:
Well if you want to get a T1 go right ahead, as people have said before, as soon as it's outside the ISPs network it's out of their control. You are NOT going to significantly reduce your latency that way. When I get back to work on Wednesday from visiting family I can ping from Redmond, WA to whatever city you'd like on T1, FiOS and Comcast and I'll be willing to bet they'll easily be within 10% of each other, which is easily within a margin of error for that test.

You seem to have already made up your mind, even with multiple people telling you otherwise.

As for 100ms being bad for gaming; I've managed to come out top of my server on 70% of the multiplayer games I've played this weekend. On a 4G tethered connection that's unreliable and shitty as hell.
Click to expand...

Great. I can top-3 on shitty pings too, but it's not fun and it's frustrating as hell when the only reason the other guy got the better of you was because of high latency.
 
What you aren't realizing, is that there isn't really a backbone. There are backbones, but no singular backbone. Level3, ATT, VZ Business, Cogent, and more, operate national and international backbones. Those backbones peer with each other at both private and public points. Thus traffic must pass from one backbone to another, unless the destination is on the same backbone you'll be routed across some peering point which will often physically extend the distance the packets must travel. Finally, over long haul, fiber rarely follows direct routes. While a traceroute may show going directly from Portland to NYC, it's bounced around fiber regen sites, possibly gone through hidden MPLS hops, etc. on it's way..
 
Radical said:
In order to get the responses you want you should consider editing the OP to read "I'm going to get T1 to my house, thank you and goodbye."
Click to expand...

Only one person has given me a useful response and that is SJ. Uber's not in a position to help me right now, but he's going to. Everyone else has just speculated or beat around the bush to no effect.

I didn't want people to say "Yeah, you can get <50ms to the east coast." I want people to say, "I have a T1 through X provider at X location, and I ping X to here." From there I can decide for myself the potential value of getting a T1 depending on the general <50s area that people are reporting. I don't need nor want people to decide things for me.
 
JTY said:
What you aren't realizing, is that there isn't really a backbone. There are backbones, but no singular backbone. Level3, ATT, VZ Business, Cogent, and more, operate national and international backbones. Those backbones peer with each other at both private and public points. Thus traffic must pass from one backbone to another, unless the destination is on the same backbone you'll be routed across some peering point which will often physically extend the distance the packets must travel. Finally, over long haul, fiber rarely follows direct routes. While a traceroute may show going directly from Portland to NYC, it's bounced around fiber regen sites, possibly gone through hidden MPLS hops, etc. on it's way..
Click to expand...

I do realize what you think I'm not realizing.

And clearly there's no single line, or bundle of lines, that run from Portland to NYC. The signal would degrade a long time before it even got there.
 
In that case, update your original post with the list IPs you want people to check. As it is, there are IPs in Chicago I can get 50ms to, while others over 90ms due to being routed via LA to Dallas, to Chicago.
 
An entire thread of network admins and people in the telecom business and the OP still things a T1 might be worth it.

Its amazing how absent minded people will be sometimes.

I work in telecom at one of the CLEC's mentioned in this thread and I wouldn't be interested in a T1 if it was the last wireline based internet connection on earth.

Let me ask you, do you get killed in a game and blame it on your ping being ~100ms to the server? If so, you have some OCD problems to be worked out.. Psssst.. its NOT your internet connection.
 
JTY said:
In that case, update your original post with the list IPs you want people to check. As it is, there are IPs in Chicago I can get 50ms to, while others over 90ms due to being routed via LA to Dallas, to Chicago.
Click to expand...

If we know how it's being routed, the additional distance can be adjusted for. If you need a list of servers, then I can get some.
 
NExUS1g said:
100ms is very bad for gaming.

Who's to say that the DSL doesn't get routed the same as my cable eventually? It probably does considering all traffic connects to a backbone that goes to Seattle and back to Portland when I'm connecting to a local machine, but even then cable provides a consistently and significantly lower latency.
Click to expand...

It's not that bad. I'm in Canada so I ping 100-200 on most US game servers and manage fine. If anything, you get so used to it that if you end up with a low ping you kick everybody's asses much more easily. :D Of course it depends on the game, but still, 100 is not that bad.
 
Red Squirrel said:
It's not that bad. I'm in Canada so I ping 100-200 on most US game servers and manage fine. If anything, you get so used to it that if you end up with a low ping you kick everybody's asses much more easily. :D Of course it depends on the game, but still, 100 is not that bad.
Click to expand...

I ping 100 and I can prefire someone around a corner and still die because my shots didn't register and interpolation allowed them to see me before I saw them.
 
NExUS1g said:
If we know how it's being routed, the additional distance can be adjusted for. If you need a list of servers, then I can get some.
Click to expand...

Routing can and does change from day to day. Plus the path from the destination to you can vary just as much. I love asymetric routing. There are game servers in Seattle that from my ISP to them, is 4 hops direct. But from the server to me, is 15 hops, and goes all the way to LA, and then back.
 
criccio said:
Its been answered 100+ times in this thread.

You know.. the one you created and haven't listened to anything written in it so far.
Click to expand...

The people who are posting seem to expect me to take for gospel what they say when they say, "It's crap. A waste of money."
 
NExUS1g said:
The people who are posting seem to expect me to take for gospel what they say when they say, "It's crap. A waste of money."
Click to expand...

Those people are people with experience, something you're lacking. That is, after all, the entire purpose of posting a question like this on a public forum.
 
criccio said:
Those people are people with experience, something you're lacking. That is, after all, the entire purpose of posting a question like this on a public forum.
Click to expand...

That's not the thread's purpose at all. That's why I think people aren't reading the OP.
 
See, most of us are network admins, telecom guys, etc. we deal in expensive connectivity, we build networks, etc. For some of us, we spend time every day doing route/path optimization trying to squeeze out the best we can from what we have. So, we all know a thing or two about this, because we do this work everyday, and have done it for a long time. I don't know about others, but I've been in the ISP field for 10 years doing admin work, and then network engineering.
 
JTY said:
Routing can and does change from day to day. Plus the path from the destination to you can vary just as much. I love asymetric routing. There are game servers in Seattle that from my ISP to them, is 4 hops direct. But from the server to me, is 15 hops, and goes all the way to LA, and then back.
Click to expand...

If it changes like that, then it doesn't make much of a difference in most circumstances. The game servers I've been connecting to for the last few years still have roughly the same latency today as they did a year ago.
 
JTY said:
See, most of us are network admins, telecom guys, etc. we deal in expensive connectivity, we build networks, etc. For some of us, we spend time every day doing route/path optimization trying to squeeze out the best we can from what we have. So, we all know a thing or two about this, because we do this work everyday, and have done it for a long time. I don't know about others, but I've been in the ISP field for 10 years doing admin work, and then network engineering.
Click to expand...

Same. I see no reason for running through the "tests" the OP is interested in because I know from experience that its futile. A 30 year old T1 in 2012 is an absolute waste of money.
 
Since you use Comcast, you have an advantage. Comcast operates a national backbone. So, they peer at a large number of locations, and due to that, you generally won't see huge path changes unless something happens like a fiber cut.
 
JTY said:
Since you use Comcast, you have an advantage. Comcast operates a national backbone. So, they peer at a large number of locations, and due to that, you generally won't see huge path changes unless something happens like a fiber cut.
Click to expand...

When I was on DSL, same thing. Latencies were higher, but generally consistent over long periods of time. FIOS was the same way when I lived in Verizon's area.
 
Major telcos like Qwest and Verizon both operate backbones. You probably had Qwest which by default uses interleaving on their DSL circuits to improve reliability. They will remove interleaving if requested, and the line is stable, the result is low latency from you to their nearest IP POP.
 
JTY said:
Major telcos like Qwest and Verizon both operate backbones. You probably had Qwest which by default uses interleaving on their DSL circuits to improve reliability. They will remove interleaving if requested, and the line is stable, the result is low latency from you to their nearest IP POP.
Click to expand...

And so you're saying that things like residential DSL provided by Qwest/Centurylink are awesome and stable, but a T1 (a business line with an SLA) provided by the same company that's up to 5x higher cost is not? That seems backwards.
 
Just to add, DSL from Qwest without interleaving is probably as good as you'll get without spending some serious money on a T1 with a tier1 national carrier. The reason being, Qwest doesn't really operate a separate residential network. Their DSL circuits once terminated into the ATM cloud go directly to the nearest Qwest IP POP.
 
NExUS1g said:
And so you're saying that things like residential DSL provided by Qwest/Centurylink are awesome and stable, but a T1 (a business line with an SLA) is not? That seems backwards.
Click to expand...

All the budget T1s are offered by CLECs, and others that operate budget networks. Many of which suck. If you want a T1 that will deliver the lowest possibly latency, you need to get it from a Tier1 carrier like Level3, which is expensive.
 
JTY said:
All the budget T1s are offered by CLECs, and others that operate budget networks. Many of which suck. If you want a T1 that will deliver the lowest possibly latency, you need to get it from a Tier1 carrier like Level3, which is expensive.
Click to expand...

You said about $500 a month for Level 3 right? Check my edit on my last post.
 
In some cases it is a bit better. For example with Verizon, they operate Verizon Internet Services which is consumer stuff. That network then connects to Verizon Business which is where they offer T1s.

That said, T1s are expensive because they can. They have SLAs, which means when a squirrel eats your overhead line at 5AM on a Sunday, a guy will be out there in a few hours to replace it. WIth a T1, the bulk of the cost is just for the SLA, and often better support. It's meant to give you that warm cozy feeling, that when things break, someone will fix them without hassle.
 
JTY said:
In some cases it is a bit better. For example with Verizon, they operate Verizon Internet Services which is consumer stuff. That network then connects to Verizon Business which is where they offer T1s.

That said, T1s are expensive because they can. They have SLAs, which means when a squirrel eats your overhead line at 5AM on a Sunday, a guy will be out there in a few hours to replace it. WIth a T1, the bulk of the cost is just for the SLA, and often better support. It's meant to give you that warm cozy feeling, that when things break, someone will fix them without hassle.
Click to expand...

If T1 is such a bad service, then why do telcos use it for their backbones?
 
Nobody uses it for backbones. They are just too slow. Backbones are built using SONET, and Ethernet these days. T carrier is mostly used for delivering a large number of phone lines to offices for PBXs, or dialup access servers.
 
JTY said:
Nobody uses it for backbones. They are just too slow. Backbones are built using SONET, and Ethernet these days. T carrier is mostly used for delivering a large number of phone lines to offices for PBXs, or dialup access servers.
Click to expand...

Ah right. I even mentioned that earlier and screwed it up. Someone had mentioned early on something about being able to connect via Ethernet through my ISP for a good price. Could you teach me about that a little bit?
 
Comcast and several other ISPs offer metro ethernet service. Typically delivered via fiber directly to you. Comcast would probably be the cheapest to get it from, since they would most likely have fiber the closest to you. But, I have no idea how much different they route their ethernet service vs their cable.
 
NExUS1g said:
What do you get to the following IP: 68.232.182.53

Could you do a tracert?
Click to expand...

Via Comcast from Sammamish, WA
C:\Users\Tawnos>ping 68.232.182.53

Pinging 68.232.182.53 with 32 bytes of data:
Reply from 68.232.182.53: bytes=32 time=75ms TTL=48
Reply from 68.232.182.53: bytes=32 time=74ms TTL=48
Reply from 68.232.182.53: bytes=32 time=74ms TTL=48
Reply from 68.232.182.53: bytes=32 time=73ms TTL=48

Ping statistics for 68.232.182.53:
Packets: Sent = 4, Received = 4, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
Minimum = 73ms, Maximum = 75ms, Average = 74ms

C:\Users\Tawnos>tracert 68.232.182.53

Tracing route to 68.232.182.53 over a maximum of 30 hops

1 1 ms <1 ms 1 ms gohome.gohome.tawnos.com [192.168.1.1]
2 * * * Request timed out.
3 11 ms 12 ms 11 ms te-4-2-ur03.bellevue.wa.seattle.comcast.net [68.
86.113.145]
4 13 ms 10 ms 9 ms te-2-3-ur01.bellevue.wa.seattle.comcast.net [68.
86.177.185]
5 16 ms 23 ms 11 ms ae-5-0-ar03.burien.wa.seattle.comcast.net [68.86
.96.213]
6 28 ms 34 ms 37 ms pos-3-14-0-0-cr01.sacramento.ca.ibone.comcast.ne
t [68.86.93.109]
7 30 ms 33 ms 33 ms pos-0-3-0-0-cr01.sanjose.ca.ibone.comcast.net [6
8.86.87.178]
8 57 ms 41 ms 40 ms pos-0-9-0-0-pe01.11greatoaks.ca.ibone.comcast.ne
t [68.86.88.110]
9 33 ms 39 ms 34 ms 208.178.58.1
10 34 ms 38 ms 33 ms ae9-20G.scr3.SNV2.gblx.net [67.16.145.117]
11 71 ms 74 ms 73 ms ae0.csr1.CHI2.gblx.net [67.16.164.78]
12 74 ms 84 ms 83 ms 69.31.105.21
13 87 ms 79 ms 79 ms as20473.xe-1-0-5-1615.ar2.ord1.us.nlayer.net [69
.31.111.178]
14 76 ms 75 ms 84 ms 68.232.182.53

Trace complete.


From microsoft, it appears the ICMP return packets are filtered, so I can't check, and really, it's jut not worth it. The fact is, a T1 isn't going to do you much.

The fact is, I get more speed down/up on my Comcast cable connection going to the east coast than you would with your T1, and I would wager the latency is within 15% of a T1: http://www.speedtest.net/result/1643642974.png
 
I don't want to say too much about this but. If the OP wants to post a list of IPs I will gladly post ping times from a TWC 30/5 cable connection and a T1. The connections are about 10 mi apart in the same city. The T1 is from a tier 1 carrier.

Also I won't speak for all metro-E connections but I am getting circuit for a lab in January and the problem I see here for most home installs is the delivery. Mine is coming in via 1000Base LX. This is a single mode fiber gige connection with a 25Mb port speed over an MPLS cloud. For us this meant a Cisco 7600 router + modules.I don't recall the price as it was quoted with some other gear for the project but it was not cheap.
 
NExUS1g said:
I don't know of any games that actually use up anywhere near 1.544Mbps. It's my understanding that T1's are much more reliable and provide a near direct access to the internet backbone. And hello from Milwaukie.

Oh, /edit, I live on the east side (as you saw Milwaukie) so no Verizon and no FIOS. Comcast is the closest we got here. The DSL pings from Qwest (now Centurylink) are double that of Comcast, so that's definitely a no-go from personal experience.
Click to expand...

Oh god, please don't tell me your on Qwest. I work on River Road btw. Qwest in your area is god awful. It was our main provider for years. Have you looked in to comcast business? Oh and Frontier is coming in to the area! I stopped and talked with a couple of installers and over the next 6 months (at least my side) will have fios. I will test our comcast business line for you tomorrow. I really do not think you will be seeing a difference though.
 
The Spyder said:
Oh god, please don't tell me your on Qwest. I work on River Road btw. Qwest in your area is god awful. It was our main provider for years. Have you looked in to comcast business? Oh and Frontier is coming in to the area! I stopped and talked with a couple of installers and over the next 6 months (at least my side) will have fios. I will test our comcast business line for you tomorrow. I really do not think you will be seeing a difference though.
Click to expand...

Cool, I live right down the road from where you work. But, no, I'm not on Qwest. I was for a bit, but I got on Comcast ASAP. Comcast business seems to be their standard cable connection. I may not have seen all of their offerings though. Having FIOS over here on the east side is exciting news. I had it for a bit when I was living in Cornelius and it was like a wet dream.

If we're getting a true FTTH service out here, then I'd definitely be going with that.
 
Last edited:
My FIOS pings from Bothell are on par with Tawnos' Comcast in Sammamish FWIW.
 
Nicklebon said:
I don't want to say too much about this but. If the OP wants to post a list of IPs I will gladly post ping times from a TWC 30/5 cable connection and a T1. The connections are about 10 mi apart in the same city. The T1 is from a tier 1 carrier.

Also I won't speak for all metro-E connections but I am getting circuit for a lab in January and the problem I see here for most home installs is the delivery. Mine is coming in via 1000Base LX. This is a single mode fiber gige connection with a 25Mb port speed over an MPLS cloud. For us this meant a Cisco 7600 router + modules.I don't recall the price as it was quoted with some other gear for the project but it was not cheap.
Click to expand...

8.3.5.16
68.232.161.42
8.6.74.195
8.9.37.202
 
Beyond the mere ping thing, you do know that most predictive ping algorithms don't work the way you described, right?

I suggest you practice your gaming more.
 
Tawnos said:
Beyond the mere ping thing, you do know that most predictive ping algorithms don't work the way you described, right?

I suggest you practice your gaming more.
Click to expand...

I don't know what a predictive ping algorithm is.

And you really should mind your baseless insults regarding my, as far as you're concerned, unknown skills in gaming.

This is what 100ms gets you. I prefired this guy in the game before he even started peeking after he went behind the box to give you an idea of the lag.
http://www.youtube.com/watch?v=GwcP__uI3zY

Game KDR's like 28-4, 27-3, 43-2, 44-2, 32-2, 36-2, 19-0 are indication you're wrong.
 
Last edited:
kill14me said:
A T1 is is not the same as a residential DSL line. They are provisioned differently, have different encoding schemes, error correction and bit rates. You would need a HDSL2 or SDSL line coming into your house to get the equivalent of a T-1 in that regard.

Remember also that ADSL is asynchronous (hence the "A") which means the line speeds are different for inbound and outbound. ADSL has a higher download speed than upload speed. While any T-1 line is synchronous, you get 1.544 Mbps both ways.

Lastly remember that the Service Provider is only concerned with the line speed from the CO (Central Office) or Head End for Cable to the Demarc at your premise. Anything equipment inside of the Demarc causing issues is the customers responsibility unless you are using their equipment.

Past the the Central Office you might be able to submit a trouble ticket if the lines are still on their network.

I guess what I'm trying to say any provider may/will promise a certain line speed with your connection but that isn't to every place on the Internet you might go. Once the traffic is off their Network there isn't much they can do about it.
Click to expand...

Thanks for clarifying that, it was a great explanation. :)

EDIT:
After reading through this thread, I'm sorry, but even when I have gamed in the past, 100ms was acceptable.

NExUS1g, do you game professionally?
I understand that having low latencies is your top priority over high data transfer rates, but don't you think you've pushed this far enough?

Why not just get the T1 line for a month or two, see if it actually will improve or degrade your latency, then move on.

Everyone here has a ton of experience in this area.
I know you don't like the answers you are seeing, but there are just too many variables within the infrastructure of the Internet to give you a completely solid/truthful answer.

I tracerouted and pinged your IP that you gave, and I get between 80-90ms, and I and others are able to game on my connection very well.

I'm not trying to be rude, but when you are going through all of this to simply shave off ~10ms, I agree with the others, it's time to improve your game. ;)
 
Last edited:
Status
Not open for further replies.
Back
Top