would using a proxy server improve my routing/ping?

[Ohrami]

I live in Louisville, Kentucky, and I usually get ~45 ms of latency to Chicago servers (quite suboptimal considering the distance). I found a website advertising a proxy server which claims that it has servers in both Louisville and Chicago and that the latency between the two is around 12 ms. This website also claims to get ~25-30 ping to Texas servers, which I usually ping around 60 to. https://wondernetwork.com/pings/Louisville

If I were to subscribe to this proxy server or a similar service, would my latency be similarly reduced, or would my latency remain at 45+?

 

[yinan]

It has the potential to get worse. Going through another middleman is never a good idea.

 

[eton975]

Depends on the source of the latency. Check your router first by pinging it.

 

[Ohrami]

http://puu.sh/kjbU9.png Here is a full traceroute to a Chicago server from my router. Seems to have gone down from last I checked (a few months ago) but only by a few milliseconds.

 

[dave_the_nerd]

http://puu.sh/kjbU9.png Here is a full traceroute to a Chicago server from my router. Seems to have gone down from last I checked (a few months ago) but only by a few milliseconds.

There's always some variability.

But honestly? It's taking 29ms just to get out of your ISP's network (rr.com) so except for complaining, I'm not sure there's a lot you can do about it.

 

[Mushkins]

I would question if this is even something worth worrying about, much less spending money on. If you're browsing the internet, checking email, or even playing online games 45 ms vs 13 ms is a meaningless distinction.

Are you just trying to get the numbers down, or is there some mission critical business case here where an extra 30 milliseconds is a legitimate issue?

 

[Ohrami]

45 ms vs 13 ms is not a meaningless distinction in online games with no lag compensation. I want the numbers down for playing online games primarily. 30 ms vs 40 ms is very very clearly noticeable for me, so 12 ms vs 45 ms would be night and day. $15 per month isn't very significant for me so I wouldn't mind paying it to decrease latency by that much if possible.

 

[dave_the_nerd]

45 ms vs 13 ms is not a meaningless distinction in online games with no lag compensation. I want the numbers down for playing online games primarily. 30 ms vs 40 ms is very very clearly noticeable for me, so 12 ms vs 45 ms would be night and day. $15 per month isn't very significant for me so I wouldn't mind paying it to decrease latency by that much if possible.

Well, proxy servers wouldn't help.

Consider switching ISPs.

Also, calling bullshit on the bolded part until there's a double-blind test.

 

[Ohrami]

I would switch ISPs but Time Warner is all we have in Louisville other than Verizon which is just as bad if not worse.

Try playing Tribes: Ascend and if you can find two servers with a 10 ms difference, you can clearly feel it. There's no lag compensation at all and most of the weapons are projectiles which need to be aimed further ahead (more leading required) with higher ping, and they also shoot slower.

 

[dave_the_nerd]

I would switch ISPs but Time Warner is all we have in Louisville other than Verizon which is just as bad if not worse.

Try playing Tribes: Ascend and if you can find two servers with a 10 ms difference, you can clearly feel it. There's no lag compensation at all and most of the weapons are projectiles which need to be aimed further ahead (more leading required) with higher ping, and they also shoot slower.

What's it take to host your own server? It'd solve the lag problem for you and the local buddies you play with, if any.

'course, visiting teams would be screwed, but maybe that's not so bad, eh?

 

[Ohrami]

Can't host your own servers. Is there any way to remove some of that 30 ms of latency between me and my ISP's own network?

 

[dave_the_nerd]

Can't host your own servers. Is there any way to remove some of that 30 ms of latency between me and my ISP's own network?

Well, it's total latency between different points on their network, not between you and them. Your ISP would have to do some network changes/tuning. It can't hurt to ask, but you'd have to figure out which network engineer to bribe.

That said, your ping is well within spec for a residential connection, so realistically, they'll tell you to go sit on it.

 

[Mushkins]

45 ms vs 13 ms is not a meaningless distinction in online games with no lag compensation. I want the numbers down for playing online games primarily. 30 ms vs 40 ms is very very clearly noticeable for me, so 12 ms vs 45 ms would be night and day. $15 per month isn't very significant for me so I wouldn't mind paying it to decrease latency by that much if possible.

Ah, I had a feeling it was one of these superhuman gamers that insists their body can clearly distinguish a 10 millisecond difference. For the record the average human reaction time to visual stimulus is 250 milliseconds, your brain and body can't even react to the difference between 45ms and 10ms much less process and act upon it.

Regardless there's nothing you're going to be able to practically do to turn that 40ms into 15ms. You're well within spec for residential internet and your ISPs infrastructure is what's causing the difference in latency. Buying a new router or paying for a VPN service will make no difference.

The only thing you *could* do is subscribe to a business class fiber ethernet service for essentially a dedicated path to the server. Last time we had a quote for one of these cross my desk it was about $2000 a month for a 50Mbps connection, and even that was only guaranteed to be less than 50ms. A consistent 10ms connection to a remote server over the internet is a pipedream.

 

[fkoehler]

I wouldn't call BS on his claims of noting a difference.

I remember going from 60 to 30-35ms on Quake back in the day, and it was noticeable, and I was able to do better.
Tribe is ancient, and I never got into it, however unlike Q2 I think, lag prediction was non-existent IIRC. So, I believe he most likely can tell a difference.
His only solution is to find closer servers, like in KY.

 

[azazel1024]

The thing is small amounts can be noticeable, but your body still can't react that fast. Of course the difference is, with your body's reaction time, 200-250ms, you add the lag time on to that. So if there is 10ms of lag, it'll take you 210-260ms to react to what the actual movement was. 40ms of lag means you'll actually react 240-290ms after the fact.

That said, 10-20ms difference in lag is effectively an imperceptible difference in lag for a human. Even a super perceptible one. You can say you see a difference, but I'd promise you if you were tested on it without knowing the lag numbers involved, you'd never test to a statistically significant result in determining which is which. 50ms difference, sure, 10-20ms, no. That is roughly 1 frame at 60Hz.

But it doesn't matter a whole lot as others have mentioned, there is very little you can do about it in this case since nearly 30ms of the lag is generated by your ISP's network. Now if you could find a different ISP, that might change.

 

[Ohrami]

I don't much care to discuss the "difference" some number of milliseconds matters, especially if you start to act like reaction time has anything to do with perception time. I have done input lag tests where I can clearly feel 20 ms of difference on mouse lag (the test lets you wave your mouse around, and one half of the screen has artificial input lag, and you have to choose which side of the screen was given artificial input lag, I got 25/25 right), but that was a long time ago and I can't find the test anymore. Long story short, if you can't perceive 20 milliseconds of lag (especially in terms of Tribes: Ascend where you shoot slower and literally have to aim farther ahead of people with higher amounts of lag), I'd say you are below average at perception. Even on a much more basic level, I'm sure you can realize that with a ping of 45, a person with a ping of 10 will be able to kill you in a FPS game even if they shoot you slightly later. This is much less noticeable than in games which have no lag compensation, but it still makes a difference.

You say 30 ms of the lag is from my ISP's network but I have another traceroute (to a city much farther from me than Chicago) where I get 30 ping overall but yet it supposedly takes over 30 ms to leave my ISP's network. How could this be possible? http://puu.sh/km7DA.png

 

[mxnerd]

If you have a Super speed railroad between New York and Los Angeles, you will arrive faster than driving from Los Angels to Sacramento via local routes.

ISP providers make their decisions how they build the routes. Closer cities not necessary have direct route between them.

 

[Ohrami]

I understand that, but how can my ISP be the limiting factor in the Chicago case, when in the case of the Manassas server I posted, it takes over 30 ms to leave my ISP, and I still have below a 30 ms overall latency?

 

[mxnerd]

I'm in Southern California

and this is my ping time for google.com (San Jose) and leaseweb.com (Virginia)

maybe only 10ms diff? sometime almost the same.

Customer has no control.

Ask if they have a trial.

 

[Ohrami]

Are you sure google.com is in San Jose? Google has servers around the world and sometimes it doesn't route you directly to the one closest to you. Sometimes I ping 28 (meaning it's definitely in Virginia, the only common place with lots of servers in it that I get a ping that low), and other times I ping much higher despite my pings on all other servers remaining the same.

Edit: I looked up that exact IP and it seems you are right, it's in San Jose. Still, the explanation doesn't really make sense. It takes 30 ms for me to leave my ISP's network in a server where I get below 30 ping, as well as a server where I get above 30 ping.

 

[mxnerd]

I don't know. I never worked for ISP. Maybe someone here has better explanation.

 

[mxnerd]

I was curious about NYSE (New York Stock Exchange) and ping it.

What a surprise! It's in San Francisco!

Only 14ms!

Tracert result, only 7 hops! google and nyse are in about the same area, yet the huge ping time difference.

Save 3 meals and you have the budget to test.

 

[sdifox]

Go ftth.

 

[Mushkins]

I understand that, but how can my ISP be the limiting factor in the Chicago case, when in the case of the Manassas server I posted, it takes over 30 ms to leave my ISP, and I still have below a 30 ms overall latency?

Because it's two entirely different routes despite being still in the network of your ISP.

Connection one might be going You > Local Node > Building A > Router B > {Internet} > ISP B > Chicago.

Connection two might be going You > Local Node > Building C > Router D > dedicated AT&T fiber line > ISP C > Manassas.

Your ISP locally has hundreds, if not thousands of network devices spread across their entire local service area. Congestion on specific parts of that network, misconfigured devices, routing protocols working as designed, and a dozen other things all uniquely determine the path for every single connection, and a connection from A to B isn't even necessarily exactly the same every time you make it. That's just how the internet functions, there's nothing you can do about it to manage it at the level you'd like.

EDIT: And it just struck me. This is now the third time we've had this exact same discussion on this forum, with the exact same answers being provided, and each time you've been extremely resistant to believe the answers being given to you :/

 

[Ohrami]

How does a traceroute even work? I know plenty of people who use proxy services to drastically decrease their latency, however you insist that it will not work for me because of the amount of time (30 milliseconds) it takes to leave my ISP's network. However, when I get below 30 ping, it still takes over 30 ms to leave my ISP's network. How is this possible?

Go ftth.

I would but I don't think there are really any real residential options here in Louisville yet. Google Fiber announced that they are coming to this city a couple of weeks ago so I'm hoping for that, but even if they do come I'm imagining it will take over 1 year, maybe 2, so I'd like a solution between now and then.