Arument for a 10mb line vs 4mb

[Sinbad]

Question - the OP is asking about accounts. Assuming a 10mb line and a 4mb account, is the uplink really throttled to 512k at the ISP end? Wouldn't he get the benefit of 1mb up anyway?

 

[Chevron]

Just thought I'd mention the 6Meg Silver is on special now:

http://openweb.co.za/special.html

 

[w1z4rd]

Sadly it is not as simple as you make it out to be, and because I can not find the article I used in the past as reference, I will stop the troll here and say, the evidence I have seen has shown me that higher speed links give lower latency. What I can tell you is, the time on ADSL changes when you go to faster speeds, there are more lanes for the packets to be split up in to. For example:a 64byte ICMP ping from windows won't fit into a 53byte ATM cell, and ATM is still used for backhaul in many places. So what happen is, when you go from a 384k line which has a 128k upload speed, to a 10mb line with a 1024k upload speed, there are more lanes for packets to be split up into, which mean they can be pushed out at the same time as opposed to waiting in a queue to be pushed through limited lanes.

Again, you are ignoring that its a tiny bit of data. 32 bytes! (not the 64 bytes you are misleading people with) If I have a 64Kb line 32bytes can go through unimpeded. If I have a 10Mb line that 32bytes can go unimpeded.

This is basically the logic. With 32bytes on a 64k lane there are already MANY lanes open for it that is not using. Lets say 10 lanes are unused. Adding another 100 lanes is not going to make that one piece of traffic go faster (if there is no load).

If there was congestion. IE you had a virus you could not pin down or didnt know about... and data was being used so there were lots of pieces of traffic then yes. More lanes = more traffic.

I understand there are different back-haul technologies that may be used, but I have already covered that in a previous post.

Either way, its so insignificant, that if you don't want to believe me, it is not the end of the world, I'll rather just leave you spreading what you believe in.

We are having a discussion about data rates. Not who your fking deity is.

 

[Tinuva]

Again, you are ignoring that its a tiny bit of data. 32 bytes! (not the 64 bytes you are misleading people with) If I have a 64Kb line 32bytes can go through unimpeded. If I have a 10Mb line that 32bytes can go unimpeded.

False. No misleading, I do my tests from linux, where it is 64bytes or 56bytes depending on which version of ping you have.

This is basically the logic. With 32bytes on a 64k lane there are already MANY lanes open for it that is not using. Lets say 10 lanes are unused. Adding another 100 lanes is not going to make that one piece of traffic go faster (if there is no load).

Yes and thousands of years ago the logic was the earth is flat because that is all you could see, but here we are now, and we know better.
Who gives a **** about windows icmp latency. Let's get real. Games use more than 32byte packets, they are guaranteed to be at least 1kbyte/sec, generally closer to 5kbyte/sec, at which point it most definitely matters how many lanes you have available.

While that may not explain why latency even on windows icmp pings are better on faster lines, a faster line does mean better latency, if location and distance isn't part of the equation.

I had 2x adsl lines at my place at one stage, a 384k one, and a 4096k one, go take a guess which one had better latencies, when ONLY a game is played on either one with no other traffic.

Nothing in your posts were ignored, but hey, like I said, keep on spreading what you believe, I won't be replying again.

 

[NomNom]

Actually it is not misleading.

When my line went from 384k to 4096k latency dropped by 5ms to the first hop, on an idle line. Speed most definitely does make a difference to latency, although 5ms to most is insignificant. Now that my line is 10Mbps, latency sometimes are 4ms to the first hop.

Agreed same happened to me, when I upgraded from 384k to 1mb my ping to most local servers halved.

 

[Heisenborg]

When I upgraded from a 0.5U to 1.0U the difference was noticeable, especially because I use teamspeak and Skype while gaming

U talking BS my friend. Pings on a 384k and 10mb line would be the same under identical load circumstances

 

[w1z4rd]

False. No misleading, I do my tests from linux, where it is 64bytes or 56bytes depending on which version of ping you have.

Which version runs 64 byte pings, and yeah, most of the people here run windows. Hence we using that as a standard.

Yes and thousands of years ago the logic was the earth is flat because that is all you could see, but here we are now, and we know better.

Completely pointless statement.

Who gives a **** about windows icmp latency.

Those of us talking about icmp latency. Duh. Thats what a ping is... EDIT: I see now you are now changing the topic to in game ping which is different as there is load. I stated in almost every post that I was talking about no load examples.

Let's get real. Games use more than 32byte packets, they are guaranteed to be at least 1kbyte/sec, generally closer to 5kbyte/sec, at which point it most definitely matters how many lanes you have available.

No ****, and completely irrelevant to my points. We all know conditions change under load (AS I HAVE SAID) and no one said games dont use more than 32byes.

I had 2x adsl lines at my place at one stage, a 384k one, and a 4096k one, go take a guess which one had better latencies, when ONLY a game is played on either one with no other traffic.

A games ping is different. Sorry, I was working with a standard. I was not working with gaming pings. I think thats where your confusion comes in (though its easy to understand that confusion especially considering how have responded previously referring to 64 bytes). Twisty arnt ya?

Nothing in your posts were ignored, but hey, like I said, keep on spreading what you believe, I won't be replying again.

Yes, a game that uses a lot of bandwidth, that has its own internal "ping" system will show a lower "ping". However, I was talking about icmp ping(and I assume up until this post so where you as you used examples of 64 bytes and talk about ping versions). But yeah, an ingame "ping" system like that. BECAUSE there is load!!!! Will show a lower ping with more bandwidth.

Please note, I have stayed consistant with my icmp pings and I clearly stated many times under no load... and specifically I spoke about single packets. You seem to have ignored "my beliefs" on that matter. Hence your confusion.

 

[iJasper]

If I have a 0.5Mbs UPLINK and Play BF3 with Skype on, connected to 3 people in my squad + my sister on FaceTime and have a 1Mbs UPLINK under the same load. It just seems obvious that the 1Mbs would give me the advantage in-game.

Again, I'm talking about REAL WORLD performance. The above situation is what my usual Friday nights are.