Blizzard Server

[thiartc]

is there enough SA players to fully populate a server?

How many peeps on a server?

 

[charlieharper]

Not gonna happen - give up hope now, dont waste your time...

I also think so. But do you think the latency of the game will be better??

 

[werries2]

Taking a look at the Australians trying to get a server for the last 4 years now. They have significantly more people playing wow than us and had cheaper and better broadband for a long time already. I really doubt that even if we fill a cluster of servers we will get it.
Nice dream though.

 

[werries2]

is there enough SA players to fully populate a server?

How many peeps on a server?

20000 people per server?
Guess we have enough people but is that economically viable to put a server here and have its own maintenance crew here?

 

[werries2]

I also think so. But do you think the latency of the game will be better??

Latency may improve with a less congested cable. But only time will tell.

 

[BrianStephan]

Latency should actually improve 5 fold

 

[werries2]

Latency should actually improve 5 fold

Not

 

[BrianStephan]

Not what I have been told by a good friend very high up in Telkom

 

[Xavierza]

yeah, would be awesome to have an official server in SA. Blizzard are just lazy, they make millions in profits and we most definitely do have enough players to fill a server, you can be sure every SA player would transfer back here. Besides, Blizzard can't bitch about costs, they would be paying maintenance staff in rands, all their costs in rands, so would work out rather cheap for them to run a server here.

Honestly I think it's more of an admin problem, do they really want to go to all the effort required to come run a server here just for us...

 

[Jase]

Why would Blizz bother to spend on that kind of infrastructure when we all play on the EU / US servers anyway.

Would subscriptions jump in SA if we had our own server?

 

[werries2]

Not what I have been told by a good friend very high up in Telkom

Find some friends that actually know something about the Internetet perhaps?

For latency to improve 5 fold (from best levels at present 250ms) would imply that we will have ~50ms latency or more conservatively 100ms latency... The problem with those speeds however is that it is physically impossible to achieve over a distance of 15 000km.

We discussed this some years back on these forums already.

The Internet operates at near light speed, which on a planet the size of Earch often practically amounts to near real-time.

Digital information such as Internet packets travel at 2/3 of the speed of light on copper wire and on fiber optic cables. Since light speed is about 300,000 kilometers a second, this means digital communications travel at about 200,000 kilometers a second, slowing down only because because copper and fiber optic materials are about one-third thicker than a vacuum.

At this speed and neglecting switching delays, two computers have to be more than ten thousand kilometers apart, or almost half way around the world, before they experience a tenth of a second in communications delay.

With fixed near-optimal transmission speed, there are only two ways to make Internet networks faster -- increase the number of bits that are traveling at once down the connection, or increase the speed at which you switch them from one connection to another at the junction points.

Internet routers are getting faster and faster with switching speeds nearing instantaneous, while fiber optics and wireless technologies are enabling networks to send much larger numbers of bits at once. The Internet is getting even faster.

Thus what I understand from this is that yes as you said it is about 2/3 the speed of light, but Copper and Fibers have the same travel time. Fibre is only better in that you can send more data on them. (that is A LOT more!) Latency or Pings however will be the same with either, and will be slowed down due to switching and the actual travel distance.

Ok, so lets make a few calculations...
The speed data travel via cable is 200,000,000 meters per second
thus 200.000 km/ms
this would give 75 ms travel time for 15 000 km.
Latency however requires a send and response thus 150ms is the theoretical best latency achievable. Line size or capacity have nothing to do with this calculation so the best latency you can achieve now will also be the best latency you can achieve with a thousand new undersea cables. A bit hard to break the laws of physics.

My comment that we may achieve better latency is that with a less congested cable from here to europe, there may be a lot less packet losses and we finaly can get closer to the theoretical best latency. A 5 fold improvement however is a bit silly.

 

[BrianStephan]

Thanx for the calculations but I phrased it wrong it will improve significantly to what we are used to say in terms of 50ms so if u had 250-400 u can do the math. Which in my book is great. And where just dreaming about getting a server here wont happen unless it snows in Durban and how many times has that happend