NS2 Server Comparison
ScardyBob
ScardyBob Join Date: 2009-11-25 Member: 69528Forum Admins, Forum Moderators, NS2 Playtester, Squad Five Blue, Reinforced - Shadow, WC 2013 - Shadow
ScardyBob Join Date: 2009-11-25 Member: 69528Forum Admins, Forum Moderators, NS2 Playtester, Squad Five Blue, Reinforced - Shadow, WC 2013 - Shadow
So there has been a fair bit of discussion about NS2 server availability, hardware requirements, performance, etc. lately that I thought it might be useful to do a little analysis with the current server data available. Unfortunately, the only data I've been able to find is max and avg playercount per server (from <a href="http://www.play4dead.com/ns2/" target="_blank">http://www.play4dead.com/ns2/</a> and game-monitor.com). However, that doesn't mean we can't make some conclusions about the aforementioned details. These two variables are indirectly related to the things we do care about in a NS2 server (i.e. players tend to play on the highest playercount servers with good performance).
Here's the data I'm talking about (ranked by region and avg playercount)
<a href="http://imgur.com/eNvmX" target="_blank"><img src="http://i.imgur.com/eNvmXl.png" border="0" class="linked-image" /></a>
As you can see, it largely confirms our anecdotal information that Australia-Pure dominates in AUS/EU, HBZ in EU, and Inversion in the US (Darkside is run off the same server as the Inversion ones, so I'm clumping them together). If we just consider the max playercount, we get the following info
<ul><li>Average max playercount: 13.94</li><li>Median max playercount: 14</li><li>Mode max playercount: 16</li></ul>
Meaning that the most common server max playercount is 16, the one in the middle is 14, which is basically the same as the average (13.94).
Another way to look at this data is to do a weighted average based on the average playercount (i.e. we want to see what are the stats of the servers people actually play on rather than the ones available). I've done that below and also broke it down to the weighted average by region.
<ul><li>All NS2 Servers (weighted average): 14.65</li><li>AUS/NZ Servers (weighted average): 14.31</li><li>EU Servers (weighted average): 13.45</li><li>US/NA Servers (weighted average): 15.74</li></ul>
This is where it gets interesting. Essentially, the weighted average by region is showing a rough estimate of the max playercount of the servers people generally play on, and its basically saying that players in the US/NA are playing on servers that are over 1 max playercount higher than in AUS/NZ and over 2 max playercount higher than in EU. While there are many reasons for server owners to limit their max playercount, the biggest is likely performance issues. This implies that EU NS2 players are more or less getting screwed on the availability of high performance NS2 servers in comparison to US/NA and AUS/NZ players.
The bottom line is that EU players probably need another high quality server similar to the HBZ/Inversion/Australia-Pure servers to satisfy the playerbase at the moment.
Also, for people who want to play with the data, I've uploaded it to <a href="http://dl.dropbox.com/u/5490563/NS2%20Server%20Comparisons.xlsx" target="_blank">dropbox (xlsx)</a>
Here's the data I'm talking about (ranked by region and avg playercount)
<a href="http://imgur.com/eNvmX" target="_blank"><img src="http://i.imgur.com/eNvmXl.png" border="0" class="linked-image" /></a>
As you can see, it largely confirms our anecdotal information that Australia-Pure dominates in AUS/EU, HBZ in EU, and Inversion in the US (Darkside is run off the same server as the Inversion ones, so I'm clumping them together). If we just consider the max playercount, we get the following info
<ul><li>Average max playercount: 13.94</li><li>Median max playercount: 14</li><li>Mode max playercount: 16</li></ul>
Meaning that the most common server max playercount is 16, the one in the middle is 14, which is basically the same as the average (13.94).
Another way to look at this data is to do a weighted average based on the average playercount (i.e. we want to see what are the stats of the servers people actually play on rather than the ones available). I've done that below and also broke it down to the weighted average by region.
<ul><li>All NS2 Servers (weighted average): 14.65</li><li>AUS/NZ Servers (weighted average): 14.31</li><li>EU Servers (weighted average): 13.45</li><li>US/NA Servers (weighted average): 15.74</li></ul>
This is where it gets interesting. Essentially, the weighted average by region is showing a rough estimate of the max playercount of the servers people generally play on, and its basically saying that players in the US/NA are playing on servers that are over 1 max playercount higher than in AUS/NZ and over 2 max playercount higher than in EU. While there are many reasons for server owners to limit their max playercount, the biggest is likely performance issues. This implies that EU NS2 players are more or less getting screwed on the availability of high performance NS2 servers in comparison to US/NA and AUS/NZ players.
The bottom line is that EU players probably need another high quality server similar to the HBZ/Inversion/Australia-Pure servers to satisfy the playerbase at the moment.
Also, for people who want to play with the data, I've uploaded it to <a href="http://dl.dropbox.com/u/5490563/NS2%20Server%20Comparisons.xlsx" target="_blank">dropbox (xlsx)</a>
Comments
(Hardware = i5 4,5Ghz etc. like most of the top servers)
I'll try to keep it running more often...
Not a solution, but i guess it could help - i see way to many ppl playing on crappy low tickrate servers. :X (which is more or less all eu servers except HBZ)
PS: no mini timeouts anymore on my connection (ppl that played on my server will know what i mean)
(Hardware = i5 4,5Ghz etc. like most of the top servers)
I'll try to keep it running more often...
Not a solution, but i guess it could help - i see way to many ppl playing on crappy low tickrate servers. :X (which is more or less all eu servers except HBZ)
PS: no mini timeouts anymore on my connection (ppl that played on my server will know what i mean)<!--QuoteEnd--></div><!--QuoteEEnd-->
Frankly, just running it for a couple of days to a week after each build is released would be enough. There probably is enough of a EU playerbase to keep it going 24/7, but that can be hard if you have limited upload or a bandwidth cap (my ISP has an informal 250GB/month limit).
I don't think that it implies that at all. It doesn't matter what server you are playing on, a 16 player server will drop in tickrate once a lot of buildings start going up.
Personally I start getting poor FPS when playing 7v7 and above even if the tickrate is fine. I think it makes sense to keep the limit at 14 because it will be a better experience overall for the majority of players.
HBZ actually run 2 servers sometimes. I guess they might put the second one online when a new build gets released. I don't see a need for it right now as it's hard enough keeping one server full. With Kuruyo's server as well I think the EU is fine for now. By the time a lot more people start playing the server requirements should be quite a bit lower and you'll start to see more servers pop up.
HIP HIP, HURRAY! :D
<!--quoteo(post=1886441:date=Nov 24 2011, 05:34 AM:name=Wilson)--><div class='quotetop'>QUOTE (Wilson @ Nov 24 2011, 05:34 AM) <a href="index.php?act=findpost&pid=1886441"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->I don't think that it implies that at all. It doesn't matter what server you are playing on, a 16 player server will drop in tickrate once a lot of buildings start going up.
Personally I start getting poor FPS when playing 7v7 and above even if the tickrate is fine. I think it makes sense to keep the limit at 14 because it will be a better experience overall for the majority of players.
HBZ actually run 2 servers sometimes. I guess they might put the second one online when a new build gets released. I don't see a need for it right now as it's hard enough keeping one server full. With Kuruyo's server as well I think the EU is fine for now. By the time a lot more people start playing the server requirements should be quite a bit lower and you'll start to see more servers pop up.<!--QuoteEnd--></div><!--QuoteEEnd-->
I think you are dead on with 14 being the sweet spot right now. It allows for 7v7 pub and 6v6 matches with casters, and a stable tickrate even in long games. Perhaps endar's 5ghz OC will allow for more?
I think we would see a much larger active community if half the people that try out a new patch didn't join a garbage server, see the game rubberbanding and decide to wait for the next patch. I only played for the first 10mins and saw 5 people leave as the server was struggling (HBZ didnt auto update it seems).
+1 No more Koruyo-lag.
The only thing that's is missing in the analysis of the server is the average tick rate while in load.
We could then see which server should be used. In Europe we sometimes see medium-low servers full when better ones are empty.
The #156 West server has now been updated to a i7 2600K. It has a slight overclock at the moment but will be ramping it up to around 4.5 GHz in the next few days.
It will run a Summit map and also a Tram map.
Server might be down most of the day on Nov 26th due to moving it into a computer room.