2000 Players on one server w/ no lag

[ry60003333]

Only the registered members can see the link.
This is the most powerful RuneScape private server. Ever.

You can check out an awesome video of this here:
YouTube - 2000 Players on Blitz (Only the registered members can see the link.)


Update: Part 2, w/ chatting/walking: YouTube - 2k Players Part 2 (Only the registered members can see the link.)

Only the registered members can see the link.


Many of you know that today's private servers aren't capable of holding that many players, and can easily be crashed by bots that connect massive amount of players to them, like so:

Only the registered members can see the link.

You also probably know that the real RuneScape worlds can each hold 2k (2000) players each. I think that makes 2000 players an excellent goal amount for private servers, and a great test to see how powerful a server really is.

Hyperion was the second server to achieve this goal, the first being Frugooscape, but the CPU usage of the server was getting extremely close to 100% (with multithreading), and the bots were doing nothing but standing. The bots were also distributed around the world to lessen the server load.

Now, my Blitz Server is the third server to ever do this, but I have put all 2,000 of the bots into the Castle Wars arena, and we've done it with less then 40% CPU usage on the Core handling the player code thread. Hyperion employes multiple task threads, while I employ one for player updating, and the other cores are used for IO.



2000 players in one general area on one server with no lag
Only the registered members can see the link.

Same server with a 508 client
Only the registered members can see the link.


Even with the 2,000 bots connected to it, each server tick still took only 100 milliseconds to complete, leaving the server completely lag-free.



You can check out an awesome video of this here:
YouTube - 2000 Players on Blitz (Only the registered members can see the link.)

This is the most powerful RuneScape private server. Ever.

[Billy]

Well usually 317 is non-lag anyways ;P 503+ is lag.

[Codeusa]

Very nice, I haven't seen any of your work in a long time. So in your general opinion IO is better for a rsps? Because atm I'm using NIO.


Well usually 317 is non-lag anyways ;P 503+ is lag.

Which explains why every 317 I go on lags with 25 players. But my 508 can hold 600 players before it starts to lag? It's not the revision, it's how you program it.

[Anthony`]

This is quite magnificent ry, great job.

[Joshua F]

Wow, very nice, mind putting in your check lag command? :)

[samuraiblood2]

If the bots are connecting only from your own computer then the lag wouldn't surmount to the same as if you truly had two thousand players. Normally each player would be doing a different action each, and probably more then one action at a time. Also, they would be connecting from different computers which is what might lag a player more so then another.

Also, me and blake made a server that could hold over two thousand bots, using my computer connecting from his.

[Brayden]

pretty amazing I wish I could do that.

[Trapt]

L0l Nice I like all the bots :p

[Fenway`]

pretty cool but i dont see any server being bigger than frugooscape was tbh

[Codeusa]

If the bots are connecting only from your own computer then the lag wouldn't surmount to the same as if you truly had two thousand players. Normally each player would be doing a different action each, and probably more then one action at a time. Also, they would be connecting from different computers which is what might lag a player more so then another.

Also, me and blake made a server that could hold over two thousand bots, using my computer connecting from his.
Your server couldn't render 2000 players let alone hold them.

[Trapt]

Hey duchebag, go do something usefull and clean your moms pussy.

[samuraiblood2]

Your server couldn't render 2000 players let alone hold them.

Your a complete idiot who cant even provide sufficient code to prove yourself knowledgeable in anything except proving your own idiocy and ignorance.

Edit: I think ill play the same game you played with me except I wont be an asshole about it. Tell me what this code does and your not an idiot.


List<Object> list = Arrays.asList(a);
Collections.reverse(list);
return list.toArray();

[Fenway`]

Your a complete idiot who cant even provide sufficient code to prove yourself knowledgeable in anything except proving your own idiocy and ignorance.

Edit: I think ill play the same game you played with me except I wont be an asshole about it. Tell me what this code does and your not an idiot.


List<Object> list = Arrays.asList(a);
Collections.reverse(list);
return list.toArray();


my hopes of being a halfway decent coder just went down the drain

[Eugene]

This is magnificent. Whats your RAM? and Processor?

[Aaron]

Very nice source:D

[Razor]

Lol nice, and funny seeing all the bots xD

[Sunni]

This isn't really that "great" if there bots off of your computer then yes, your not going to lag.If there not off your computer indeed there is going to be a slight lag if not a major lag issue.

[Mohammad]

It's pretty easy just make the connection to 0 and then there and no lags let me log in and put 100ms connection and your computer will start to show some smoke ;).

[What teh Fux]

2000 bots is different to 2000 players you realise lol.

[TehCow]

One word.
WOW.

[Ryan]

Which explains why every 317 I go on lags with 25 players. But my 508 can hold 600 players before it starts to lag? It's not the revision, it's how you program it.

Because, your probably using Delta.

[Jusstin]

holy shitttt, this is awesome. i dont care how many players this ACTUALLY is.. this is epic !@ that last pic, its like a war..

[jrune_]

Not bad, but remember they are just bots.

[Glen]

Very nice source:D

Your an fucking idiot. Why don't you lay off and stop spamming the Community Forums and actually read the original post for once? He wasn't showing the source.

@topic, thats pretty cool. You code or no?

[Mike]

y would u do that, like ppl make servers and have alot of work in it, while they fecking keep getting attacked by those bots the day, dont do this crap, i had this on my server..

[leanbow12]

nice localhost

[bloodpk3r]

y would u do that, like ppl make servers and have alot of work in it, while they fecking keep getting attacked by those bots the day, dont do this crap, i had this on my server..

lol the point of this thread is the stability of his server not how many bots hes attacked other servers with.. i think you missed the point of the first post

[Shishir G]

This is BS all the bots were run through his own Computer, which don't cause as much lag as if they were from another one half way across the world.
+
They're just standing there, not moving if im correct.
REAL users move, they do things which will cause lag.
So this stability test is extremely un-accurate.

[Waynex]

nice localhost

That's what I was thinking.

[Codeusa]

Because, your probably using Delta.

"Which explains why every 317 I go on lags"

In this form of speech go on means to play. l2englishgrammar.

[What teh Fux]

"Which explains why every 317 I go on lags"

In this form of speech go on means to play. l2englishgrammar.

dude shut up, you dont know what your talking about. so much for your post on my thread saying you dont do rsps anymore its a rsps forum you moron. you seem to go around acting like a jerk thinking your alot better than you are. lol'd at you claiming u can write gpi on my thread you wouldnt have a shit show. and u think i leeched mine? your wrong i can even get vouches for it moron.

[Ryan]

"Which explains why every 317 I go on lags"

In this form of speech go on means to play. l2englishgrammar.

No shit.

You're a complete ass hole.

[Ramsin]

Let me send you my 2000 bots.... bet cant handle 600-800. And all it does is talk and run around.. and a command if i choose to.

[Robbie]

Its impressive, you dont have to be assholes about it.

2000 players all rendering in the same location without lag? I know they are inactive but still, Jagex can't render that many players in a small location.

[Ryan]

Its impressive, you dont have to be assholes about it.

2000 players all rendering in the same location without lag? I know they are inactive but still, Jagex can't render that many players in a small location.

Umm, I'm pretty sure Rune Scape is much more stable then this.

[Robbie]

Umm, I'm pretty sure Rune Scape is much more stable then this.

Ofc, and RuneScape has all that content to manage too :P

About the character rendering, eg; if you go to Runescape W2 grand Exchange, you can only see the characters within around 10 squares of you, even though the entire grand exchange is covered in players.
I believe its the 256 closest players to you that are displayed.

[Marc]

If they were all bots you bet your sweet ass Jagex could do it.

I want to see you have 900 players in a functional Soul Wars style minigame, with all of them performing multiple actions and multiple itterations for the minigame itself. Not that this isn't impressive, but you're making it seem quite more powerful then it actually is.

[Codeusa]

dude shut up, you dont know what your talking about. so much for your post on my thread saying you dont do rsps anymore its a rsps forum you moron. you seem to go around acting like a jerk thinking your alot better than you are. lol'd at you claiming u can write gpi on my thread you wouldnt have a shit show. and u think i leeched mine? your wrong i can even get vouches for it moron.

When I said I used NIO it's for my 2d game.. hurrrrrr

[Robbie]

If they were all bots you bet your sweet ass Jagex could do it.

I want to see you have 900 players in a functional Soul Wars style minigame, with all of them performing multiple actions and multiple itterations for the minigame itself. Not that this isn't impressive, but you're making it seem quite more powerful then it actually is.

Most servers couldn't manage 10% of that.

[Ramsin]

10% is only 200, most server could handle 200 bots trust me.

Probably 40%.

[Ryan]

200? Then isn't 100% like 2,000?

Oh, never mind. The thread is talking about 2,000 players, witch 10% is 200.

I was thinking about what Marc said about 900 players.

[Faab234]

If the bots are connecting only from your own computer then the lag wouldn't surmount to the same as if you truly had two thousand players. Normally each player would be doing a different action each, and probably more then one action at a time. Also, they would be connecting from different computers which is what might lag a player more so then another.

Also, me and blake made a server that could hold over two thousand bots, using my computer connecting from his.

Did you released that server?

[samuraiblood2]

Did you release that server?

No, we got bored of doing the essentials (shopping, banking, etc) and never finished it.

[dragonkk]

Tryed that game and got banned for pj.... It has a noob community and its borring pk as dam ^^.

[Emperor]

What's so special about this?
Hyperion can hold 2000 players aswel, localHost also holds way more then actually putting it online does.

Anyone understanding what they're doing (Or using Hyperion as base, and not fucking it up) can do this.

[Codeusa]

What's so special about this?
Hyperion can hold 2000 players aswel, localHost also holds way more then actually putting it online does.

Anyone understanding what they're doing (Or using Hyperion as base, and not fucking it up) can do this.
Hyperion networking & threading is shit, yes it can hold 2000 players. but it will also take your computers process to the maximum.

[Emperor]

Correct.

[Relapse]

Umm, I'm pretty sure Rune Scape is much more stable then this.

In Runescape you only see 255 (or something like that) of people that are in your area, 2000 bots is similar to 2000 players..

2000 players doing things like pking, pking is just animations and hit updates every few seconds, the thing that would lagg a server the most is if 2000 players were running around

[fail brid`]

very nice ry, what is it hosted on?

[Canownueasy`]

What networking api are ya using?

[ry60003333]

Jus to clarify some things, this is using the Java NIO networking, and was NOT done using localhost. samuraiblood2, words mean nothing. You need proof that your server could do this. Also, many of you think that there is some kind of magical difference between bots and players, and you need to wake up. To the server, connections are connections, and packets are packets. The bots send the exact same packets as a player idling, and that is a fact. I've also updated the post with a link to the second video where the bots walk, talk and chat.


Update: Part 2, w/ chatting/walking: YouTube - 2k Players Part 2 (Only the registered members can see the link.)

Only the registered members can see the link.

[Divine]

Rofl you guys realize that he literally either ddos'd himself or ddos'd someone else to get those pics -- right?

[Rog3r]

Rofl you guys realize that he literally either ddos'd himself or ddos'd someone else to get those pics -- right?

What........

[Divine]

What........

Are those pics suppose to be real, or just showing how much the server can hold?

[Glen]

Read the OP. He is showwing how stable the server is.

[Divine]

Read the OP. He is showwing how stable the server is.

Lol well dat was a fail on my part. sowwy.

[ry60003333]

Are those pics suppose to be real, or just showing how much the server can hold?

If you watched both of the videos you would realize very quickly that the pictures are real. ;)

[samuraiblood2]

samuraiblood2, words mean nothing. You need proof that your server could do this.I would give proof but I cant publicly and that's assuming I still have the specific base in question. Though if I cant find that specific one I do have others that can compete with it.
Also, many of you think that there is some kind of magical difference between bots and players, and you need to wake up. To the server, connections are connections, and packets are packets. The bots send the exact same packets as a player idling, and that is a fact.

Yes but you fail to realize the fact that normal connections aren't coming from localhost, they are coming from (usually) long distances away. The longer the distance between the server and the client, the longer the upload and download time will take. This is why we take speed tests from long distances and not locally. Not only that but the computer and the internet connection involved also take part in possible lag.

[ry60003333]

Yes but you fail to realize the fact that normal connections aren't coming from localhost, they are coming from (usually) long distances away. The longer the distance between the server and the client, the longer the upload and download time will take. This is why we take speed tests from long distances and not locally. Not only that but the computer and the internet connection involved also take part in possible lag.
If you had even watched the second video you would have seen me Teamview into the computer making the bot connections. The server is on a VPS in Europe and the bots were connected from my friends computer in California. :)

[samuraiblood2]

If you had even watched the second video you would have seen me Teamview into the computer making the bot connections. The server is on a VPS in Europe and the bots were connected from my friends computer in California. :)

Then that's obviously different from the first one now isn't it? By the way, are the bots using proxies?

[Wise Old Man]

Actually, each runescape world can hold 2048 players. You're client would crash when the 2047th player logs in though.[If they didn't set the max at 2000] ;D

[ry60003333]

Then that's obviously different from the first one now isn't it? By the way, are the bots using proxies?
No, it was the same connection setup in the first video, we just didn't anticipate that everyone would go "o itz localhost". All the connections were done directly from my friend's house in California to the VPS, over the internet, with no proxies. You can see that in the second video, where I did Teamviewer just to prove that very point.

[samuraiblood2]

No, it was the same connection setup in the first video, we just didn't anticipate that everyone would go "o itz localhost". All the connections were done directly from my friend's house in California to the VPS, over the internet, with no proxies. You can see that in the second video, where I did Teamviewer just to prove that very point.

Use proxies then and see how the server holds up after it being connected from random IP's. I also looked at the video, the "mutli-version support" has been done before so you know. By the way, I don't mean to offend, I'm just naturally sceptical (especially with anything involving the RSPS community).

[ry60003333]

Use proxies then and see how the server holds up after it being connected from random IP's. I also looked at the video, the "mutli-version support" has been done before so you know. By the way, I don't mean to offend, I'm just naturally sceptical (especially with anything involving the RSPS community).

You think changing the IP address that they are coming from will make a difference? You obviously don't understand how the internet or TCP/IP works if you do. The server is running on a VPS, and the connections came from over the internet. The destination IP doesn't matter as long as the packets have to leave the computer, then its up to internet routers to get the packets to their destination.

No, I was actually the first with multi-version, and now other people have started on it. I started this project three years ago, I was definitely the first to have it, and I am still the only one who has been completely successful with it.
Only the registered members can see the link.
Only the registered members can see the link.

You are way beyond sceptical, you can't even see what is obviously right in front of you.

[samuraiblood2]

You think changing the IP address that they are coming from will make a difference? You obviously don't understand how the internet or TCP/IP works if you do. The server is running on a VPS, and the connections came from over the internet. The destination IP doesn't matter as long as the packets have to leave the computer, then its up to internet routers to get the packets to their destination.

Actually I do understand how TCP works, its like its brain science. Right now I'm not in the mood to argue, Ill edit this post or respond later to the flame/spam/troll Ill get for not making a rebuttal now. As for the multi-client thing, I think your client may actually be the one I was referring to though I don't remember exactly (the pictures look familiar).

[Najca]

Thats many noobs!

[_1Greg1_]

You used a mac :). I have a mac to host my server.

[Graham_]

Hyperion networking & threading is shit, yes it can hold 2000 players. but it will also take your computers process to the maximum.

It is designed to do that to run faster... the servers which use single-threaded updating systems aren't using all the available power in a multiple core system.

The networking code in Hyperion (MINA) has been used by many commercial projects and can handle loads well above what you get in a typical RuneScape server.


Yes but you fail to realize the fact that normal connections aren't coming from localhost, they are coming from (usually) long distances away. The longer the distance between the server and the client, the longer the upload and download time will take. This is why we take speed tests from long distances and not locally. Not only that but the computer and the internet connection involved also take part in possible lag.

That doesn't make the server itself lag though. While those packets are taking their 200 milliseconds to get across the world, the server doesn't know about them yet and is happily processing away. Unless your server sucks and blocks something for one player while waiting for a response from another player, one player having a slow connection won't make a difference to anyone else or the server itself.

The only thing I can think about that would skew localhost tests is the fact that you can send much more data locally than you can on your standard home ADSL connection within a time period. However, on the Internet, many servers have 100 mbps or 1000 mbps links which are definitely adequate. In fact, having the bot and server running on the same PC is going to take CPU power away from the server so the bot can run, so tbh it probably evens itself out.


Use proxies then and see how the server holds up after it being connected from random IP's. I also looked at the video, the "mutli-version support" has been done before so you know. By the way, I don't mean to offend, I'm just naturally sceptical (especially with anything involving the RSPS community).

Random IP addresses won't make a difference as ry60003333 said. All internet traffic has equal priority, unless you have some firewall on your end that suddenly went 'omg why is there loads of packets from this same IP? better drop those packets'. As far as I know, he was also the first one to do the multi-version support.


Actually, each runescape world can hold 2048 players. You're client would crash when the 2047th player logs in though.[If they didn't set the max at 2000] ;D

2046 actually. 0 is not used in the lists and 2047 is used as a magic number to signify that some block updates occurred.