|
On May 30 2011 12:48 Cambium wrote:Show nested quote +On May 30 2011 12:41 VIB wrote:On May 30 2011 12:17 L3gendary wrote:On May 30 2011 11:49 VIB wrote:On May 30 2011 11:20 darkness wrote:As far as I know, chasing maphacks is like cats'n'mice. In other words, if you patch a maphack exploit, there will be new one. Correct me if I'm wrong.  Technologically, making the game 100% maphack free is pretty trivial. Just change the network architecture from p2p to client-server. Or at least give players that option if they so choose. The problem is economical and not technological. Blizzard only doesn't do it because it's not worth the cost vs benefit. They would spend considerable money to implement it. And how much more money do they think they would make from a maphack free game? Very little. They don't think it's worth it. This kinds of thing would have to go through the CFO which would simply analyse the ROI (return of investiment) and say no. A maphackable game is more profitable. Huh? The game is not p2p it is client-server. And there are plenty of games that are client-server and hacking is rampant. You're probably misunderstanding what I said. I'm talking about the specific problem of maphacks. Not about general hacks. Maphacks specifically are only possible because the player data are shared between both clients in p2p instead of being handled by a third party server. This problem is 100% solvable with a server as referee. On May 30 2011 12:34 Cambium wrote:On May 30 2011 12:17 L3gendary wrote:On May 30 2011 11:49 VIB wrote:On May 30 2011 11:20 darkness wrote:As far as I know, chasing maphacks is like cats'n'mice. In other words, if you patch a maphack exploit, there will be new one. Correct me if I'm wrong. Technologically, making the game 100% maphack free is pretty trivial. Just change the network architecture from p2p to client-server. Or at least give players that option if they so choose. The problem is economical and not technological. Blizzard only doesn't do it because it's not worth the cost vs benefit. They would spend considerable money to implement it. And how much more money do they think they would make from a maphack free game? Very little. They don't think it's worth it. This kinds of thing would have to go through the CFO which would simply analyse the ROI (return of investiment) and say no. A maphackable game is more profitable. Huh? The game is not p2p it is client-server. And there are plenty of games that are client-server and hacking is rampant. Sorry to say but it's just a reality of online games. Idk any games that handle hacking very well. Most tournaments have their own anti-cheats and lan is obviously not a problem so its mostly just annoying on ladder in that u might run into a hacker every 1 in 100 games or something. I don't see how making it client-server would solve the problems though, it has to do what's being transmitted, not where it's coming from, no? Yes it's what's being transmitted. The problem is that if it's p2p then both clients have to tell each other what they're doing to keep synced. If the game was client-server then the clients would only tell what they're doing to the server and would never know what the other is doing. The server would receive data from both sides, calculate what both can see, and only share with the players what's in their vision. And not what's under the fog of war. that's an interesting idea, i'd never thought about it that way. However, wouldn't that put an enormous amount of strain on the server? I always thought of the client-server model as the server relaying the data and does no calculations. The server would obviously keep track of the frame counter, but that's it. Yes it would. Not enormous. But some strain is more than no strain. And it would cost blizzard money to keep the servers running. They would need to buy more servers and bandwidth to handle it. Like I said. It's not a technical problem, it's an economical problem.
They could even allow us to use our own computers as hosts for custom games. So, for example, on TSL we would use one of the admin's computer as host instead of blizzard's. So it wouldn't strain their servers and it would have 100% guarantee unmaphackable online tournaments.
Blizzard could cut costs by making only important games unhackable instead of all of them. But, honestly, with today's hardware power and price. The cost to keep enough servers running to keep all ladder games above gold unhackable would be certainly small compared to what they make. Problem is that the gains from it would be even smaller (ammount of people who would buy the game only because there's no maphacks). So on the cost vs benefit it's still not profitable to make the investment.
It's not too different from how WoW works btw. WoW is unmaphackable. There's no hacks that allows you to see stealthed rogues in wow, for example. (there are other hacks tho, but those wouldn't affect SC). The technology is old and blizzard knows how do it. It's not like it's a brand new ingenious idea I just brought up.
Just made this image to explain it better:
+ Show Spoiler +
|
On May 30 2011 13:15 VIB wrote:Show nested quote +On May 30 2011 12:48 Cambium wrote:On May 30 2011 12:41 VIB wrote:On May 30 2011 12:17 L3gendary wrote:On May 30 2011 11:49 VIB wrote:On May 30 2011 11:20 darkness wrote:As far as I know, chasing maphacks is like cats'n'mice. In other words, if you patch a maphack exploit, there will be new one. Correct me if I'm wrong. Technologically, making the game 100% maphack free is pretty trivial. Just change the network architecture from p2p to client-server. Or at least give players that option if they so choose. The problem is economical and not technological. Blizzard only doesn't do it because it's not worth the cost vs benefit. They would spend considerable money to implement it. And how much more money do they think they would make from a maphack free game? Very little. They don't think it's worth it. This kinds of thing would have to go through the CFO which would simply analyse the ROI (return of investiment) and say no. A maphackable game is more profitable. Huh? The game is not p2p it is client-server. And there are plenty of games that are client-server and hacking is rampant. You're probably misunderstanding what I said. I'm talking about the specific problem of maphacks. Not about general hacks. Maphacks specifically are only possible because the player data are shared between both clients in p2p instead of being handled by a third party server. This problem is 100% solvable with a server as referee. On May 30 2011 12:34 Cambium wrote:On May 30 2011 12:17 L3gendary wrote:On May 30 2011 11:49 VIB wrote:On May 30 2011 11:20 darkness wrote:As far as I know, chasing maphacks is like cats'n'mice. In other words, if you patch a maphack exploit, there will be new one. Correct me if I'm wrong. Technologically, making the game 100% maphack free is pretty trivial. Just change the network architecture from p2p to client-server. Or at least give players that option if they so choose. The problem is economical and not technological. Blizzard only doesn't do it because it's not worth the cost vs benefit. They would spend considerable money to implement it. And how much more money do they think they would make from a maphack free game? Very little. They don't think it's worth it. This kinds of thing would have to go through the CFO which would simply analyse the ROI (return of investiment) and say no. A maphackable game is more profitable. Huh? The game is not p2p it is client-server. And there are plenty of games that are client-server and hacking is rampant. Sorry to say but it's just a reality of online games. Idk any games that handle hacking very well. Most tournaments have their own anti-cheats and lan is obviously not a problem so its mostly just annoying on ladder in that u might run into a hacker every 1 in 100 games or something. I don't see how making it client-server would solve the problems though, it has to do what's being transmitted, not where it's coming from, no? Yes it's what's being transmitted. The problem is that if it's p2p then both clients have to tell each other what they're doing to keep synced. If the game was client-server then the clients would only tell what they're doing to the server and would never know what the other is doing. The server would receive data from both sides, calculate what both can see, and only share with the players what's in their vision. And not what's under the fog of war. that's an interesting idea, i'd never thought about it that way. However, wouldn't that put an enormous amount of strain on the server? I always thought of the client-server model as the server relaying the data and does no calculations. The server would obviously keep track of the frame counter, but that's it. Yes it would. Not enormous. But some strain is more than no strain. And it would cost blizzard money to keep the servers running. They would need to buy more servers and bandwidth to handle it. Like I said. It's not a technical problem, it's an economical problem. They could even allow us to use our own computers as hosts for custom games. So, for example, on TSL we would use one of the admin's computer as host instead of blizzard's. So it wouldn't strain their servers and it would have 100% guarantee unmaphackable online tournaments. It's not too different from how WoW works btw. WoW is unmaphackable. There's no hacks that allows you to see stealthed rogues in wow, for example. (there are other hacks tho, but those wouldn't affect SC). The technology is old and blizzard knows how do it. Just made this image to explain it better: + Show Spoiler +
yep, crystal clear
thanks
|
Does it not go thru Blizzard's servers? (sc2)
|
On May 30 2011 13:43 L3gendary wrote:
Does it not go thru Blizzard's servers? (sc2) That's irrelevant at this point. I don't even have sc2 so I can't be sure. But even if it does they probably only keep it for archiving. They aren't really doing any filtering of the data in the server. So in practice it's as if the game is p2p.
But to answer your question: When battle.net disconects for some reason. Does the game continues or does it stops? (I remember in bw it would continue, because it was p2p - but if it isn't then the game would stop)
|
There is an extra cost of latency to a client-server architecture. So even ignoring economy, it may be undesirable to implement it.
|
On May 30 2011 14:01 stenole wrote:
There is an extra cost of latency to a client-server architecture. So even ignoring economy, it may be undesirable to implement it. Not really, there would be less latency for some. More for others. Latency would change from between 2 players, to between each player and the server. You might have a better latency to the server than with other player.
But if the other guy is right. And the game is already client-server (which is very plausible). Then the additional delay would be the time the server's CPU take to calculate each player's vision. Which is a few 1/100000th of a second or something negligible like that.
Doesn't your in-game latency change if you're playing on a different server in SC2? (ex between US, euro or korean bnet). If that's true, it means that the game is actually already client-server. And blizzard is just not doing any filtering in the server.
|
I think R1ch would get a kick out of the armchair programmers in this thread.
|
Germany2896 Posts
Currently the commands given by the players are relayed through bnet servers. This results in very low, constant rate traffic and the server doesn't need to simulate the game. Now if you change to a model where the server only sends the necessary information, it would have several negative consequences:
* The server need to simulate the game. So Blizzard needs *many* more servers to handle the load. I'd guess something like a factor of a few hundred over the current amount of servers.
* The average traffic increases. Costs a bit of money, but blizz should be able to handle that.
* You can't record a small replay yourself and need to DL it from the server afterwards. No problem.
* In complex situations you need to send information about several thousand objects at a time. So you get sudden traffic spikes. Here Blizzard is lucky that SC2 is very primitive for an RTS. It's a bit hard to estimate how big such traffic spikes are. But Blizzard might need to raise the bandwidth requirements, especially for 4vs4s or complex UMS maps.
And to those guys comparing ICCup Antihack with bnet2 Antihack, that's no fair comparison. ICCup has no public hacks because the hackers don't bother writing them, not because their antihack is so good. The early implementations of ICC Antihack were trivial to circumvent(it basically sent an "Antihack is on and has found nothing" packet to the server which was trivial to spoof), and even the current implementation is less powerful than the technology built into Warden.
The only way to reliably prevent maphacks is using the architecture VIB is talking about, but it has disadvantages too. And it seems like Blizzard preferred maphacks of these disadvantages.
|
On May 30 2011 17:14 MasterOfChaos wrote:
* The server need to simulate the game. So Blizzard needs *many* more servers to handle the load. I'd guess something like a factor of a few hundred over the current amount of servers. Remember that the only expensive algorithm the server would need to do is the path finding. But fortunately path finding algorithms are very flexible. You can easily adjust it to be cheaper. (ie. pathfinding in sc1 is much cheaper than in sc2) Blizzard only made sc2's search algorithm as cpu intensive as it is today because they know it would be running on the client side with today's powerful cpus. So if you know you'll also be running that on the server you can just adjust it a bit to cost less cpu cycles. It would have a "negative" impact on gameplay, but not much -- some people still say they like sc1's pathfinding better  .
And of course. The ideal solution on a pro-gaming point of view. Is if Blizzard would only use this for select matches. And allowed us to use our own computers as host servers. This way online tournaments would be 100% maphack free and cost blizzard zero extra cash to run. They would only have to invest resources in developing this in the first place, but no extra infra-structure cost to maintain it. Then later they could slowly expand it to blizzard sponsored online tournaments. Then to only top ladder games (say top 1% ranked players or so). This would be a much more scalable solution.
So yea. There are economical problems. But on the technical side it's definitely solvable.
|
Looks like a proxy barracks / all in with SCVs. I don't see anything that might indicate map hacking.
|
|
|
|
|
|
EPT![[image loading]](http://i.imgur.com/RSxAe.png)
