Mr Ekted Posted August 17, 2003 Report Posted August 17, 2003 We've been all through this before... The ONLY security is obscurity. If you do not understand this, then you are not knowledgeable enough to even be discussing the issue. Cont will never be open-source for security reasons. Cont was not made using any existing source code. JeffP (et al) would never risk their careers.
etrigan Posted August 17, 2003 Report Posted August 17, 2003 We've been all through this before... The ONLY security is obscurity. If you do not understand this, then you are not knowledgeable enough to even be discussing the issue. Cont will never be open-source for security reasons. lol im guessing you're not in the security industry. I am. And that statement would be considered false by many. But you are en!@#$%^&*led to your own opinion.
Guest Liddo-Kun Posted August 17, 2003 Report Posted August 17, 2003 I wonder what PriitK is working on now ^_^
Smong Posted August 17, 2003 Report Posted August 17, 2003 Liddo-Kun: By now it's pretty clear that nothing will ever be added to this game anymore if it requries over 1 line of typing for PriitK.I don't think they even add 1 line, the file size of the pre-releases are always the same (but that could be something to do with security/checksums).
Yupa Posted August 17, 2003 Report Posted August 17, 2003 I wonder what PriitK is working on now ^_^ probably .39pr1 bugs
»SOS Posted August 17, 2003 Report Posted August 17, 2003 The file size of programs is often the same between minor changes. The compiler sometimes leaves in "empty space" for optimization (or bad compiler) reasons.
Mr Ekted Posted August 18, 2003 Report Posted August 18, 2003 etrigan, please post your ideas to prove me wrong, but be aware you will be making a fool of yourself. If you are in the security industry, then you can explain in excruciating detail how to protect a client from the user of that client through encryption (or whatever your ideas are). Please, we are all on the edge of our seats.
Coconut emulator Posted August 18, 2003 Report Posted August 18, 2003 I think etrigan can't prove what he's saying. The ONLY security is obscurity. If you do not understand this, then you are not knowledgeable enough to even be discussing the issue. Cont will never be open-source for security reasons. You're absolutely right, Ekted. Continuum will never be "open-source" due to security reasons... OK. The real problem is that the code at execution time is always "open-exe", so the code can always be cracked/reversed and that obscurity that you mentioned disappear. You can always get an asm source of any x86 machine code. Even if the code is encrypted or whatever security techniques are applied you can still use a system level debugger like SoftICE to trap the code as it is executed. In the case of Continuum, SoftICE protection (just running or breakpointing) must be cracked before by using a diss!@#$%^&*embler and a hex editor/cracking program. We can discuss this issue for years and years, while all new modifications and security features got cracked systematically. I think the security should reside in the logic of the game protocol. The game should be played at server side mostly. The game is played a lot at client. All the client side features are more susceptible of cheating.I will use greens as an example. Greens are placed by the client on screen, so the server has no control at all of greens earned by players. Using the actual philosophy the game server will never be able to detect client cheats based on greens. If the client says it has gained an item by greening, the server has no way to validate it. Note that some greens are distributed by server to other clients (killing greens) but the green that your own client places randomly on screen when other green has expired is not distributed the same way. As a side result of this, you can have greens on your map that other players does not have on their maps. Also, the possibility of green based cheats is open. The server must accept every item you use because it has no mean to validate how you earned it. Why cheat with greens having the ?buy feature? Most people will use ?buy instead of cheating, but greening a brick, etc on the run (out of safe) could save your life many times. Setting the server to show ship state to spectators could help. But many zones won't do it because everybody will see smods/sysops cheating (abusing the *prize command). It is not the matter, but I also want to say that this game is designed for staff to abuse powers (not very nice). My point is on how the logic of the game protocol is open to client cheats. On the other hand, play the whole game at server side will introduce lag in some routines that are triggered locally at the client by now. This way, the client performance will be affected by lag. More game traffic will be introduced. It is just an idea that I present here: client performance against game security.I know it is crazy to pretend to play the whole game at server side, but some things must be done this way or there will always be security concerns with some aspects of the game. You can forget about people developing bots... this will always happen (even with a double key encryption system). Bots can be a lot of more stronger than the client could be by taking advantage of server features and bugs/holes and also by automating commands. Once again, playing the game at server side will reduce the power of clients/bots. Coconut emulator
Yupa Posted August 18, 2003 Report Posted August 18, 2003 good read nothing is 100% secure, I think Continuum does alright, but as always I will leave the heavy programming thought to people that actually know a programming language
Mr Ekted Posted August 18, 2003 Report Posted August 18, 2003 Granted CoCo, but Subspace will always be client-authoritative, and Jeffp's/PriitK's solutions are valid and necessary given that. I cracked all of subspace.exe in about 3-4 days. Although I've never tried or looked, I would guess it would take me more than a month to do the same with continuum.exe. I'm not willing to put in the time, but if I did and was able to execute any obvious cheat, it would take a day to change everything. How many people are skilled/malicious/persistent enough to spend a month hacking after each modification comes out a day later?
Yupa Posted August 18, 2003 Report Posted August 18, 2003 well I would be persistent enough, but first I'd have to learn the language, and I'm not THAT bored
etrigan Posted August 18, 2003 Report Posted August 18, 2003 I think etrigan can't prove what he's saying.dont assume. thanks. I think the security should reside in the logic of the game protocol. The game should be played at server side mostly. The game is played a lot at client. All the client side features are more susceptible of cheating. and there's the fix. The way SS is client-side is an unfortunate and inadequate security model. I think the VIE guys originally created what they thought was best, and unfornately it didnt have a strong security model. Server side stress and network utilization are a problem of course, but im thinking SS/Cont needs a change from the ground up in that area. The way its done right now doesnt seem to give us room for the future. The game protocol was made before internet security was an issue for many. please post your ideas to prove me wrong, but be aware you will be making a fool of yourself. If you are in the security industry, then you can explain in excruciating detail how to protect a client from the user of that client through encryption (or whatever your ideas are). Please, we are all on the edge of our seats. Ahhh thanks for keeping an open mind. I think this subject is moot at this point, if indeed there's no changing of the protocol. The protocol is the problem, perhaps you could see that instead of attacking people.
etrigan Posted August 18, 2003 Report Posted August 18, 2003 Granted CoCo, but Subspace will always be client-authoritative, and Jeffp's/PriitK's solutions are valid and necessary given that. I cracked all of subspace.exe in about 3-4 days. Although I've never tried or looked, I would guess it would take me more than a month to do the same with continuum.exe. I'm not willing to put in the time, but if I did and was able to execute any obvious cheat, it would take a day to change everything. How many people are skilled/malicious/persistent enough to spend a month hacking after each modification comes out a day later? eh figured i'd better address this also... Im not entirely sure of the detection mechanisms(other than the obvious), that are used by sysops(im just a player ;P ). But as far as I can tell they are inadequate. And detection is a problem. If relying on people to report cheaters, is one of the main methods in some of the zones, then it sounds like a fricking problem to me. Just my opinion. Well over 1000 players at peak times, im sure there are some that have the ability. -*BAD WORD*- I wouldnt be surprised if there were MITM attacks being messed with. Mr Ekted, do the servers have logging? and if so what type?(this is a serious question, try not to talk down to me when you answer, sorta makes you look like a defensive a-hole) Thanks!
numpf Posted August 18, 2003 Report Posted August 18, 2003 given the performance/timing requirements, it is necessary for a significant portion of the game state to be managed by the client. Therefore the client must be protected so the server can trust it. The only means of protecting the client is obfuscation. -numpf
»nintendo64 Posted August 18, 2003 Report Posted August 18, 2003 etrigan, SubSpace has a lot of security features(Packet Encryption, Security Checksum, Position Packet Checksum, SoftIce detection...) to annoy any cracker to stop them from disambling it, and continuum has even more. Some Security features of Continuum is that doesn't allow Full access to its proccess and if you make it, it completely disables the Winsock support, plus it won't run if softice is running. Yes those are minor security features, but for a game that dates from 1996, they were enough for their time(until VIE dropped support and SS got cracked) and now they're been more than enough with some new features added by PriitK. The best security it's what Ekted said, sure you can dissamble Continuum but as soon as it's done a new version will be there. For the Cheater Detection, the Subgame (Game Server) does a few checksums over the data sent by the client, so yes it can detect certain type of packet tampering or memory editing, it detects for fast bombing, having items that you're not allowed, or high amounts and etc.., yes some stuff like the Greens location on the Client cannot be detected easily if manipulated, but with the new Server in development ASSS, maybe that can be fixed soon. -nintendo64
etrigan Posted August 18, 2003 Report Posted August 18, 2003 etrigan, SubSpace has a lot of security features(Packet Encryption, Security Checksum, Position Packet Checksum, SoftIce detection...) to annoy any cracker to stop them from disambling it, and continuum has even more. Some Security features of Continuum is that doesn't allow Full access to its proccess and if you make it, it completely disables the Winsock support, plus it won't run if softice is running. Yes those are minor security features, but for a game that dates from 1996, they were enough for their time(until VIE dropped support and SS got cracked) and now they're been more than enough with some new features added by PriitK. The best security it's what Ekted said, sure you can dissamble Continuum but as soon as it's done a new version will be there. For the Cheater Detection, the Subgame (Game Server) does a few checksums over the data sent by the client, so yes it can detect certain type of packet tampering or memory editing, it detects for fast bombing, having items that you're not allowed, or high amounts and etc.., yes some stuff like the Greens location on the Client cannot be detected easily if manipulated, but with the new Server in development ASSS, maybe that can be fixed soon. -nintendo64 thank you for the educational post w/o bashing(some people here are incapable of that). Is the bullet damage still done clientside?(ie. shrap, burst since they are under the same hood)
Dire Wolf Posted August 19, 2003 Report Posted August 19, 2003 I agreez0rz... grabvity b0mbz twould roxz0rz j00 s0xz0rz! -DW
Coconut emulator Posted August 19, 2003 Report Posted August 19, 2003 Subspace will always be client-authoritative, and Jeffp's/PriitK's solutions are valid and necessary given that.Yes... I'm talking about other game... I understand that my proposals are about rewriting the whole game, in a way that will be nearly impossible to write a game server to accept the actual client and "my" client at same time... So I'm talking of other game I guess. And yes, I have to agree, Jeffp's/PriitK's solutions are valid and necessary. The game is getting better, in the technical side I mean. We all hope Priit will hear you "one of these days" asking for enhanced zone settings and such. That kind of discussion is much more valuable than talking about rewriting the whole thing. But, oh -*BAD WORD*- it comes the "but", I think that passing the green control from client to server won't surprise anyone. Does people usually complain about how laggy is to get the soccerball sometimes? That's because the server must assign the ball to you and tell it to everyone so all can see ball carriers on their maps (if the corresponding setting is on). By controlling the greens at server side, the most general kind of cheaters (but not all, sigh, I know) will be stopped. Fortunately, twister times passed by. It doesen't seems to be any cheating progs around... huh? Well... there always will be someone screwing here and there, but all of we know how sad Twister times were... tones of people jumping, bricking, thoring like mads. Oh well, there's a few I can do or say to help the game in its actual state. I don't mean to harm it in any way too so I better shut up. Another "last" thing: I'm not pretending to teach you how the game is, Ekted, neither to anyone else, I'm just trying for most people to understand what are we talking about... I think this subject is moot at this point, if indeed there's no changing of the protocol. So true.But oh well, this topic is about gravity bombs... vote yes!! vote yesssssssss!!!!!!! Coconut emulator
etrigan Posted August 19, 2003 Report Posted August 19, 2003 But oh well, this topic is about gravity bombs... vote yes!! vote yesssssssss!!!!!!! I voted yes!! Always yes for new features.
Coconut emulator Posted August 19, 2003 Report Posted August 19, 2003 How many people are skilled/malicious/persistent enough to spend a month hacking after each modification comes out a day later? What it took me two months to crack, two years ago, it takes 1 week to me now. I try to be in the skilled/persistent side, not in the malicious one. But, anyhow, I have to agree again. Boredom, frustration, paranoia, are the best weapons against crackers/hackers/whatevers. I agree to what nintendo64 said too.
»CypherJF Posted August 19, 2003 Author Report Posted August 19, 2003 You guys sure have a lot of good arguments and comments on several matters.... but I still think gravity/repel bombs would rule >: ) harhar!- CypherJF
Mr Ekted Posted August 19, 2003 Report Posted August 19, 2003 Not trying to be mean, although it's hard not to. It's so easy for people to come in here and say "open source is the only way" or "obscurity is not security" who have never written a single line of code, or written a game, or have the slightest clue. I would expect someone with a serious idea to actually propose it in detail. I have spent the last several years, on and off, considering ways to protect Subspace better. I have seen nothing yet other than the typical technological rhetoric. The following are NOT serious/valid ideas... - re-do entire protocol for server-authoritative- use RSA/PGP/3DES/blah encryption- open-source the client John Carmack (Quake) himself has given in to the fact that closed-source obscurity is the ONLY solution to "good enough" security in fast-action games. Btw, gravity bombs are kinda cool.
»nintendo64 Posted August 19, 2003 Report Posted August 19, 2003 Is the bullet damage still done clientside?(ie. shrap, burst since they are under the same hood) Yes, it's done Client-side. -nintendo64
Coconut emulator Posted August 19, 2003 Report Posted August 19, 2003 Not trying to be mean, although it's hard not to. It's so easy for people to come in here and say "open source is the only way" or "obscurity is not security" who have never written a single line of code, or written a game, or have the slightest clue. I would expect someone with a serious idea to actually propose it in detail. I have spent the last several years, on and off, considering ways to protect Subspace better. I have seen nothing yet other than the typical technological rhetoric. The following are NOT serious/valid ideas... - re-do entire protocol for server-authoritative- use RSA/PGP/3DES/blah encryption- open-source the client John Carmack (Quake) himself has given in to the fact that closed-source obscurity is the ONLY solution to "good enough" security in fast-action games. Btw, gravity bombs are kinda cool. OK, here I stop agreeing. And I want to be mean too. A couple of things.... You think my ideas are not serious, they are the typical technological rhetoric... OK. First: I'm not talking about rewrite the whole protocol to be server authoritative. I'm talking about redoing parts of it which will always be a problem in its actual state. I can discuss it in detail if you want. PGP encryption and open-source ideas sucks, I still agree on that. Second: I haven't heard any argument, good or bad, besides the tipical "oh no no no, this just can't be done" thing. I would like to hear a good reason which makes SO impossible to let the server take care of a few things that are client side at the moment. No wonder that you've been guessing for years how to enhance game security. I don't see any significative advance on that. You are just lucky because there seems to be no people like sage386 around the game at the moment. But your "work" on the SS/Cont protocol deservers a plague of sage386s flying around. Talking about Continuum protocol as a better thing than the SS protocol is just an illusion. I say closed source is the best way it can be done. I say obscurity is your only weapon (a poor weapon, but it is almost the only one). And yes, I wrote some more than a single line of code. One of the last personal projects I've been working on is a 3D engine based on DirectX (COM objects fully written in !@#$%^&*embly) which can be called from a lot of languages like C++ and Delphi. The core is finished at the time and it is fast as -*BAD WORD*- and also very customizable. I have built some pretty demos with it. That's why I arrived here to discuss the SS protocol. I'm planning a 3D multiplayer game with it and some more things. I'm not going to start the multiplayer game until I design a better protocol than the SS/Cont one. I know well some security issues can't be solved that's why I'm not currently developing my own internet game. I feel like spending some more time analyzing SS protocol, to learn how things must not be done. And it's not so easy to come here and say things like I said. Come here is easy, yeah, but the things I said aren't easy. It took me months to crack this -*BAD WORD*-. I know very well what I'm talking about. If you think I'm fake and bluff you're plain wrong. So John Carmack says that closed-source obscurity is the ONLY solution to "good enough" security in fast-action games... I admit it is the ONLY solution but I seriously doubt it is good enough. -*BAD WORD*- John Carmack is all I can say. Also (and now I'm really mean), from a cracker to a cracker, I seriously doubt you cracked the whole game (a lot of more than encryptions and protocols) in 2 months and didn't feel the urge to rewrite parts of it that remains the same by now, Ekted. given the performance/timing requirements, it is necessary for a significant portion of the game state to be managed by the client. Therefore the client must be protected so the server can trust it. The only means of protecting the client is obfuscation. I see you guys will be trapped here for years... good luck with it. The client just can't be protected from its user. Period. Have a nice journey. Coco
Recommended Posts