Client/Server Model Idea

[Zachstar]

What is the thought on the client/server model idea?

[Dr.Sid]

Haven't thought much about this ..

For sure it will be more master-clients .. one of the game participants will e selected as master. I know no game which works different.

Now the question is what all will be client simulated, what master (server) simulated.
Movement and hits for sure needs to be simulated on every computer (so on master too) and it's easy to transfer to clients, or to compare the values with the clients.
Then there are sensors. Sonar or radar will produce much more data. Whole the simulation too must be made on all computers, so that is not a problem. But the transfer of results could be. The most cheat-safe way would be to transfer just the outputs to the client. There is a chance it won't be possible or easy.

Anyway naval simulation does not depend on low pings as other simulations does, so I guess it won't be much problem.

Even with no-save model-client model, most MP cheating consists of packet modification, or even sending completely fabricated packets (i've killed you !) , memory modifications (my HP is still 100%), and very often resource modifications (I have 100 grenades).

We have to somehow solve resource matching anyway because it is expected that different users will have different mods installed. Packet modification can be made pretty hard with encryption, even simple. Games vulnerable to this does not have packet encryption. Memory modification can't be completely avoided, but as this project will be quite live with many updates, it too will not be that easy, and will have to be done again and again.

Also with steep learning curve, the community always be small and will know each other.

Btw. is cheating problem in DW ? I don't play it in MP that much.

[Zachstar]

Well cheating will be a low problem obviously. However this sim so far talks to be a tad bit more "Sandbox" than DW meaning that it is possible that some jack*** will pop in and cause havoc for the fun of it and leave. Also keep in mind the price range.

Light anticheat abilities built in will deter most cheaters and jack***** without attracting cheaters. Heavy anticheat causes many cheaters to pop in just for bragging rights.

That is a side problem anyway. What I speak of is making the different aspects to be in plugins instead of plugins.

However, I can see how that can quickly cause problems and ruin standards.

[Phil21]

At first: I'm no expert about this, so maybe some of the following could be just...'wrong'...so sorry in advance :wink:

I just thought about what is needed for MP-Clients & Server here is what i think:

1. Generally you need to transfer the environment parameters (e.g. position/status of ships/subs, the 'map' with the terrain and changes in weather etc., other small pure AI objects like torpedos and their movement). This shouldn't be to hard since you only need to submit changes in their behavior plus checksums (to see if they are on course) etc.

2. I don't know how complex you modelled the sensors, but would not transfering the results be enough? Every result would be checked with the server if its valid (e.g. detection of a sub by FFG occures on several clientsplus server almost simultaniously) and then confirmed. This would cause the game to destabalize if there are differences in the sensors between clients, but i think it keeps traffic quite low...

3. Dont think to much about cheating right now. I dont think that cheating is a large issue in the DW community. Just concentrate on finding a reliable method which has no bugs and is stabel/fast rather than making it 100% cheat proof...

These are my thoughts about this. Afterall this is just an idea.
Phil

[OneShot]

On the cheating issue : I don't think this is or has been a problem with DW. To my knowledge there are no "Cheats" in DW. However what does exist is "cheating" or "gamey" gameplay by using gamey tactics or rigged maps where you know the start positions and such. Which does not exist (to my knowledge) are things comparable to wallhacks or glitching in FPS games (lets call them "real" cheats).

As far as those gamey cheats go, you can only avert them by :
1. not including functions that would allow them
2. carefully selecting the people you play with

I'd say No.2 is a lot easier to achieve then 1. as some people will get creative and use tactics or functions in the game in a total unexpected and gamey manner. Fortunatly, as Sid pointed out above, our community is quite small and such people are fast and easily identified.

Bottom line : as Phil said, don't worry bout making everything Cheat Proof so much, whats more interesting is a stable and fast connection.

Cheers
OS

[XabbaRus]

If you're looking for decent MP look at how IL-2 MP was implemented. Granted there was little sensor data being transferred but it works well even on bad connections.

[Zachstar]

IL-2 has to send ALOT more info. Gun fire rates, LOTS of Maneuvering, ground fire, etc...

Granted that sonar data will be difficult to transfer with good compression. However, I do not think anything more than 56k is truly needed with a good effort to compress the data.

There are multiple papers out there on ways to compress and manage multiplayer data. When it comes to that point I will begin a search for it.