ThinkTank - Persistant Servers

[VictorFarbau 0]

Hello,

I just wanted to stimulate some brainstorming about the upcoming "Persitant = 1" feature of the 1.05. In my understanding this will enable a dedicated server to continue running a mission even after all human players left. So theoretically any player can resume at any time to continue the mission and influence the environment.

I must say that I am very excited about this feature and I want to make sure to use its potential.

While I was thinking of making my first big battlefield mission (which is work-in-progress) compatible to that concept I already had some ideas what to look out for. I don't want to return to my private war at day 5 only to find out that everybody else ran out of ammo, has no gasoline or bleeded and starved to death.

Anybody interested in throwing in their concerns and ideas? I will maintain a list of items in this very first post then.

Vehicles

1. AI vehicles driving / flying around will run out of gasoline after a while. Need some smart scripting to either refuel them per script (lame), have them refuelled by supply trucks (better) or have them find and use gas stations (best).

2. Same as 1. but for Ammo!

Human players

1. If a continuous war scenario is required there need to be sufficient playable soldiers available. Think about respawning them to battlegroups at given positions in time.

2. Human players will need medics and ammo at a higher rate than in typical missions now. Think about how to make these available (EvacHelis, Ammo chaches etc).

AI players

1. Make sure they can get Ammo and medical treatment as well.

Battlefield organisation

1. Nowadays battlegroups often just appear somewhere on the map. When thinking of a more realistic approach this could include proper deployment strategies of enemy units using Heli insertion, moving out from barracks etc.

2. Resource / capacity planning. Nobody has endless supplies so attacking vital supply routes or warehouses for example should show an effect. Which would need a proper supply management to be in place in the first place. Same for both sides.

[paragraphic l 2]

With ideas like this there have to be bases don't you think?

A base for both sides where you can have teams/groups spawn as reinforcement, with active transport heli's who have to check their own status before deploying the teams; if low on fuel/ammo first get that before picking up the team

And if you are sure both side are almost unable to advance to far you can place fieldtents for medical situations and ammo

just my 2 cents love this idea

[Espectro (DayZ) 0]

Yea

Check this thread, to see the possibilities around data storage and data exhange in ArmA/Windows.

http://www.armedassault.eu/forum....=#post_

This means that you can connect to a database on your website, checking which ID the player connecting is, and then giving him the gear etc. accordingly using scripts.

So you can have a true persistent online world with RPG features.

On top of this you can have a map on your website always showing the current state of the battlefield.

[Heatseeker 0]

Thoughts..

Imagine a persistent battle using a customised C&H mode (with several hold areas), everyone leaves, 3 guys join and beat it. Goodbye persistent mission...

I believe that a large 'persistent' mission should have a minimum amount of players to be finished (end), for example you have 3 areas to capture and you must have atleast 2 or 3 players (squad) per area for the area to be "captured", in adition the areas should be held for a considerable amount of time.

No point about having a large persistent battle with few players playing and beating it.

[VictorFarbau 0]

Espectro, this was not my intention. To keep the server alive and equip soldiers won't be a big issue anymore once the server can run continuously.

My idea here was more to define the requirements to keep the environment alive and realistically reflect changes through time. Also a supply system is needed.

To have combat groups sitting in bases, run out from time to time to fight and then return home is neither fun nor realistic in my mind. I want infantry, tank groups and SpecOp groups to be out and about all over the place with a fair chance to be on their own for days, cut off from supplies, and be able to meet with other groups to form new battlegroups and get back into the game. So the real madness of war. Tecnically it should be possible. Just a lot of things to think about.

Cheers,

Victor

[VictorFarbau 0]

Heatseeker,

good point indeed. I admit that I only had "Coop" in mind. This way you could control the enemy side much easier if you set a scene in which they just control and defend the whole island for example.

I am not much into CTF and Sector control missions, here I want to explore a long-term war mission (well, long term, talking about days here).

Victor

[Heatseeker 0]

Well, a persistent coop doesnt seem so interesting imo .

Even if the mission is complex it might not be appealing enough, the usual kill enemy here and there against respawning a.i.

It might get too repetitive for a persistent mission, i think coop does best with diferent objective based events and goals (hostage rescue, convoy ambushes, etc) and no respawns, plus random events and replayability options.

Thats how i always liked coop seeing the end of a lenghty and challenging coop map with no respawns.

[[aps]gnat 28]

i.e. Dynamic Afghanistan by honchoblack or like Dynamic Libya

Love to see DA in persistent mode, with a few extra twists, like the enmy starting an base attack just after you've been sent on a distance attack mission yourself.

Or during several missions the enmy AI are building a stockpile or base of their own, hidden away, and not on the list of random go-get-them missions. Hence, if you are not patrolling during or before/after missions, you may completely miss the developing enmy force and thus never be able to stop them when they finally attack.

Of course you could just assign more and more resources to base defence (man, machines and static), just to hold off such an attack ...

[VictorFarbau 0]

Hm, most of the discussion I see is what type of mission everybody would prefer to run on a continuously running server.

The original idea was to think about requirements for mission makers when thinking about long-term missions which is a new thing for most of us. So feel free to add any ideas or concerns that come to your mind. I know this is more of a theoretical excercise; just thought it could be useful.

Regards,

Victor

[UNN 0]

The first thing that springs to mind, is saving data. As Espectro already mentioned. If the server crashes eight hours into a mission, then you want some way of returning to a reasonable point. Not to mention the fact that it might be a good idea to reset the server after so long, anyway.

So save and restore must be at the top of the list?

[crashdome 2]

I think you could very easily create very simple to very complex systems. I don't see the need to discuss this as if there is one grand idea which will appeal to all.

For example, I can see a system which merely spawns units as needed and deletes them when they die. this type of system would be good for very fast public servers. Like a continuous battle...

Also, I could see a simple multi-base system which spawns units as they arrive and deletes when they die... again very hip to those who like fast public servers.

On the more complex end, I can think of a system which handles missions based on predefined context but with dynamic objectives. For example, a typical sweep mission is a series of waypoints (three are target waypoints) and the mission AI generates them in areas where intelligence says it saw the enemy. It would account for unit, fuel, etc and schedule the mission. That way players can jump into any mission for any time or even into existing units on mission.

If units die, run out of fuel, etc.. than it is the result of action. They are deleted when dead, etc.

It would be very interesting to see a recon mission from a littlebird scheduled and when the heli is shot down, the mission AI schedules a SAR recovery team and schedules some Blackhawks to transport the team. If the unit dies, the mission is cancelled. If the team makes it back to base, they are deleted. If the SAR team dies on the way, schedule a new one.

Make two mission AIs and you got two sides working against each other. Make a dialog for players to schedule their own missions if needed (if they are high enough rank maybe?) and you have a great way to defeat redundancy.

Because each side is fighting either AI or Players it satisfies everyone's needs in terms of coop or PvP. You could make a small version with players only for the hardcore PvP.

The problem with this complex of a system is you need ALOT of maintenance. Saving data is necessary. Also, you'd need to write ALOT of FSMs to cover the AI being able to respond to the mission AI.

We can talk all day about how to make a persistant server, but the more complex we make it - the more likely it will never happen.

[Hovis 0]

I think a co-op persistent server could be good, and despite being an ongoing concern it need not involve huge numbers of players. A game based on fighting a guerilla war could be good, with players able to spawn at a base camp, tool up a squad and then make moves on the established enemy occupation forces as and when they see fit.

I think however something that could really be useful is some sort of implementation of the Dungeon Master concept from the Neverwinter Nights games. For those unfamiliar with the idea basically the Dungeon Master is a player who can spawn monsters, trigger effects, take control of NPCs and basically add the human factor to the module that is created.

Giving the AI army an overall commander, with the ability to spawn units and whatever (note though it'd have to be somebody more interested in seeing the opposing force have fun than just wiping them out) would make a persistent mission a lot more fun. It'd mean you could basically set up the islands military to function passively, but if you want to whip up a company of men to put the hurt on a player held base you can.

At the risk of sounding too RPG oriented I'd also like to see a rank system and an experience points system cooked up in persistent games, maybe use them to determine who has access to what items. This kind of character progression gives people an incentive to play on, but I don't think there is any capacity to store character data in Arma, so that might be a nonstarter.

[Nutty_101 0]

Thats called VBS2/1.

[crashdome 2]

First check out CE3... before you say that

http://www.thechainofcommand.net/cgi-bin....2;top

[VictorFarbau 0]

Actually, now that patch 1.05 is out and my dedi server has been updated to that: the "persistent = 1" parameter seems to have no effect. After I leave the server still closes the mission.

Is there anything known about the parameter or maybe requirements for the server to run in this mode? As usual documentation is not available as it seems.

Cheers,

Joerg

[Nutty_101 0]

I know this topic might be a little old but just some new ideas to toss at people.

One thing i have done with my servers is the fact that almost all the missions and scripts that cause events are stored in a database. There is one main script that checks every five minutes to determine if there are any new scripts to launch. If so they are added executed on the server. So pretty much you can add scripts to the server any time or schedule them when to go off. I use it so that i can update my scripts without restarting the server at all (Unless i update a client side dialog).

[=dB=O8.Ed 0]

I would like confirmation that using software that would hook into the games executable to allow database integration is allowed. (Don't need any trouble) if so I would love to use a database backend with this game!

[Nutty_101 0]

I would like confirmation that using software that would hook into the games executable to allow database integration is allowed. (Don't need any trouble) if so I would love to use a database backend with this game!

I really have no idea if this is allowed or not. What is being done is not used to support cheats or hacks or break us into the VBS market (It does not have any database connections that I know of anyway). That I can promise anyone, I want this company to profit so we can continue this genre of the game. I get bored to death with the limited abilities of other games and most games in this area are to linear to me. Bis has taken a huge step in a direction I have always wanted to see and they offer us power to do things thru the tools that other companies do not even look at and more than likely will not.

The software was designed to extend the dedicated server (Though the client app works as well) to allow external data storage. In my case I use it for the ability to modify scripts without having to shut my game down. So if there is a bug that needs resolving or a new addition I just set a flag and all selected scripts are terminated, updated, and restarted while people are playing. I also use it as a mission manager since I use external software to monitor the games world and create situations at random based on information, conditions and time.

As for the constant running of a server game (persist). Why do we need AI that just re-spawns anywhere it wants or in a set location all the time? That's boring. Why not have something that lets people play whatever way they want while utilizing the AI for functions that most online players would bore of. I am trying to offer the ability to do whatever your team or you enjoy doing. That might be capturing a base/town, providing transport to the front lines, being the guy who drives the M113 with units in it for re-spawning players, or even be something so transparent as transporting fresh weapons and ammo to the front bases or combat lines. In my world any base or town on the map can be captured and converted. They all can hold units (limited to housing) along with vehicles (Again limited to space). So you can move units from town to town (Only way to get people to the front lines anyway. No units no support/backup. No spawning). The bases also are self healing, so during attacks they go into defense mode but if they survive the attack then the engineers will run around and fix all the damage, repair any vehicles that are able to be; other wise they are buried and gone for good. Expired units are picked up by medics and take to a hospital where they are cooked into mush for the troops. All weapons and ammo are also collected and returned to the base as are limited as well; no they don't appear out of nowhere. So as you knock off the enemy you can recover their goods into the base/towns stocks. The goods can also be delivered to different towns/bases by land or air (though not in large quantities).

It's starting to be fun to watch the whole thing play out of late as I have all the towns alive. Last night by leaving the game running I came back this morning and see a few towns switched controlling forces. This is mostly because the AI manager will launch attacks based upon supplies at the forward bases. I also have some minor implementations of civilian revolts working, if they can sneak into a guarded area with supplies or have the chance to grab stuff from the dead they will mount small attacks that can cause issues.

An example from Wednesday nights logs.

A group of four civilians ambushed an AI reinforcement convoy of 22 (West) units traveling from Geraldo to Corazol in two 5ton opens. Not sure where the civies got the supplies from (I am guessing around Corazol as that is where the majority of the battle has been since it is the dividing point between the two sides). They managed to knock out the two vehicles without any incident with the use of 3 satchel charges and 1 round of an RPG.

This is all dynamic thru the scripting and integration of a database and some extended functions. Where as without this connection, doing something like this would be impossible. If we are not allowed I cannot see a real reason why not. The game is marketed to a group of people who bought it because it offers a more realistic environment than any other game on the market. The game has vast landscapes that at the moment only small portions of are even being used in missions. If we are expected to make any real use of the vast world being offered then we need something else to store information. With the reliability of the dedicated server being as it is, we cannot expect to run a persistent game that would offer any of this. I know what I am doing would be impossible as I use two servers and in the future I want to expand out to maybe four that all communicate with each other.

Sorry for the long winded response but that’s why I need this link and use it.

[jerryhopper 285]

geez that was a long read.

i support the idea nutty has, and what he says is true.

if you want to use all of that BIG sahrani island, the current 'Persistent' function just dont suffice.

[UNN 0]

Quote[/b] ]if you want to use all of that BIG sahrani island

It's actualy quite small when your flying over it in a jet

But thanks to the work Nutty is doing, it should be possible to extend the length of mission indefinetly, without running the server into the ground and loosing all your data. With a mamoth, none stop, gaming session

[Nutty_101 0]

Quote[/b] ]if you want to use all of that BIG sahrani island

It's actualy quite small when your flying over it in a jet

Your telling me, I have been trying to figure out a way to do some nice AA scripts for bases but to be honest by the time the planes leave the ground you would have to shoot at them. I am sure we will get our huge maps sometime. Then it will be fun.

[UNN 0]

Quote[/b] ]Your telling me, I have been trying to figure out a way to do some nice AA scripts for bases but to be honest by the time the planes leave the ground you would have to shoot at them.

Even before we got proper multiple airbases, I did a quick experiment to see how much fuel you would use flying from one small island airbase to the other and back again. It used about 1%. Kind of removes any sense of urgency. Unless you reduced the fuel capacity to 60 seconds of flying time

[Nutty_101 0]

heh, I am just waiting for the tools to get released. Curious if we can build a huge ocean with some decent sized islands to work on. Then we could get the naval, air and other sections running with a real purpose.

[UNN 0]

With your system, how practicle would it be to store details on every active unit in a long running mission?

Also what would be the basic hardware and software configuration. Could it be done with a single, standard server, outputting to a text file. Or do you need multiple servers and a database engine?

[Nutty_101 0]

With your system, how practicle would it be to store details on every active unit in a long running mission?

Also what would be the basic hardware and software configuration. Could it be done with a single, standard server, outputting to a text file. Or do you need multiple servers and a database engine?

Ehh, i dunno. If you are careful you could keep track pretty well. If you want to talk about it more, post up on my forums. You don't need to register to put up anything in the support area. I don't want to flood this topic anymore than i have. Ill have a response to this msg there.