My assumption is that the typical cycle still holds true? Whenever they add some new feature, it becomes super buggy, they squash bugs for a while and it becomes relatively stable, only for new content to be added and it becomes buggy again.
Lancet in 3.18 and honest ever since the mining rework has been a problem because it's supposed to be a 'support' mining laser with a weak output, but its bonuses to support were so good that its lack of power didn't matter. It had a -75% resistance and -75% instability buff on any rock it was touching, and with some mods (since it could equip 3) you could get that up to, like, -95% resistance and -77% instability with 2x FocusIII modules (-10% resist and +4% instab each) and 1xVaux C3 module (-10% instab, -4% optimal charge window). Or could go up to -100% if willing to eat the instability hit.
Turns out even a weak pissbaby laser can mine quant ez when it can lower the rock's resistance by that much, and it kind of became the go-to for most everything.
The Lancet in 3.19 is, uh... Not that. For some reason it's still referred to as a support laser, but tbh it's more statted like a budget Helix. It has the second highest laser output in terms of mins and maxes, a small reduction to instability, no reduction to resistances, middle of the road mod slotting, and crushes your optimal charge window by 60%. Its big bonus is ... 40% optimal charge rate. Which is... not good. Being able to get the rock to the green zone and reduce the risk of overcharging it is just so much more valuable than getting the rock to charge a few seconds faster.
I think I remember hearing SaltEMike say that all mining lasers were being rebalanced in such a way that each laser now has a specialty for what it can mine.
The idea being that you are supposed to have a target resource in mind beforehand, and you match the laser to the type of resource you’re mining.
To go along with that laser rebalance, Mike also said they did a big resource placement shuffle. Certain resources are now only found at certain locations.
So… if you know you want to mine X, you equip the laser best suited to X and then you fly to the planet/moon/asteroid where X is known to spawn.
Has there been any word on the sort of plants we will be able to grow on the land plots we purchased a way back? Idk if it's planned for the next system or its a sq42 thing but i love cross breeding plants ala animal crossing
NEO|PhyteThey follow the stars, bound together.Strands in a braid till the end.Registered Userregular
There are possible features that aren't already must-haves?
It was that somehow, from within the derelict-horror, they had learned a way to see inside an ugly, broken thing... And take away its pain.
Warframe/Steam: NFyt
There are possible features that aren't already must-haves?
But of course.
Robot companion and therefore pets and therefore the petting, feeding, housing, and persistence thereof.
(TW for AI, Creature, Item, Vehicle, and Story teams)
I think pets with all the things you mention was one of the early stretch goals ~10 years ago.
He'll probably see the build-your-own-spaceship interface and want that, requiring a full rework of all spaceships
While I don’t think sentient AI are part of the setting, I’m pretty sure an astromech type repair droid was a stretch goal or bonus for buying some ship or other.
There are possible features that aren't already must-haves?
But of course.
Robot companion and therefore pets and therefore the petting, feeding, housing, and persistence thereof.
(TW for AI, Creature, Item, Vehicle, and Story teams)
When your pet craps on a moon, you can leave and come back weeks later and it will still be there.
Looking forward to the tier 2 implementation where the poop animations are implemented, but the poop itself is still represented as a pile of 1/64 SCU Fertilizer crates.
It's CitizenCon right now and they're showing all sorts of shit. Apparently this game that is never going to come out might perhaps come out (but maybe not). Maybe though. We'll see.
Yeah, there was actually a lot of pretty amazing stuff at CitizenCon this year. I'm sure people will still find a way to be haters, non-believers, skeptics, cynics, and scam-claimers. But this year's con certainly made it a lot harder to be any of those things.
Big items from the con include:
-A live, onstage demonstration of Server Meshing + Replication Layer at work. The tech is real. It exists. They did it. They got it working. Yes, it took forever. But they have delivered a new form of server tech/architecture that has never existed in video games before now.
-An announcement that Pyro (and probably Server Meshing by necessity) will be going up on a PTU server on Oct 31. This is huge. Also, the Pyro system was playable at the show on demo stations.
-The announcement that Squadron 42 is feature complete and has moved into the final stages of production, which is quality pass, optimization, iteration, balance, and finishing touches. No release date given, but at this point the release is an inevitability. Maybe 2024. Definitely no later than 2025.
-Engineering gameplay is coming along quite nicely. They showed an in-game demonstration set on board a Crusader Hercules starship. They walked us through routing power, repairing and replacing components, turning components on and off via the engineering control station, opening and closing doors remotely, and also fire propagation when all hell breaks loose on the ship.
I can certainly criticise them for taking 10 years to make a game any half decent studio would have done in only a couple with a faction of the money they raised. and i will.
It’s honestly kind of hard to believe that Squadron 42 might actually be released in the next year or two or three. I had written it off as a casualty of Roberts’ crusade for perfection. Kinda weird to have positive feelings about a project I’ve been mostly ignoring for years. The hype for this game was stupid high like almost a decade ago. Fast forward to today and I’ve built a family and a career while waiting on this!
I can certainly criticise them for taking 10 years to make a game any half decent studio would have done in only a couple with a faction of the money they raised. and i will.
I mean, sure. Go right ahead. You're free to voice your opinion.
But doing so makes you look like a bit of a goose. It demonstrates that you have no comprehension of what an incredible feat of network engineering the effort has been to create server meshing. It took 11 years because this is something that has never been done before. This is literally brand new technology that they have had to invent. And yes, they have had some false starts and they had to entirely scrap an earlier implementation of it, because it wasn't working. But that's not because they're a bad studio. It's because they are doing something that has never been done before. This isn't a matter of something "a half decent studio" should have done in a couple of years. They have some of the top network engineers in the world working on this, and they are doing something in gaming that has never been done before. (Maybe outside of gaming too).
Other video games use technology such as sharding, phasing, and other tech to instance people and keep servers from exploding from having too many people in a space at one time. That's not new. But what *is* new is the way they're allowing people who are on different servers to be able to interact and see each other across those servers.
You know how games like WoW have zone lines and load screens? SC will have none of that. You know how in WoW, if a player walks from the world map into a dungeon, it loads them into an instance server and removes them from the world map? That won't happen in SC. In SC, if one person was standing on the world map and another person was standing inside a dungeon instance, they would be able to look at each other across the zone line, and they would be able to shoot each other. From different servers. That tech does not exist today. Servers don't allow interaction across them like this. That's the new part.
Having multiple servers, instances, sharding, and phasing is old tech that exists today. Allowing people to interact and see each other across different servers has never been done before.
You're free to criticize them. But being critical indicates that you don't have an understanding or an appreciation for what they have managed to accomplish. They aren't just delivering a video game here. They are delivering the future of server technology.
+1
Options
NEO|PhyteThey follow the stars, bound together.Strands in a braid till the end.Registered Userregular
They might be referring to Squadron 42, which notably is single player and thus hasn't needed any of the networking architecture that has held up Star Citizen
It was that somehow, from within the derelict-horror, they had learned a way to see inside an ugly, broken thing... And take away its pain.
Warframe/Steam: NFyt
In most projects i've been on feature complete is the first 80% and the final cleanup is the other 80%.
I backed in the initial campaign and haven't been paying attention in the last couple years. More on foot than I expected, but what was shown was pretty awesome looking to me. I hope actual gameplay is much heavier on the flight aspect than the shooter side.
Yeah, there was actually a lot of pretty amazing stuff at CitizenCon this year. I'm sure people will still find a way to be haters, non-believers, skeptics, cynics, and scam-claimers. But this year's con certainly made it a lot harder to be any of those things.
Big items from the con include:
-A live, onstage demonstration of Server Meshing + Replication Layer at work. The tech is real. It exists. They did it. They got it working. Yes, it took forever. But they have delivered a new form of server tech/architecture that has never existed in video games before now.
-An announcement that Pyro (and probably Server Meshing by necessity) will be going up on a PTU server on Oct 31. This is huge. Also, the Pyro system was playable at the show on demo stations.
-The announcement that Squadron 42 is feature complete and has moved into the final stages of production, which is quality pass, optimization, iteration, balance, and finishing touches. No release date given, but at this point the release is an inevitability. Maybe 2024. Definitely no later than 2025.
-Engineering gameplay is coming along quite nicely. They showed an in-game demonstration set on board a Crusader Hercules starship. They walked us through routing power, repairing and replacing components, turning components on and off via the engineering control station, opening and closing doors remotely, and also fire propagation when all hell breaks loose on the ship.
Server meshing won't be in the Pyro preview channel. It won't have jumping between Stanton & Pyro, so it won't need server meshing. Chris Roberts said it was only shown working 3 weeks ago. Making it similar to when he made the PES letter to the chairman. Which we then waited another year for that to release, and that was in a terrible state.
_____________________________________
-'Squadron 42' trailer impressed me. I didn't really care about 42 before this, now I'm kinda interested. Trailer exceeded my expectations.
-Base building was just more theory crafting. Same old, same old for 'Star Citizen'. Won't even begin development until 2024.
-Freight Elevators When!?!
-Distribution centers & raids seem cool. They looked way too small to necessitate roads or encourage vehicle usage though.
-Very happy to see free roaming quantum is returning.
Yeah, there was actually a lot of pretty amazing stuff at CitizenCon this year. I'm sure people will still find a way to be haters, non-believers, skeptics, cynics, and scam-claimers. But this year's con certainly made it a lot harder to be any of those things.
Big items from the con include:
-A live, onstage demonstration of Server Meshing + Replication Layer at work. The tech is real. It exists. They did it. They got it working. Yes, it took forever. But they have delivered a new form of server tech/architecture that has never existed in video games before now.
-An announcement that Pyro (and probably Server Meshing by necessity) will be going up on a PTU server on Oct 31. This is huge. Also, the Pyro system was playable at the show on demo stations.
-The announcement that Squadron 42 is feature complete and has moved into the final stages of production, which is quality pass, optimization, iteration, balance, and finishing touches. No release date given, but at this point the release is an inevitability. Maybe 2024. Definitely no later than 2025.
-Engineering gameplay is coming along quite nicely. They showed an in-game demonstration set on board a Crusader Hercules starship. They walked us through routing power, repairing and replacing components, turning components on and off via the engineering control station, opening and closing doors remotely, and also fire propagation when all hell breaks loose on the ship.
Server meshing won't be in the Pyro preview channel. It won't have jumping between Stanton & Pyro, so it won't need server meshing. Chris Roberts said it was only shown working 3 weeks ago. Making it similar to when he made the PES letter to the chairman. Which we then waited another year for that to release, and that was in a terrible state.
Ahh, ok. I missed that comment. I didn't realize that they were not going to connect it back to Stanton at this stage.
0
Options
NEO|PhyteThey follow the stars, bound together.Strands in a braid till the end.Registered Userregular
Assuming they didn't bullshot us and actually got a functional prototype of server meshing running, that does seem like actual big capital n News. Especially for multiplayer gaming at large if they are willing to sell the tech once they've got it all hammered out.
It was that somehow, from within the derelict-horror, they had learned a way to see inside an ugly, broken thing... And take away its pain.
Warframe/Steam: NFyt
"But doing so makes you look like a bit of a goose"
I don't think this is true...Roberts and co have to prove something as the current track record is abysmal. If anything it's more goosery just blindly believing what is said.
It needs to be in players hands for me to take any of it seriously. There is comprehension of the engineering, we just also comprehend what has been said in the past vs what was delivered.
In most projects i've been on feature complete is the first 80% and the final cleanup is the other 80%.
Yes. The quote "Ultimately, this is the final phase of gameplay iteration before we fully transition into optimization and stability on the road to release" does not, to me, suggest it will be released in the next year.
In most projects i've been on feature complete is the first 80% and the final cleanup is the other 80%.
Yes. The quote "Ultimately, this is the final phase of gameplay iteration before we fully transition into optimization and stability on the road to release" does not, to me, suggest it will be released in the next year.
As a reminder, the last time they said Squadron 42 would be feature complete by the end of the year and ready to release in the following year was 2019. And the same thing in 2016. I'll believe it's done when it's actually available for sale.
Also Feature Complete does not mean Content Complete. A game could be feature complete and still have a long time in development to complete all the art, sound, maps, ships, items, missions, writing, dialog, cutscenes, etc.
SiliconStew on
Just remember that half the people you meet are below average intelligence.
I still don’t fully understand a) why server meshing is such a huge focus for this game, and b) why successfully implementing any form of server meshing for this game would necessarily have any significant implications outside of this game.
As best as I understand it (and having come at this a few times, I do think I get the idea, because it’s the execution that seems to be the actually complicated part), there are going to be a series of trade offs and design decisions as a result of pursuing maximum seamlessness for a maximum number of people. Those decisions are going to be very specific not just for reaching that goal, but also for reaching that goal in the context of a space shooter/FPS. And any decisions about how to prioritize a specific agent or event on a given server over other servers is going to be pretty dependent on the context and goals.
So unless a dev is looking to make something pretty close to Star Citizen, I’m not sure all that much work has been done for them.
And admittedly, I’m still skeptical that any implementation is going to survive at scales far beyond current MMO instance levels without doing the EVE time dilation thing, so even allowing a higher player cap to interact doesn’t seem like it’s inherently opening up gameplay doors?
And I feel more confident that I’m not missing some tremendous implication of this technology when the features they’re emphasizing in their buzz reel are variations on ‘Press F to perform canned animation.’
Like, I’m not invested in this aside from finding the project fascinating, wanting it to succeed way back in the day, and the niggling feeling I must be missing something about this specific element that gets emphasized so hard, so definitely feel free to dismiss all this if those are your litmus tests. But if I am wrong about how this is meaningful, I would genuinely appreciate an explanation because the premise is interesting and I’m really not just here to cheer on a car crash.
Edit: To be clear, it would be fun to be wrong and we’re on the verge of a transformational technology for gaming. Genuinely, that’s exciting news. I just can’t see how that is what this is, and I swear I have tried with some real effort multiple times. Part of the problem may be the lack of details on implementation mean most explainers are speculative?
OneAngryPossum on
+2
Options
MonwynApathy's a tragedy, and boredom is a crime.A little bit of everything, all of the time.Registered Userregular
It's a marketing buzzword intended to distract you from the fact that the game world crashes if you look at it wrong because they do not currently have Server Meshing (TM) enabled
I still don’t fully understand a) why server meshing is such a huge focus for this game, and b) why successfully implementing any form of server meshing for this game would necessarily have any significant implications outside of this game.
It's entirely for a practical reason.
Right now, if you log into the game and play in the live Persistent Universe (PU), the game world is being hosted on a single server, with up to 100 players per identical copy of the world server. And if you were to log in right now, with a single server hosting the entire game world, you would notice a lot of problems...
-NPCs AI gets broken very easily. The server is incapable of simulating thousands of NPCs across the entire game world, because that's just too many processes to handle for a single server.
-One single server is trying to host the entire entity graph (aka persistent entity streaming). The more items that are spawned, the more ships, the more players, the more entities it has to track the more it gets bogged down.
-They want to simulate economy, with fluctuating prices and game events having a tangible effect on the in-game markets. But they can't do any of that stuff because the one single game server is already maxed out with everything else it's doing.
And a lot of other things similar.
The problem is that one server cannot handle the load they're putting on it. That's why there's a need for server meshing. It's entirely a practical one.
It's not some industry buzzword as implied. It's a real fix to a real problem. They're already pushing the capabilities of a single server to the absolute max.
The idea of Star Citizen is that it is supposed to a seamless world. It cannot stay seamless and also grow. Stanton is already maxing it out. And they can't even have Stanton do everything they want it to do, let alone adding additional star systems.
It's a marketing buzzword intended to distract you from the fact that the game world crashes if you look at it wrong because they do not currently have Server Meshing (TM) enabled
This is an incredibly cynical take and also 100% wrong. Server meshing is a real thing, not just a buzz word. And it actually has nothing to do with crashes. I believe you're looking for the term "replication layer" if you want to talk about crash protection. The replication layer creates redundancy to help protect against crashes.
Server meshing is exactly what the name implies. It is a mesh of multiple servers all working in tandem to create the playable game world, or in this case game universe. Server meshing on its own has nothing to do with crash protection.
You are right that the servers tend to fall over if you look at them wrong. But that's because of the reason I mentioned about already being strained to the max. They literally cannot fit any more data on the ONE server that is currently trying to simulate the star system. Thus the need to integrate multiple servers into a seamless mesh.
I still don’t fully understand a) why server meshing is such a huge focus for this game, and b) why successfully implementing any form of server meshing for this game would necessarily have any significant implications outside of this game.
Server meshing is a scaling technology intended to allow an unlimited number of players to interact seamlessly in a single shared world. The ability to do this would benefit any multiplayer game, hence why it seems CIG's future business plans are to licence this engine technology to others.
There is only so much computing power you can spend simulating the area around a player or group of players. To maintain performance you traditionally need to split up your world to either limit the number of players in a single area or make sure players cannot interact.
Let's say you have a map the performs well if less than 10 players are in it. What if 10 more players join? Or 100 or 1000? If that map is still being run on a single server, your performance tanks. But what if your system could automatically scale to add more servers for every 10 players? Great, easy to do except your players are still segregated into individual servers. But what if you have 10 servers with 1 player each? Do you waste money running 90% of your infrastructure idle or do you merge the servers and have players see 9 other players pop into existence from nowhere. It would be better if things were simulated in a shared way such that those other 9 players seemed to be there from the beginning even if they were really on different servers. And what if you want those players on different servers to interact as though they were on the same map? Now you have to design an inter-server communication system that tracks all interactions that might occur across those server borders. This is not a trivial problem to solve. Made even harder when your servers are constantly splitting or merging to maintain performance as player density in a given area changes.
It's a great technology if they've gotten it to work. They've just been saying it was nearly ready for release for the past 5 years straight so they've built up a lot of scepticism.
Just remember that half the people you meet are below average intelligence.
I still don’t fully understand a) why server meshing is such a huge focus for this game, and b) why successfully implementing any form of server meshing for this game would necessarily have any significant implications outside of this game.
It's entirely for a practical reason.
Right now, if you log into the game and play in the live Persistent Universe (PU), the game world is being hosted on a single server, with up to 100 players per identical copy of the world server. And if you were to log in right now, with a single server hosting the entire game world, you would notice a lot of problems...
-NPCs AI gets broken very easily. The server is incapable of simulating thousands of NPCs across the entire game world, because that's just too many processes to handle for a single server.
-One single server is trying to host the entire entity graph (aka persistent entity streaming). The more items that are spawned, the more ships, the more players, the more entities it has to track the more it gets bogged down.
-They want to simulate economy, with fluctuating prices and game events having a tangible effect on the in-game markets. But they can't do any of that stuff because the one single game server is already maxed out with everything else it's doing.
And a lot of other things similar.
The problem is that one server cannot handle the load they're putting on it. That's why there's a need for server meshing. It's entirely a practical one.
It's not some industry buzzword as implied. It's a real fix to a real problem. They're already pushing the capabilities of a single server to the absolute max.
The idea of Star Citizen is that it is supposed to a seamless world. It cannot stay seamless and also grow. Stanton is already maxing it out. And they can't even have Stanton do everything they want it to do, let alone adding additional star systems.
It's a marketing buzzword intended to distract you from the fact that the game world crashes if you look at it wrong because they do not currently have Server Meshing (TM) enabled
This is an incredibly cynical take and also 100% wrong. Server meshing is a real thing, not just a buzz word. And it actually has nothing to do with crashes. I believe you're looking for the term "replication layer" if you want to talk about crash protection. The replication layer creates redundancy to help protect against crashes.
Server meshing is exactly what the name implies. It is a mesh of multiple servers all working in tandem to create the playable game world, or in this case game universe. Server meshing on its own has nothing to do with crash protection.
You are right that the servers tend to fall over if you look at them wrong. But that's because of the reason I mentioned about already being strained to the max. They literally cannot fit any more data on the ONE server that is currently trying to simulate the star system. Thus the need to integrate multiple servers into a seamless mesh.
It's a cynical take because the Star Citizen devs deserve nothing but cynicism.
I appreciate the responses, and that’s what I understood the goal and technology to be, so that’s reassuring.
I’m still not sure how that tech has broad applications outside of this niche given it still imposes significant design constraints outside of the ‘seamless with large number of players’ aspect unless somehow they’re maintaining states across a dynamic number of servers without significant processing/transmission time. Or how whatever solution they have could truly apply at a significantly larger scale without vastly limiting player interaction (which is the thing they seem to want to avoid most), or implementing an EVE like slowdown system when certain thresholds are crossed.
I guess my hang up is that this tech is discussed as kind of revolutionary, but every practical implementation I can think of feels iterative at best, despite the difficulty of the problem.
0
Options
MonwynApathy's a tragedy, and boredom is a crime.A little bit of everything, all of the time.Registered Userregular
I still don’t fully understand a) why server meshing is such a huge focus for this game, and b) why successfully implementing any form of server meshing for this game would necessarily have any significant implications outside of this game.
It's entirely for a practical reason.
Right now, if you log into the game and play in the live Persistent Universe (PU), the game world is being hosted on a single server, with up to 100 players per identical copy of the world server. And if you were to log in right now, with a single server hosting the entire game world, you would notice a lot of problems...
-NPCs AI gets broken very easily. The server is incapable of simulating thousands of NPCs across the entire game world, because that's just too many processes to handle for a single server.
-One single server is trying to host the entire entity graph (aka persistent entity streaming). The more items that are spawned, the more ships, the more players, the more entities it has to track the more it gets bogged down.
-They want to simulate economy, with fluctuating prices and game events having a tangible effect on the in-game markets. But they can't do any of that stuff because the one single game server is already maxed out with everything else it's doing.
And a lot of other things similar.
The problem is that one server cannot handle the load they're putting on it. That's why there's a need for server meshing. It's entirely a practical one.
It's not some industry buzzword as implied. It's a real fix to a real problem. They're already pushing the capabilities of a single server to the absolute max.
The idea of Star Citizen is that it is supposed to a seamless world. It cannot stay seamless and also grow. Stanton is already maxing it out. And they can't even have Stanton do everything they want it to do, let alone adding additional star systems.
It's a marketing buzzword intended to distract you from the fact that the game world crashes if you look at it wrong because they do not currently have Server Meshing (TM) enabled
This is an incredibly cynical take and also 100% wrong. Server meshing is a real thing, not just a buzz word. And it actually has nothing to do with crashes. I believe you're looking for the term "replication layer" if you want to talk about crash protection. The replication layer creates redundancy to help protect against crashes.
Server meshing is exactly what the name implies. It is a mesh of multiple servers all working in tandem to create the playable game world, or in this case game universe. Server meshing on its own has nothing to do with crash protection.
You are right that the servers tend to fall over if you look at them wrong. But that's because of the reason I mentioned about already being strained to the max. They literally cannot fit any more data on the ONE server that is currently trying to simulate the star system. Thus the need to integrate multiple servers into a seamless mesh.
Here's an idea, instead of spending ten years trying to create new technology to solve problems that don't matter after taking over half a billion dollars from your target audience, scope your game appropriately so it doesn't fall the fuck over if someone drops a Peacebloom on the ground
Monwyn on
+6
Options
NEO|PhyteThey follow the stars, bound together.Strands in a braid till the end.Registered Userregular
I appreciate the responses, and that’s what I understood the goal and technology to be, so that’s reassuring.
I’m still not sure how that tech has broad applications outside of this niche given it still imposes significant design constraints outside of the ‘seamless with large number of players’ aspect unless somehow they’re maintaining states across a dynamic number of servers without significant processing/transmission time. Or how whatever solution they have could truly apply at a significantly larger scale without vastly limiting player interaction (which is the thing they seem to want to avoid most), or implementing an EVE like slowdown system when certain thresholds are crossed.
I guess my hang up is that this tech is discussed as kind of revolutionary, but every practical implementation I can think of feels iterative at best, despite the difficulty of the problem.
There's not really too much I can think of where this technology would actually matter, but I suspect that's a lot of cases of because the tech hasn't existed there haven't been a real push to do things it would benefit. The closest thing I can think of is EVE Online. When you get big fleet battles, the servers tend to start melting for that system, even if they had advanced warning and were able to allocate more hardware ahead of time, which they were able to Band-Aid with time dilation; rather than just letting the server melt, they slow down time in that node so it has more time to process everything.
What server meshing as envisioned would enable you to do is just dynamically add more server processing to that in-game system, so as players drop in for the fleet battle, the back end spins up server clusters to hold them and keeps them talking to each other so all the players can see each other without pop-in, and keeping any specific server from melting down and degrading gameplay.
It was that somehow, from within the derelict-horror, they had learned a way to see inside an ugly, broken thing... And take away its pain.
Warframe/Steam: NFyt
I appreciate the responses, and that’s what I understood the goal and technology to be, so that’s reassuring.
I’m still not sure how that tech has broad applications outside of this niche given it still imposes significant design constraints outside of the ‘seamless with large number of players’ aspect unless somehow they’re maintaining states across a dynamic number of servers without significant processing/transmission time. Or how whatever solution they have could truly apply at a significantly larger scale without vastly limiting player interaction (which is the thing they seem to want to avoid most), or implementing an EVE like slowdown system when certain thresholds are crossed.
I guess my hang up is that this tech is discussed as kind of revolutionary, but every practical implementation I can think of feels iterative at best, despite the difficulty of the problem.
I think a large part of the concept here is that they still want to fit as many people into a single server as they can.
Server Meshing, as a tech for SC is basically broken into two major phases:
Static Server Meshing - In which the servers are pre-defined, and the mesh is logically defined around places of high player concentration. Or the likelihood of high player concentration.
Dynamic Server Meshing - It works on an automated scale, where servers are spun-up and dropped into the mesh as needed. This is a reactive system, and the reaction is handled by some sort of automated governing logic.
With static server meshing, they might break the Stanton system up by planets. That's a fairly logical way they could break it up. So instead of one server trying to run the entire Staton system, now they have 6 servers running it. A server dedicated to Hurston. And a server dedicated to ArcCorp. And a server dedicated to MicroTech. And so forth. Each server would still cover a very large amount of playable space, but it would still be a much more manageable load. And if there was a particular space station or location that sees a huge amount of traffic, like for example if they are doing one of their annual Expos in a convention hall, they could dedicate a server just to the convention hall or something. The main idea being that in static server meshing, it's all pre-defined and setup by network engineers.
Once they move to dynamic server meshing, the idea is that the system will have the ability to self-regulate, and if a particular area gets too crowded, they can spin up extra servers dynamically. For instance, let's say a player Org decides they want to have an Org party on board an 890 Jump, the dynamic server mesh system could theoretically spin up a new server just for the 890 Jump, and it could host the 75 players who are partying together on their own dynamic server. And then once the crowd disperses and that extra server is no longer necessary, the dynamic system would remove that server from the mesh. Or if there's a Jump Town PVP event happening at a specific base on a planet, they could drop in a server for that event.
0
Options
HardtargetThere Are Four LightsVancouverRegistered Userregular
I can certainly criticise them for taking 10 years to make a game any half decent studio would have done in only a couple with a faction of the money they raised. and i will.
I mean, sure. Go right ahead. You're free to voice your opinion.
But doing so makes you look like a bit of a goose. It demonstrates that you have no comprehension of what an incredible feat of network engineering the effort has been to create server meshing. It took 11 years because this is something that has never been done before. This is literally brand new technology that they have had to invent. And yes, they have had some false starts and they had to entirely scrap an earlier implementation of it, because it wasn't working. But that's not because they're a bad studio. It's because they are doing something that has never been done before. This isn't a matter of something "a half decent studio" should have done in a couple of years. They have some of the top network engineers in the world working on this, and they are doing something in gaming that has never been done before. (Maybe outside of gaming too).
Other video games use technology such as sharding, phasing, and other tech to instance people and keep servers from exploding from having too many people in a space at one time. That's not new. But what *is* new is the way they're allowing people who are on different servers to be able to interact and see each other across those servers.
You know how games like WoW have zone lines and load screens? SC will have none of that. You know how in WoW, if a player walks from the world map into a dungeon, it loads them into an instance server and removes them from the world map? That won't happen in SC. In SC, if one person was standing on the world map and another person was standing inside a dungeon instance, they would be able to look at each other across the zone line, and they would be able to shoot each other. From different servers. That tech does not exist today. Servers don't allow interaction across them like this. That's the new part.
Having multiple servers, instances, sharding, and phasing is old tech that exists today. Allowing people to interact and see each other across different servers has never been done before.
You're free to criticize them. But being critical indicates that you don't have an understanding or an appreciation for what they have managed to accomplish. They aren't just delivering a video game here. They are delivering the future of server technology.
how does any of what you said have anything to do with Squadron 42?
Posts
Enlist in Star Citizen! Citizenship must be earned!
3.19 mining laser changes
Pffft Lancet. Lol.
Rofl
Etc.
PSN: ShogunGunshow
Origin: ShogunGunshow
What was it before? What changed?
Origin ID: Discgolfer27
Untappd ID: Discgolfer1981
Turns out even a weak pissbaby laser can mine quant ez when it can lower the rock's resistance by that much, and it kind of became the go-to for most everything.
The Lancet in 3.19 is, uh... Not that. For some reason it's still referred to as a support laser, but tbh it's more statted like a budget Helix. It has the second highest laser output in terms of mins and maxes, a small reduction to instability, no reduction to resistances, middle of the road mod slotting, and crushes your optimal charge window by 60%. Its big bonus is ... 40% optimal charge rate. Which is... not good. Being able to get the rock to the green zone and reduce the risk of overcharging it is just so much more valuable than getting the rock to charge a few seconds faster.
PSN: ShogunGunshow
Origin: ShogunGunshow
The idea being that you are supposed to have a target resource in mind beforehand, and you match the laser to the type of resource you’re mining.
To go along with that laser rebalance, Mike also said they did a big resource placement shuffle. Certain resources are now only found at certain locations.
So… if you know you want to mine X, you equip the laser best suited to X and then you fly to the planet/moon/asteroid where X is known to spawn.
PSN: ShogunGunshow
Origin: ShogunGunshow
Warframe/Steam: NFyt
Robot companion and therefore pets and therefore the petting, feeding, housing, and persistence thereof.
(TW for AI, Creature, Item, Vehicle, and Story teams)
I think pets with all the things you mention was one of the early stretch goals ~10 years ago.
He'll probably see the build-your-own-spaceship interface and want that, requiring a full rework of all spaceships
I could see it. You know. In another ten years when it's implemented. Right around the time we get our third star system lololol
PSN: ShogunGunshow
Origin: ShogunGunshow
When your pet craps on a moon, you can leave and come back weeks later and it will still be there.
and the game crashes with all the poops
Looking forward to the tier 2 implementation where the poop animations are implemented, but the poop itself is still represented as a pile of 1/64 SCU Fertilizer crates.
Big items from the con include:
-A live, onstage demonstration of Server Meshing + Replication Layer at work. The tech is real. It exists. They did it. They got it working. Yes, it took forever. But they have delivered a new form of server tech/architecture that has never existed in video games before now.
-An announcement that Pyro (and probably Server Meshing by necessity) will be going up on a PTU server on Oct 31. This is huge. Also, the Pyro system was playable at the show on demo stations.
-The announcement that Squadron 42 is feature complete and has moved into the final stages of production, which is quality pass, optimization, iteration, balance, and finishing touches. No release date given, but at this point the release is an inevitability. Maybe 2024. Definitely no later than 2025.
-Engineering gameplay is coming along quite nicely. They showed an in-game demonstration set on board a Crusader Hercules starship. They walked us through routing power, repairing and replacing components, turning components on and off via the engineering control station, opening and closing doors remotely, and also fire propagation when all hell breaks loose on the ship.
I mean, sure. Go right ahead. You're free to voice your opinion.
But doing so makes you look like a bit of a goose. It demonstrates that you have no comprehension of what an incredible feat of network engineering the effort has been to create server meshing. It took 11 years because this is something that has never been done before. This is literally brand new technology that they have had to invent. And yes, they have had some false starts and they had to entirely scrap an earlier implementation of it, because it wasn't working. But that's not because they're a bad studio. It's because they are doing something that has never been done before. This isn't a matter of something "a half decent studio" should have done in a couple of years. They have some of the top network engineers in the world working on this, and they are doing something in gaming that has never been done before. (Maybe outside of gaming too).
Other video games use technology such as sharding, phasing, and other tech to instance people and keep servers from exploding from having too many people in a space at one time. That's not new. But what *is* new is the way they're allowing people who are on different servers to be able to interact and see each other across those servers.
You know how games like WoW have zone lines and load screens? SC will have none of that. You know how in WoW, if a player walks from the world map into a dungeon, it loads them into an instance server and removes them from the world map? That won't happen in SC. In SC, if one person was standing on the world map and another person was standing inside a dungeon instance, they would be able to look at each other across the zone line, and they would be able to shoot each other. From different servers. That tech does not exist today. Servers don't allow interaction across them like this. That's the new part.
Having multiple servers, instances, sharding, and phasing is old tech that exists today. Allowing people to interact and see each other across different servers has never been done before.
You're free to criticize them. But being critical indicates that you don't have an understanding or an appreciation for what they have managed to accomplish. They aren't just delivering a video game here. They are delivering the future of server technology.
Warframe/Steam: NFyt
I backed in the initial campaign and haven't been paying attention in the last couple years. More on foot than I expected, but what was shown was pretty awesome looking to me. I hope actual gameplay is much heavier on the flight aspect than the shooter side.
Server meshing won't be in the Pyro preview channel. It won't have jumping between Stanton & Pyro, so it won't need server meshing. Chris Roberts said it was only shown working 3 weeks ago. Making it similar to when he made the PES letter to the chairman. Which we then waited another year for that to release, and that was in a terrible state.
_____________________________________
-'Squadron 42' trailer impressed me. I didn't really care about 42 before this, now I'm kinda interested. Trailer exceeded my expectations.
-Base building was just more theory crafting. Same old, same old for 'Star Citizen'. Won't even begin development until 2024.
-Freight Elevators When!?!
-Distribution centers & raids seem cool. They looked way too small to necessitate roads or encourage vehicle usage though.
-Very happy to see free roaming quantum is returning.
-Maps looks great. Downloadable areas, placeable markers, searchable locations.
-Dunno how I feel about them making fps portion more CoD like. Will have to see how that feels.
-Water wake looks great.
-Inventory looks functional.
Ahh, ok. I missed that comment. I didn't realize that they were not going to connect it back to Stanton at this stage.
Warframe/Steam: NFyt
I don't think this is true...Roberts and co have to prove something as the current track record is abysmal. If anything it's more goosery just blindly believing what is said.
It needs to be in players hands for me to take any of it seriously. There is comprehension of the engineering, we just also comprehend what has been said in the past vs what was delivered.
Not to be too flippant, but I really just wanted a video game.
Yes. The quote "Ultimately, this is the final phase of gameplay iteration before we fully transition into optimization and stability on the road to release" does not, to me, suggest it will be released in the next year.
As a reminder, the last time they said Squadron 42 would be feature complete by the end of the year and ready to release in the following year was 2019. And the same thing in 2016. I'll believe it's done when it's actually available for sale.
Also Feature Complete does not mean Content Complete. A game could be feature complete and still have a long time in development to complete all the art, sound, maps, ships, items, missions, writing, dialog, cutscenes, etc.
As best as I understand it (and having come at this a few times, I do think I get the idea, because it’s the execution that seems to be the actually complicated part), there are going to be a series of trade offs and design decisions as a result of pursuing maximum seamlessness for a maximum number of people. Those decisions are going to be very specific not just for reaching that goal, but also for reaching that goal in the context of a space shooter/FPS. And any decisions about how to prioritize a specific agent or event on a given server over other servers is going to be pretty dependent on the context and goals.
So unless a dev is looking to make something pretty close to Star Citizen, I’m not sure all that much work has been done for them.
And admittedly, I’m still skeptical that any implementation is going to survive at scales far beyond current MMO instance levels without doing the EVE time dilation thing, so even allowing a higher player cap to interact doesn’t seem like it’s inherently opening up gameplay doors?
And I feel more confident that I’m not missing some tremendous implication of this technology when the features they’re emphasizing in their buzz reel are variations on ‘Press F to perform canned animation.’
Like, I’m not invested in this aside from finding the project fascinating, wanting it to succeed way back in the day, and the niggling feeling I must be missing something about this specific element that gets emphasized so hard, so definitely feel free to dismiss all this if those are your litmus tests. But if I am wrong about how this is meaningful, I would genuinely appreciate an explanation because the premise is interesting and I’m really not just here to cheer on a car crash.
Edit: To be clear, it would be fun to be wrong and we’re on the verge of a transformational technology for gaming. Genuinely, that’s exciting news. I just can’t see how that is what this is, and I swear I have tried with some real effort multiple times. Part of the problem may be the lack of details on implementation mean most explainers are speculative?
It's entirely for a practical reason.
Right now, if you log into the game and play in the live Persistent Universe (PU), the game world is being hosted on a single server, with up to 100 players per identical copy of the world server. And if you were to log in right now, with a single server hosting the entire game world, you would notice a lot of problems...
-NPCs AI gets broken very easily. The server is incapable of simulating thousands of NPCs across the entire game world, because that's just too many processes to handle for a single server.
-One single server is trying to host the entire entity graph (aka persistent entity streaming). The more items that are spawned, the more ships, the more players, the more entities it has to track the more it gets bogged down.
-They want to simulate economy, with fluctuating prices and game events having a tangible effect on the in-game markets. But they can't do any of that stuff because the one single game server is already maxed out with everything else it's doing.
And a lot of other things similar.
The problem is that one server cannot handle the load they're putting on it. That's why there's a need for server meshing. It's entirely a practical one.
It's not some industry buzzword as implied. It's a real fix to a real problem. They're already pushing the capabilities of a single server to the absolute max.
The idea of Star Citizen is that it is supposed to a seamless world. It cannot stay seamless and also grow. Stanton is already maxing it out. And they can't even have Stanton do everything they want it to do, let alone adding additional star systems.
This is an incredibly cynical take and also 100% wrong. Server meshing is a real thing, not just a buzz word. And it actually has nothing to do with crashes. I believe you're looking for the term "replication layer" if you want to talk about crash protection. The replication layer creates redundancy to help protect against crashes.
Server meshing is exactly what the name implies. It is a mesh of multiple servers all working in tandem to create the playable game world, or in this case game universe. Server meshing on its own has nothing to do with crash protection.
You are right that the servers tend to fall over if you look at them wrong. But that's because of the reason I mentioned about already being strained to the max. They literally cannot fit any more data on the ONE server that is currently trying to simulate the star system. Thus the need to integrate multiple servers into a seamless mesh.
Server meshing is a scaling technology intended to allow an unlimited number of players to interact seamlessly in a single shared world. The ability to do this would benefit any multiplayer game, hence why it seems CIG's future business plans are to licence this engine technology to others.
There is only so much computing power you can spend simulating the area around a player or group of players. To maintain performance you traditionally need to split up your world to either limit the number of players in a single area or make sure players cannot interact.
Let's say you have a map the performs well if less than 10 players are in it. What if 10 more players join? Or 100 or 1000? If that map is still being run on a single server, your performance tanks. But what if your system could automatically scale to add more servers for every 10 players? Great, easy to do except your players are still segregated into individual servers. But what if you have 10 servers with 1 player each? Do you waste money running 90% of your infrastructure idle or do you merge the servers and have players see 9 other players pop into existence from nowhere. It would be better if things were simulated in a shared way such that those other 9 players seemed to be there from the beginning even if they were really on different servers. And what if you want those players on different servers to interact as though they were on the same map? Now you have to design an inter-server communication system that tracks all interactions that might occur across those server borders. This is not a trivial problem to solve. Made even harder when your servers are constantly splitting or merging to maintain performance as player density in a given area changes.
It's a great technology if they've gotten it to work. They've just been saying it was nearly ready for release for the past 5 years straight so they've built up a lot of scepticism.
It's a cynical take because the Star Citizen devs deserve nothing but cynicism.
I’m still not sure how that tech has broad applications outside of this niche given it still imposes significant design constraints outside of the ‘seamless with large number of players’ aspect unless somehow they’re maintaining states across a dynamic number of servers without significant processing/transmission time. Or how whatever solution they have could truly apply at a significantly larger scale without vastly limiting player interaction (which is the thing they seem to want to avoid most), or implementing an EVE like slowdown system when certain thresholds are crossed.
I guess my hang up is that this tech is discussed as kind of revolutionary, but every practical implementation I can think of feels iterative at best, despite the difficulty of the problem.
Here's an idea, instead of spending ten years trying to create new technology to solve problems that don't matter after taking over half a billion dollars from your target audience, scope your game appropriately so it doesn't fall the fuck over if someone drops a Peacebloom on the ground
There's not really too much I can think of where this technology would actually matter, but I suspect that's a lot of cases of because the tech hasn't existed there haven't been a real push to do things it would benefit. The closest thing I can think of is EVE Online. When you get big fleet battles, the servers tend to start melting for that system, even if they had advanced warning and were able to allocate more hardware ahead of time, which they were able to Band-Aid with time dilation; rather than just letting the server melt, they slow down time in that node so it has more time to process everything.
What server meshing as envisioned would enable you to do is just dynamically add more server processing to that in-game system, so as players drop in for the fleet battle, the back end spins up server clusters to hold them and keeps them talking to each other so all the players can see each other without pop-in, and keeping any specific server from melting down and degrading gameplay.
Warframe/Steam: NFyt
I think a large part of the concept here is that they still want to fit as many people into a single server as they can.
Server Meshing, as a tech for SC is basically broken into two major phases:
Static Server Meshing - In which the servers are pre-defined, and the mesh is logically defined around places of high player concentration. Or the likelihood of high player concentration.
Dynamic Server Meshing - It works on an automated scale, where servers are spun-up and dropped into the mesh as needed. This is a reactive system, and the reaction is handled by some sort of automated governing logic.
With static server meshing, they might break the Stanton system up by planets. That's a fairly logical way they could break it up. So instead of one server trying to run the entire Staton system, now they have 6 servers running it. A server dedicated to Hurston. And a server dedicated to ArcCorp. And a server dedicated to MicroTech. And so forth. Each server would still cover a very large amount of playable space, but it would still be a much more manageable load. And if there was a particular space station or location that sees a huge amount of traffic, like for example if they are doing one of their annual Expos in a convention hall, they could dedicate a server just to the convention hall or something. The main idea being that in static server meshing, it's all pre-defined and setup by network engineers.
Once they move to dynamic server meshing, the idea is that the system will have the ability to self-regulate, and if a particular area gets too crowded, they can spin up extra servers dynamically. For instance, let's say a player Org decides they want to have an Org party on board an 890 Jump, the dynamic server mesh system could theoretically spin up a new server just for the 890 Jump, and it could host the 75 players who are partying together on their own dynamic server. And then once the crowd disperses and that extra server is no longer necessary, the dynamic system would remove that server from the mesh. Or if there's a Jump Town PVP event happening at a specific base on a planet, they could drop in a server for that event.
how does any of what you said have anything to do with Squadron 42?
also actual main stream coverage! - https://arstechnica.com/gaming/2023/10/star-citizens-squadron-42-campaign-is-feature-complete-after-11-years/