Forget server meshing, let's talk about that opening line. Yikes! Oh also, OVERSTATED. Cannot be overstated* Finally, A Major Update: th-cam.com/video/1XXKkLjK4Vc/w-d-xo.html
Listen, you always make great videos. But come'on. You're talking as if "server meshing" is some crazy new technology that RSI has has to create/invent to make this crazy new thing happen that's new to MMOs. But this is stupid. "server meshing" is as fundamental to MMOs as a UI. I remember the 2 second pauses between servers while running around in UO. This is NOT NEW and EXTREMELY unimpressive and embarrassing that this is where they are at after 12 years and a billion dollars.
@@Vikkoify I don't know man I tried REAL hard in this video to say THIS IS NOT NEW TECHNOLOGY OTHER PEOPLE HAVE DONE THINGS LIKE THIS, but it always gets missed. Where in the video am I saying it's new? I even put EVE and Dual Universe on the screen? I highlighted the one thing that was actually impressive about this implementation, that there IS no 2 second pause. I've talked to a lot of people about this, so I understand how unimpressive it might be to you, it's why I only tried to draw attention specifically to the unique parts. What they've done is brilliant, maybe it's not the first of it's kind, perhaps it's not going to change the world, and it's definitely is not solving other games. But it's a big deal for this game, and it's an achievement that only a few in the industry can claim. I'm excited for this tech, and from what I've seen it's not easy to make, so I'm gonna give em credit on it. The title of the video isn't "this will fundamentally change gaming", it's just Star Citizen that's moving forward, and that's a positive.
@@valeriansamborski5844 eh, I'm not thinking about Musk for it, more just the amount of time it took the company to develop their tech advantage into a pepper product, ya know?
What they are attempting is nothing like what you describe in anything but the most basic level. But hey, I'm just a Sysadmin of 30 years and gamer for longer.@@Vikkoify
In the new "Letter from the chairman" Benoit said "In the coming weeks and months, get ready for more Technical Preview tests with various mesh configurations: multiple game servers per solar system and seamless transitions without gates. We're talking about layouts where servers are dedicated to entire planets and moons, others focused solely on Landing Zones or other key locations, with plenty of higher player count experiments." Lets gooooooo.
@@Fuar11dynamic would be able to change dynamically, on demand and analyzing by itself the demand. Static means the charge and size and the survers will stay the same no matter the loads on each servers or needs to aliviate one or the other.
@Fuar11 The hype is very real, but yeah, dynamic means servers will turn on or off depending on if people are in them or not. Even potentially actively split an area that says a big fleet battle is happening. In real time. Spinning up more or less servers depending on where the action is, but I suspect that is very far down the road. that all being said, why it matters is running a server costs money. So if say, every planet has a server. And every star system is supposed to have an average of 3 planets. For the 100 star systems planned, that's 300 servers. Just for one region. Say NA, then you'd need 300 more for Europe. Asia etc. You can see right away it's a huge cash sink if you can't actively turn servers on or off. Even if players aren't using them. Now multiply that if we do every landing zone having a sever and you got yourself a problem. Make no mistake. Now is an exciting time to be in star citizen. But we are still a ways off from even a dozen star systems, but static server meshing will give us a lot. It is the last hurtle before they can develop the rest of the game in earnest. Since server meshing is so fundimental, CIG basically had to gut a whole game engine and reulbuild it to make it work for this type of networking. It makes sense that star citizen development has been slow for the last few years beyond sq42. They couldn't add more that they wouldn't have to just totally redo for the new tech. So from a 'running a company' perspective, why make it in the first place? But now that we have servermeshing, that means the flood gates are open.
@Fuar11 It definitely is in preparation for the dynamic mesh. Testing different scenarios will allow them to collect data on how servers perform under different loads, server startup time based on entity counts for different sized areas, and all sorts of metrics they can use to define the rules to dynamically alter the mesh.
The beauty of dynamic is: If 1000 players decide to meet at Lorville, a server will spool up, and take over Lorville, to reduce the strain on the overall experience.
Your videos are always the ones I share with those who ask about Star Citizen. You can take complex concepts and break them down into digestible chunks. We appreciate it!
Tomato you are really on a different level in terms of SC quality videos. So we'll explained, great graphics, excellent editing and your smooth voice. It's pure enjoyment watching each and every video you put out! ❤️🙏
I was just testing static server meshing last night. Very mixed results. When it wasn't recovering repeatedly over and over it was generally quite smooth. I saw my server change as I traveled. It. Blows my mind that we’re testing it already.
@@deejagers716 there should be a drop-down in the launcher. You have to download the game again basically. It's on wave two testing. I can't remember what that means
Its just the start. Now that it's discovered and working, even if not seamless now it's gonna be very short to transition into dynamic server meshing. Then it will be even crazier. The amount of bandwidth required will be insane tho 😂
Awesome video. I am glad that they are testing meshing now. That is going to expand the possibilities of this game especially when it comes to dat load. I wonder if it will improve npc behavior.
1. Great video, as always! 2. Did you see the Chairman's Letter that just came out and the information about all the NEW upcoming server meshing tests that will see multiple servers per solar system? We're talking servers for each planet, moon, even landing zones and with even higher player counts than we just saw tested over the weekend. Your video couldn't have had better timing, every player new and old needs to watch it!! 😊
Making this happen without desync is seriously impressive. Making it happen robustly and reliably will be a miracle if they can pull it off. Assuming the meshing code is game engine agnostic, packable, configurable and licensable to other studios, that tech engineering alone could power the game for the rest of its life irrespective of players buying things. Think Physx ubiquitous for really large scale games. Total War where every unit is actually a player. GTA with hundreds or thousands of concurrent players on an instance or simply all at once in the same world. No more instancing anything by necessity, but only if it's intended.
Once again @spacetomato you have knocked it out of the park on this one. Great vid and explanation on how this all comes together. Now to see it in live!!
"Do you think we'll be able to build bases anywhere in the Star Citizen universe soon? Having multiple systems with player bases sounds amazing! I'm new to the game, and it has huge potential. The only thing that feels off is the slow pace (that train ride after death is rough!)."
Yes, base building will be a thing. As far as soon? Eh, relative to development (12 years so far) sure. But actually soon? Probably not for quite a while.
You speak with such confidence about something not even out yet, that hasn't even been truly tested, by a company who still don't have a complete game after a decade of scope creep and poor management. Its load balancing and database replication on more nodes for better performance, there summed up your entire video. Its not easy to do properly that can scale to high demand in any application let alone gaming where its never been used quite like this before. Only in gaming do you see this level of glazing from fanboys for things not even released yet. They have yet to deliver on anything outside of tinkering proof of concepts duct taped together.
This is one of the best videos i have seen. There is a lot of info in this and it is explained perfectly. Great job man. I am going to share your video.
If they pull this off it will change the future of all mmo's in the future. Or immagine battlefield game with 1000 players per battlefield, the roof is basically endless on this tech. This also means that this tech is ghonna be a lifeline for the company as many wants this tech. Letts hope, because this might be future.
Alot of people run their mouth about this game calling it "Scam Citizen" but mark my words this will be the most epic and expansive game in existence. It has been an honor to be a part of the alpha testing, helping the developers find bugs during development.
I've always really appreciated the vision CIG has. While the game is heavily meme'd on i still watch it and with the server meshing finally looking like it's going to be real. This will change gaming, it'll change everything. I'm hoping.
Having the replication layer (what the player connects to) mix and match the results from all servers to make it look like one big world is smart but also scary. I think the point is: The servers calculate the physics and object and NPC locations etc, then sends the results to the replication server, which forwards the info to the player. The danger is that now the replication server is the weak point instead. But infinitely better than what we had before. Maybe they can even solve the overall load by sharding the replication server into one per system LOL. Still gonna be rough if a lot of players share one replication server and all go to the same location though. Oh and the persistence server will get a huge workout from all the reading and writing of data, constantly changing when players move things around. But... It's already a huge step in the right direction anyway.
@@MyAmazingUsername possibly, some actions that need high fidelity like combat actions might have their simulation completely replicated, instead of just the state, similarly to how most client-server games simulate movement on both ends at a higher rate and then synchronize it at a slower rate
Got to love these free fly events when the game is not playable because of lag, I don't want to introduce the game to friends if their experience is a lag fiesta.
This sounds like memcached but hyper juiced. This is actually incredible massive leap forward. I hope they have a research paper they release on their tech. Incredible.
I've been real hesitant to get into this due to the cost of entry and the performance issues. If this turns out as well as we hope, I think I might finally take the plunge.
I'm curious where the next inevitable bottle neck will pop up when dynamic server meshing scales up. I could see the replication layer being the next to overload where they'd need to start meshing those layers as well.
I assume there is HA/DR on the Replicator? Would he awful if there was anything that went wrong on that for all the servers. Super excited about this design!
I just wanna say. Even with just static server meshing, you can do a lot. Splitting each planet to its own server would be an obvious first step. We could have upwards of 400 players in just one star system. And the thing is, even if we only do a server per star system. We could still have a handful of star systems easy. dynamic server meshing is the holy grail. I suspect it will be a while. But this moment now is very reminiscent of when object containers happened. With just object containers, we got several planets and moons. Static server meshing is a magatidue more then object containers.
Personally, I think DSM is about a year away, tops. Also, once its in Im not sure how many people the shards can handle. Not for stability reasons, but think about the amenities we have. With wrecks and trash, things get a little crowded from time to time. Say you increase Stanton to 500 players (exaggeration a bit but bare with me), does Stanton have enough hangars, bunkers, cargo collection and delivery locations? Often times when doing bounties, I come across bounties from other players missions. Id love for them to increase the player cap, but I think we'll need a few more star systems before they drastically increase it to 1000 or more per shard.
@justicebrown1077 I mean, I hope you are right about the DSM timing, but I think you are over optimistic. They said they would have static meshing in 3 years. 6 years ago. I suspect in terms of the player count, if they make each planet and moon a server and each major landing zone. It doesn't take a lot of code to just casually make some of the random buildings around the city now be a drop-off point for stuff. They can also just add more points of interest now. (Including hangers, drop off points and such) a lot of the reason we don't have the Aaron halo yet is because of simply entity count since right now, one server manages the entire stanton system. But I guess that is the test we will be doing is seeing what these meshes can handle in terms of players and all the stuff they leave around
I wonder what server meshing can do with something like 3D modelers such as blender maybe even splitting load so much that rendering may take hour or less for even large projects with enough servers. However, for meshing servers itself, I was having my one thoughts on setting my own server and game system before seeing this video that took inspiration from zero down horizon Gaia and her subroutine AIs, for serves there will be prime servers and subordinate servers, if prime is to have dynamic animal life a subordinate server will be in charge of this while prime it self is which animals will move, basically doing all calculations for prime servers and you can add as many subordinate servers as needed. Now combining my prime and subordinate servers with meshing servers may potentially go well together.
After a decade of backing star citizen (my wallet smacks me every time I pull it out) this has me over the top excited. The potential gameplay could be infinite.. full org battles with 1000s of players as a small example What excites me is the players and their ability to create gameplay (ie: Daymar rally) This sim has always had the potential to revolutionize gaming and it seems I can finally see that peaking its head on the horizon Shoutout to all the “star citizen is a scam” folks 👋
You mention that other games have used server meshing. also that is not a pivotal time in game industry (mayeybe you say that to another video) howeverI was under the impression that indeed star citizen is potentially shaping the game industry exactly for this technology. do the oversell it ? or there is actually truth to that? like the other server meshing game would have end up in a bottleneck?
It's really specific to each implementation. What they are doing is pretty monumental because of the real time transitions in space, but something like a server transition in a jump gate is not very novel at this point. Where SC is really shaping the industry is with dynamic server meshing which allows those real time transitions to happen at any point in space as server resize themselves as needed. It's certainly a big deal, but until it can be applied to the succes of the game it's good to not go through the roof with hype lol
Corrections (based on my understanding of how SM works): 2:15 - I personally don't like that statement personally as "The next generation of Loading-Screens" describes Object-Container Streaming MUCH MORE than actual Server-Meshing. In OCS you as a client constantly load/unload the world around you as you move through it which I do think is "the next generation of Loading-Screens". Server-Meshing however is more about splitting up the load on Servers from being all done on a single server to splitting it up between multiple servers (based on Location and/or Purpose). So, I would say, Server-Meshing is "the next generation of Simulation" 4:01 - The state of entities is not really replicated in other containers. If we JUST talk about Object-Container Streaming and not RepLayer or ServerMeshing, then Object-Containers are basically FOLDERS for the GameWorld. They structure the game-world so that you as a player only need to load the folders and in those folders the files that are necessary for what you are currently doing. The Server also only needs to have those Folders and Files in it's simulation that all the players are currently interacting with and nothing that is not needed for anybody. That is SS-OCS (ServerSide Object-Container Streaming) 5:20 - The state of every entity doesn't need to be replicated on other servers nearby because otherwise the players would not see what is happening in the other servers. Players get all their data from the RepLayer (they are no longer connected to the server at all) and the RepLayer knows what is happening to all entities in the shard (not just the once from specific Object-Containers). The reason Replication on other servers nearby is needed is simply for the transitions between Servers. When one server needs to take authority over an entity it needs to be basically instant, otherwise we would have lag in player-movement, physics, ... if anything changes the Server. By already replicating all entities that are simulated by nearby servers the server now doesn't need to load those anymore (which is the most time-consuming part) but simply needs to start simulating it. 6:46 - The performance problems you have with your client aren't or at least have very little to do with the Server-Performance (unless SC still has some VERY BAD backend design). That is simply optimisation of the Rendering-Pipeline that still needs to happen. The bad physics and desync during an event like Daymar-Ralley is however often a result of bad Server-Performance which SM will/can help with 11:36 - From what I have heard (this is based on a podcast where one person said something he heard from some CIG Person): Pyro is currently too big for a single server to handle. There are too many entities in that system hence the reason why they can't test the entire system on a single server (so they only test parts on the Preview-Channel without SM). Unfortunately I can't say anymore which podcast I actually heard this so take this information (even if I am pretty sure I have heard it and that it was stated by somebody at CIG) with a grain of salt.
If individual servers don't need replicated data then how will each server handle operations it has to perform that require cross-server information? For example, if server 1 has a heat seeking missile that is trying to hit something in server 2.
@@bobbycrosby9765 I expect you mean what I wrote in Point 3 Read it again. I didn't say it doesn't need to be replicated on multiple nearby servers which it obviously does for transition sake (as you said). What I was trying to say with that point is that it is not necessary (like I heard so often) because if is it is not done clients wouldn't see the changes made in the other server. Clients have their own criteria that is replicated to them completely separate from what servers replicate so not replicating on nearby servers would obviously hurt transitions but not clients seeing what changes on the other servers
Even if star citizen doesnt make its goal, selling & licensing the tech out will make them 100 billion easy if they have a revenue share system. Allowing small studios to off shift thier success instead of leveraging gigantic budgets to build the tech themself. instead focusing on art & game loops and playability & storyline
Damn it really took them almost a billion dollars to discover load balancers and database replication. Truly one of the innovators of all time, just like Tesla.
The only way to mesh servers seamlessly is to keep hot backups online at all times whether used or not due to the time required to initialize a fresh server once a crash has happened assuming crash detection requires zero time. Best of luck in reducing the time required to recover from a 30k quickly.
@@kekkopf69 Your assertion that server meshing will save money is false. Server meshing requires hot backups to be constantly available, or a finite amount of time will be needed to bring a new server online. CiG would be better off just using Redis for their server mesh, but this, too, would require some time when a new server is brought online. Server meshing will increase the cost for CiG, assuming players don't want to wait.
Or not. Another option is to mesh with some overlap, allowing "nearby" servers to take over the charge (with some degradation?) while a new one in instantiated to replaced the broken one. Like some kind of RAID 5, but for servers.
@@eldarionmarchombre4568 Ok. You assume the replacement server has few objects to load to be meshed. It would have zero objects to load until a player touches it. However, it will have to load objects to get meshed, and getting meshed will always require some time. Assuming we are talking about Cloud Servers... a replacement server would have to get provisioned and then started before it can get meshed; I am assuming the image for the replacement server has the current game installed and does not need to get updated before it can run, I am also assuming the OS does not require any updates. OS updates will be necessary, as will game updates, and this is what will eat time, assuming you have no hot backups. Players will not want to wait 5 minutes for a 30k recovery, and this supports my initial thesis because hot backups require server nodes to be made available in advance of being used, and doing so eats money, assuming your Cloud Provider will allow you to offline a server node without being billed however I am aware of one Cloud Provider who charges money for server nodes regardless of their use.
i expect them to start with something basic with server meshing to then ramp it up overtime i mean we all know this game wont survive if the server performance stays shit.. we are only here because we are confident that these issues will be fixed
You see a lot of critics saying that the concept of server meshing exists in other games already. WoW has server phasing, ESO uses layering, SWTOR uses instances. Different executions of the same idea. People who think this is the same thing as CIG taking 3000 concurrent players and serializing the client session to pass between servers seamlessly, do not know what Server Meshing is, or why this is big. No company in history has ever done that which is exactly why you see developers on the outside claim that CIG's dreams are unrealistic. Because they haven't done it.
@@Tazytots Well, maybe you should read mine? Did I mention WoW, ESO, SWTOR? No, I did not. These are clearly wrong statements made by people who have no idea what they are talking about. But if you look at the implementation of DU and SpatialOS, its pretty much the same as what CIG implement. If you dont see the similarity, then quite frankly are doing the same thing the critics are doing.
@@SETHthegodofchaos DU is not the same thing as Dynamic Server Meshing. They're categorically not the same thing. DU is the closest a GAME has ever gotten to what SC wants to reach, but no. They aren't the same. I won't touch SpatialOS. It's effectively failed to produce any actual successful projects.
@@Tazytots Hm ok. Maybe I missed something then. How does the DU tech differ? They split the game world as well and put them the sections dynamically on different servers, no? They have systems in place to optimize loading and networking too. Like CIG has with OCS and PES. And that SpatialOS doesnt have any games doesnt worry you? They too have a service inbetween game servers and clients, just like the Replication Layer. It too has a in-memory cache and database connected. It also uses gRPC for communication. It dynamically divides the game world into sections and hands of authority over them to multiple game servers. Where is the different there that will make CIG's version work (better)?
CIG is nothing like Tesla. CIG is more like Scientology, but instead of paying money for the salvation that never comes, you're paying money for pretend spaceships in a broken and featureless "game" that will likely never fully release
Just to clarify, is it going to be that servers are handling both players AND environments separately? So you could have a TON of people in game all be at the same place at the same time, but also have 50% of people in one environment "shooting cannons" at the other environment seamlessly? If im right that sounds pretty cool
Based upon the video and my rough knowledge of this subject, they are doing positional meshing. So no, you couldn't have a TON of people in the same place, because since they're in the same place at the same time they're all on the same server. I think people are massively overselling the current state of this - don't get me wrong, it's still impressive! It's like most MMOs but no loading screen between "zones". Which is super cool because you can have a seamless world. But it doesn't really allow you to have more people to be in the same place like so many are letting on.
@@bobbycrosby9765 This is true for static server meshing, which is what CIG showed us at cit con, but the next step of this tech will be "dynamic server meshing", where the servers will be automatically alocated to different places based on the amount of entities in said places. Let's say for example a ton of players decide to hang out in the same place, with dynamic server meshing, there will basically be multiple servers assigned to that location to reduce stress. But the most important step was this one I'm pretty sure, the transition from static server meshing to dynamic will take some time and effort but not nearly as much.
@@PikChouk it's still not necessarily going to help. And it may actually hurt. If a bunch of people in 1 spot interact with 1 thing, that interaction is still going to have to go through the server with that 1 thing, even if all those players are on different servers. Then you're going to have to replicate data across those multiple servers because they're close enough to eachother that players can interact with them. It's not a silver bullet. It might help in some scenarios. But there definitely will be others where it all falls down. E.g. all the players shooting a missile at 1 entity. They will all converge on the 1 entity, and they will necessarily end up all on the same server, and all the surrounding servers may need to be informed of the movements for any processing they do. It's not anywhere near as much of a sure thing as people are making it out to be.
@@bobbycrosby9765 but a server could perfectly handle this scenario tho, the "server with that 1 thing" is the replication layer, what it does is that if a server crashes, it is automatically replaced by another one preventing players from totally crashing, in effect if a server still crashes which will be unlikely, you'll just a wait a few seconds and the servers will get replaced so you can get back into the action. At the moment, servers can handle the load of the entire stanton system, even if 1000 missiles are shot at 1 entity, if the server is limited to a very small space around this entity, the load will be incredibly lighter compared to what it can handle right now.
I feel like a shareholder instead of a gamer at this point, they are pushing the tech and innovating the industry. At this point it would be better for me if they were to convert our ships to equivalent shares and went public.
Server meshing aint that special Ultima Online had a version of server meshing in 1997. Yes dynamic server meshing would be amazing but there is no set date for this technology. 4.0 is supposed to come out later this year but that only has static server meshing. So really there is all this hype for a technology thats still just a proof on concept, but hey maybe if we give star citizen another billion dollars they can give us another presentation in the next 5 years.
Wow they built all of this amazing tech and didn't consider server cold starts? Sounds too stupid to be true.... interested to know why not keep a warm pool of servers ready to replace a down one or maybe the latency is due to the syncing time it's taking with the "Replication Layer" to get to the current world state. I want to know so much mooooree about this.
This is still Static servermeshing without the hope of ACTUAL Dynamic servermeshing for quite a few more years. I've covered this at such a granular level over an entire series of videos over 3 years ago already. The server technology is the main thing holding this entire development back. Current server technology isn't quite literally designed to handle ramp up and ramp down in such concurrent cycles. It will literally kill servers physically. Current server technology is only file based technology and not virtualisation based like quantum computing is. We have to wait for quantum computing servers to become the norm still. As I said I've covered this granularly over the years already. The overhead costs of current server technology also would lead to prohibitevly expensive running costs and failure rates. Just like you're seeing now with 5 min down to 2 min ramp up and ramp down issues. That's a tiny sample size that is still firmly based on static meshing technology otherwise know as file system based server technology. That's a best possible case for current hardware technology. It seems they finally have the net code 2/3 sorted out with the replication layering code but there is a huge way to go still. Don't get your hopes up at all. This is not going to be a real solution for years to come. CIG at Citizencon 2 years ago also publically admitted that as well. Over a year after I already proved that with industry evidence and so on. Spectrum was alive with comments and questions for that year or so prior asking the same thing based on what I had already granularly proved as well. Hence why CIG finally came clean about it! We are no where near REAL Dynamic servermeshing. Please don't get peoples hopes up. We don't need another community war breaking out again! Otherwise wonderful graphics in your presentation and I can see your hearts in the right place. But please let this sleeping dog conversation lay. This is something people get very "excitable" about as if you're an OG backer like myself and a massive amount of other people are, it can seriously set off some major arguments and resentment from the community. Trust me you don't want that. Been there done that and proved my self right 100% and took down the psychos in the community as well including "The Organisation ORG" CIG protagonist Meta staff! Yes got them fired with a formal legal letter with intention to sue from myself and a list of others in the community. The bad actors (entire framework) fell apart shortly after. All the bad actors began to interfight with each other taking each other down. The community has settled down completely since then. Please do not start that all up again over the very thing that started the meta bullying along with control freaks running ORG's aone example Montoya from TEST!!!!! Just stay away from this conversation until we have official CIG confirmation on Dynamic servermeshing specifically ok? PLEASE!
Oh hey, it's you. How is it going? Long time no read. I see you are still playing the know-it-all tech police. How is that going for you? Hurting servers physically. Quantum computing to the rescue. File system based server technology. Man, you really have no clue about anything at all and need people to know, dont you? 😅 How about not catastrophizing about thinks you dont understand for attention, hm?
@@SpaceTomato lol no I didn't say you did. I'm saying that it's now truly possible with new NVIDIA server technology and soon (about 5-7 years from now) to finally be fully realised quantum internet. This will allow for actual world wide close to 0 ping Dynamic servermeshing to be achieved! No more regional servers and true dynamic meshing instantaneously through NVIDIA's Hoppa 100 (by then probably 1000) dynamic GPU server technology replacing the current file system sever technology we currently have now. I'm covering this very soon myself. It's ground breaking news for Star Citizen mate! The actual realisation of 100+ solar systems and international dynamic servermeshing world wide simultaneously without lag or anything to hold the true vision of what Star Citizen was always meant to be finally! This will not end up like freelancer did now!!! Mind blown yet!? I've just finished a series on gaming in general and it's all bad news. BUT for Star Citizen it's the dawn of its true realisation now at the same time!! Star Citizen will set a completely new president in what true future gaming will be. A new standard that if any other developers are left by 2030 will have to start way behind the 8 ball on to catch up! Star Citizen will be truly in a league of its own as the new AAA standard in gaming. What I never thought was possible with current technology was right all along but with Hoppa 100 server GPU technology and soon quantum internet, Chris Roberts extremely ambitious journey will literally come to full fruition now!!! Please check out what I've covered over the years to see why I'm so excited now!! My predictions were 100% correct when talking about Dynamic servermeshing but with these technology breakthroughs finally, Dynamic servermeshing will become a very real reality finally in the years to come now!! Finally what I'd hoped would one day happen has happened right inline with the dates I set years ago now. I know there are detractors and haters, but what I predicted was also admitted by CIG at Citizencon a couple of years back as well themselves. Now with these two major technology road blocks being removed soon Dynamic servermeshing will be technically and practically achievable finally. This is ALL good news for the developments success! And this is why Star Citizen 1.0 and Static servermeshing 1 has been implemented. Keep in mind they are doing 3-4 Static servermeshing roleouts prior to finally launching Dynamic servermeshing already. This was already told to us years ago as part of the process for this tech. I've been quietly waiting for this to happen for 5 years now! I said 7-10 years and this will fall precisely inline with my industry based facts and proof in past. Personally I'm truly excited for the developments future for the first time ever! 2015 OG backer and I.T professional by the way. The haters don't want this game to succeed by the way. Be wary of who's in your comments section spreading hate and lies. Yet another reason why I moved away when I did as there was nothing to talk about until now. And holy hell what happened to Mike? He's so angry all the time now? Oh well like you'll I'll be bringing some love back to the development for now at least. And some really needed hope it seems. So many angry TH-camrs in the Star Citizen scene right now! Hopefully what I have to share might make them happier/optimistic again? Let's see yeah?
@@SETHthegodofchaos funny thing is kid, is that you're part of the finished with meta. What I'm talking about is so far beyond your squirrel brain that when it actually happens you'll be even more irrelevant than you already are. You and the skitziens are a faded memory. Don't try to resurface, it'll be the end of your time in the community this time. You're also lurking around TH-cam channels that you have no relevance in either, sad! How's your Aurora going? 🤣🤣🤣 You are not important kid.
yes and what happens when it brakes down ? this is the reason is not use in video games cause at one point it will brake and recovering will effect all entities related to the problem, so u will slow down allot of player , eve online has a time dilatation so shit will not hit the fan, SC is a online first person shooter it will not work , online first person shooters need stability first. Any way i wish them good luck and god speed.
There is meant to be refinance built into the system. If a server breaks, the replication layer takes over, if that breaks, the entity graph storage still exists. You would just experience a true crash from the game, at that point.
Star citizen is literally an animation waiting to become a game that doesn’t run like shit and that can be the 24 hour online universe serving tens of thousands of players flawlessly. Just because Chris Roberts writes a flowery letter and they are doing basic testing doesn’t mean it’s time for a new wave of content creation to hype up the game as something that it is currently not and will probably not be for still many many more years. The notion of having multiple servers for various parts of the game is as old as dirt so to be presented as if it was just thought of the other day is disingenuous. Remember these are content creators it is their business and their interest to hype up any bit of news they can out of star citizen
Server meshing is part of the stability fix that should help there. But there is also a missions team who is working on that! You can find more info about that here: th-cam.com/video/xMN2xLmvJzc/w-d-xo.html
They’re definitely making this technology to sell it. They’re not making the game because they’re gamers they’re making the game so they can sell the servers.😂
![[Full] 4 ต่อ 4 Celebrity EP.922 | 10 พ.ย. 67 | one31](http://i.ytimg.com/vi/zmV5HRIxIBI/mqdefault.jpg)
Forget server meshing, let's talk about that opening line. Yikes!
Oh also, OVERSTATED. Cannot be overstated*
Finally, A Major Update: th-cam.com/video/1XXKkLjK4Vc/w-d-xo.html
lol SeeeeIG
Listen, you always make great videos. But come'on. You're talking as if "server meshing" is some crazy new technology that RSI has has to create/invent to make this crazy new thing happen that's new to MMOs. But this is stupid. "server meshing" is as fundamental to MMOs as a UI. I remember the 2 second pauses between servers while running around in UO. This is NOT NEW and EXTREMELY unimpressive and embarrassing that this is where they are at after 12 years and a billion dollars.
@@Vikkoify I don't know man I tried REAL hard in this video to say THIS IS NOT NEW TECHNOLOGY OTHER PEOPLE HAVE DONE THINGS LIKE THIS, but it always gets missed. Where in the video am I saying it's new? I even put EVE and Dual Universe on the screen? I highlighted the one thing that was actually impressive about this implementation, that there IS no 2 second pause. I've talked to a lot of people about this, so I understand how unimpressive it might be to you, it's why I only tried to draw attention specifically to the unique parts.
What they've done is brilliant, maybe it's not the first of it's kind, perhaps it's not going to change the world, and it's definitely is not solving other games. But it's a big deal for this game, and it's an achievement that only a few in the industry can claim.
I'm excited for this tech, and from what I've seen it's not easy to make, so I'm gonna give em credit on it. The title of the video isn't "this will fundamentally change gaming", it's just Star Citizen that's moving forward, and that's a positive.
@@valeriansamborski5844 eh, I'm not thinking about Musk for it, more just the amount of time it took the company to develop their tech advantage into a pepper product, ya know?
What they are attempting is nothing like what you describe in anything but the most basic level. But hey, I'm just a Sysadmin of 30 years and gamer for longer.@@Vikkoify
In the new "Letter from the chairman" Benoit said "In the coming weeks and months, get ready for more Technical Preview tests with various mesh configurations: multiple game servers per solar system and seamless transitions without gates. We're talking about layouts where servers are dedicated to entire planets and moons, others focused solely on Landing Zones or other key locations, with plenty of higher player count experiments." Lets gooooooo.
If that ain't Dynamic Server Meshing idk what is. The hype is REAL
@@Fuar11dynamic would be able to change dynamically, on demand and analyzing by itself the demand. Static means the charge and size and the survers will stay the same no matter the loads on each servers or needs to aliviate one or the other.
@Fuar11 The hype is very real, but yeah, dynamic means servers will turn on or off depending on if people are in them or not. Even potentially actively split an area that says a big fleet battle is happening. In real time. Spinning up more or less servers depending on where the action is, but I suspect that is very far down the road.
that all being said, why it matters is running a server costs money. So if say, every planet has a server. And every star system is supposed to have an average of 3 planets. For the 100 star systems planned, that's 300 servers. Just for one region. Say NA, then you'd need 300 more for Europe. Asia etc.
You can see right away it's a huge cash sink if you can't actively turn servers on or off. Even if players aren't using them. Now multiply that if we do every landing zone having a sever and you got yourself a problem.
Make no mistake. Now is an exciting time to be in star citizen. But we are still a ways off from even a dozen star systems, but static server meshing will give us a lot. It is the last hurtle before they can develop the rest of the game in earnest. Since server meshing is so fundimental, CIG basically had to gut a whole game engine and reulbuild it to make it work for this type of networking. It makes sense that star citizen development has been slow for the last few years beyond sq42. They couldn't add more that they wouldn't have to just totally redo for the new tech. So from a 'running a company' perspective, why make it in the first place? But now that we have servermeshing, that means the flood gates are open.
@Fuar11 It definitely is in preparation for the dynamic mesh. Testing different scenarios will allow them to collect data on how servers perform under different loads, server startup time based on entity counts for different sized areas, and all sorts of metrics they can use to define the rules to dynamically alter the mesh.
Yeah that was amazing timing 😂
This tech could literally change all MMORPG experience, not only for SC...
The beauty of dynamic is: If 1000 players decide to meet at Lorville, a server will spool up, and take over Lorville, to reduce the strain on the overall experience.
I hope it works!
Great video man. Impressed with the quality given the catastrophic water damage!
Thanks! Community got us back on our feet!
Your videos are always the ones I share with those who ask about Star Citizen. You can take complex concepts and break them down into digestible chunks. We appreciate it!
I appreciate you watching! Glad you enjoy them, and thanks for sharing. :)
Tomato you are really on a different level in terms of SC quality videos. So we'll explained, great graphics, excellent editing and your smooth voice. It's pure enjoyment watching each and every video you put out! ❤️🙏
I appreciate you, it's an absolute pleasure ❤
I was just testing static server meshing last night. Very mixed results. When it wasn't recovering repeatedly over and over it was generally quite smooth. I saw my server change as I traveled. It. Blows my mind that we’re testing it already.
how you can test that? what you need to do?
@@deejagers716 there should be a drop-down in the launcher. You have to download the game again basically. It's on wave two testing. I can't remember what that means
Its just the start. Now that it's discovered and working, even if not seamless now it's gonna be very short to transition into dynamic server meshing. Then it will be even crazier.
The amount of bandwidth required will be insane tho 😂
Thanks for this work! I shared this with a buddy of mine after trying to gush about the server tech yesterday 🤣
Glad to help!
Awesome video. I am glad that they are testing meshing now. That is going to expand the possibilities of this game especially when it comes to dat load. I wonder if it will improve npc behavior.
1. Great video, as always!
2. Did you see the Chairman's Letter that just came out and the information about all the NEW upcoming server meshing tests that will see multiple servers per solar system? We're talking servers for each planet, moon, even landing zones and with even higher player counts than we just saw tested over the weekend.
Your video couldn't have had better timing, every player new and old needs to watch it!! 😊
WHAT?! REALLY?! omw to go find that letter...
Checking out the letter now, oh boy!
Making this happen without desync is seriously impressive. Making it happen robustly and reliably will be a miracle if they can pull it off. Assuming the meshing code is game engine agnostic, packable, configurable and licensable to other studios, that tech engineering alone could power the game for the rest of its life irrespective of players buying things. Think Physx ubiquitous for really large scale games. Total War where every unit is actually a player. GTA with hundreds or thousands of concurrent players on an instance or simply all at once in the same world. No more instancing anything by necessity, but only if it's intended.
Once again @spacetomato you have knocked it out of the park on this one. Great vid and explanation on how this all comes together. Now to see it in live!!
"Do you think we'll be able to build bases anywhere in the Star Citizen universe soon? Having multiple systems with player bases sounds amazing! I'm new to the game, and it has huge potential. The only thing that feels off is the slow pace (that train ride after death is rough!)."
Yes, base building will be a thing. As far as soon? Eh, relative to development (12 years so far) sure. But actually soon? Probably not for quite a while.
give it 15 years and another 650 million dollars and anything is possible.
"Do you think we'll be able to actually login in the finished game with just ships and basic combat?" No.
Another great video making the complex simple (relatively).
Appreciate the effort you've made explaining what's what regarding everything server meshing.
You speak with such confidence about something not even out yet, that hasn't even been truly tested, by a company who still don't have a complete game after a decade of scope creep and poor management. Its load balancing and database replication on more nodes for better performance, there summed up your entire video. Its not easy to do properly that can scale to high demand in any application let alone gaming where its never been used quite like this before. Only in gaming do you see this level of glazing from fanboys for things not even released yet. They have yet to deliver on anything outside of tinkering proof of concepts duct taped together.
wdym hasn't been truly tested ?
Players have already played in the game server meshed, how has this not been tested?
This is one of the best videos i have seen. There is a lot of info in this and it is explained perfectly. Great job man. I am going to share your video.
If they pull this off it will change the future of all mmo's in the future.
Or immagine battlefield game with 1000 players per battlefield, the roof is basically endless on this tech.
This also means that this tech is ghonna be a lifeline for the company as many wants this tech.
Letts hope, because this might be future.
Alot of people run their mouth about this game calling it "Scam Citizen" but mark my words this will be the most epic and expansive game in existence. It has been an honor to be a part of the alpha testing, helping the developers find bugs during development.
I've always really appreciated the vision CIG has. While the game is heavily meme'd on i still watch it and with the server meshing finally looking like it's going to be real. This will change gaming, it'll change everything. I'm hoping.
Commenting for the support and to say that once again, especially with this tough topic for some, you and Mrs. T knocked it out of the fuckin park!
Thank you Casa!
Amazing video as always, way to take a complex topic and break it down in a way eveyrone can understand!!
Having the replication layer (what the player connects to) mix and match the results from all servers to make it look like one big world is smart but also scary.
I think the point is: The servers calculate the physics and object and NPC locations etc, then sends the results to the replication server, which forwards the info to the player.
The danger is that now the replication server is the weak point instead. But infinitely better than what we had before.
Maybe they can even solve the overall load by sharding the replication server into one per system LOL.
Still gonna be rough if a lot of players share one replication server and all go to the same location though.
Oh and the persistence server will get a huge workout from all the reading and writing of data, constantly changing when players move things around.
But... It's already a huge step in the right direction anyway.
Oh and the scariest thing will be bullets and missiles crossing the boundaries. How on earth will that work reliably. We will see.
@@MyAmazingUsername possibly, some actions that need high fidelity like combat actions might have their simulation completely replicated, instead of just the state, similarly to how most client-server games simulate movement on both ends at a higher rate and then synchronize it at a slower rate
Got to love these free fly events when the game is not playable because of lag, I don't want to introduce the game to friends if their experience is a lag fiesta.
This sounds like memcached but hyper juiced. This is actually incredible massive leap forward. I hope they have a research paper they release on their tech. Incredible.
I've been real hesitant to get into this due to the cost of entry and the performance issues. If this turns out as well as we hope, I think I might finally take the plunge.
It may be a good place to check back in a year!
I'm curious where the next inevitable bottle neck will pop up when dynamic server meshing scales up. I could see the replication layer being the next to overload where they'd need to start meshing those layers as well.
I assume there is HA/DR on the Replicator? Would he awful if there was anything that went wrong on that for all the servers. Super excited about this design!
I just wanna say. Even with just static server meshing, you can do a lot. Splitting each planet to its own server would be an obvious first step. We could have upwards of 400 players in just one star system. And the thing is, even if we only do a server per star system. We could still have a handful of star systems easy.
dynamic server meshing is the holy grail. I suspect it will be a while. But this moment now is very reminiscent of when object containers happened. With just object containers, we got several planets and moons. Static server meshing is a magatidue more then object containers.
Personally, I think DSM is about a year away, tops. Also, once its in Im not sure how many people the shards can handle. Not for stability reasons, but think about the amenities we have. With wrecks and trash, things get a little crowded from time to time. Say you increase Stanton to 500 players (exaggeration a bit but bare with me), does Stanton have enough hangars, bunkers, cargo collection and delivery locations? Often times when doing bounties, I come across bounties from other players missions. Id love for them to increase the player cap, but I think we'll need a few more star systems before they drastically increase it to 1000 or more per shard.
@justicebrown1077 I mean, I hope you are right about the DSM timing, but I think you are over optimistic. They said they would have static meshing in 3 years. 6 years ago.
I suspect in terms of the player count, if they make each planet and moon a server and each major landing zone. It doesn't take a lot of code to just casually make some of the random buildings around the city now be a drop-off point for stuff. They can also just add more points of interest now. (Including hangers, drop off points and such)
a lot of the reason we don't have the Aaron halo yet is because of simply entity count since right now, one server manages the entire stanton system. But I guess that is the test we will be doing is seeing what these meshes can handle in terms of players and all the stuff they leave around
I'm thinking Ready Player One! This new server meshing in combination with game world meshing and we got the beginnings of "The Oasis" !!
God d@mn! I love your content, it's always cinematic ;-)
About Star Citizen, this will be the year o7
Glad you enjoy it! The next couple years are looking good for SC!
The timing of this is hilarious XD
I wonder what server meshing can do with something like 3D modelers such as blender maybe even splitting load so much that rendering may take hour or less for even large projects with enough servers.
However, for meshing servers itself, I was having my one thoughts on setting my own server and game system before seeing this video that took inspiration from zero down horizon Gaia and her subroutine AIs, for serves there will be prime servers and subordinate servers, if prime is to have dynamic animal life a subordinate server will be in charge of this while prime it self is which animals will move, basically doing all calculations for prime servers and you can add as many subordinate servers as needed.
Now combining my prime and subordinate servers with meshing servers may potentially go well together.
@0:45 How close is close for servers to pull this off? How close will each player need to be to each other IRL?
this year I'm seeing that everyday something new and cool is being added and coming to SC, development really is speeding up
CitizenCon 2024 demo? Fix it in post. ^^
Love your videos. You got me into watching streamer content for the game. :-)
Living in the future! Thanks for watching 🙏🏼
Possibly one of your best videos yet, ST. Your quality is inspiring, to say the least.
The replication layer is like the blockchain
Yooo Tomato, Asmongold just did a reaction video to this one!
"... but understanding how it works is not so difficult."
famous last words
I believe the 2 minute recovery was only on an Evocati test flight.
After a decade of backing star citizen (my wallet smacks me every time I pull it out) this has me over the top excited. The potential gameplay could be infinite.. full org battles with 1000s of players as a small example
What excites me is the players and their ability to create gameplay (ie: Daymar rally) This sim has always had the potential to revolutionize gaming and it seems I can finally see that peaking its head on the horizon
Shoutout to all the “star citizen is a scam” folks 👋
Your Outro tracks hit well. Do you have a published playlist?
You mention that other games have used server meshing. also that is not a pivotal time in game industry (mayeybe you say that to another video) howeverI was under the impression that indeed star citizen is potentially shaping the game industry exactly for this technology. do the oversell it ? or there is actually truth to that? like the other server meshing game would have end up in a bottleneck?
It's really specific to each implementation. What they are doing is pretty monumental because of the real time transitions in space, but something like a server transition in a jump gate is not very novel at this point.
Where SC is really shaping the industry is with dynamic server meshing which allows those real time transitions to happen at any point in space as server resize themselves as needed.
It's certainly a big deal, but until it can be applied to the succes of the game it's good to not go through the roof with hype lol
@@SpaceTomato Thank you!! ^_^
This is the "big bang" of the star citizen universe
Corrections (based on my understanding of how SM works):
2:15 - I personally don't like that statement personally as "The next generation of Loading-Screens" describes Object-Container Streaming MUCH MORE than actual Server-Meshing. In OCS you as a client constantly load/unload the world around you as you move through it which I do think is "the next generation of Loading-Screens". Server-Meshing however is more about splitting up the load on Servers from being all done on a single server to splitting it up between multiple servers (based on Location and/or Purpose). So, I would say, Server-Meshing is "the next generation of Simulation"
4:01 - The state of entities is not really replicated in other containers. If we JUST talk about Object-Container Streaming and not RepLayer or ServerMeshing, then Object-Containers are basically FOLDERS for the GameWorld. They structure the game-world so that you as a player only need to load the folders and in those folders the files that are necessary for what you are currently doing. The Server also only needs to have those Folders and Files in it's simulation that all the players are currently interacting with and nothing that is not needed for anybody. That is SS-OCS (ServerSide Object-Container Streaming)
5:20 - The state of every entity doesn't need to be replicated on other servers nearby because otherwise the players would not see what is happening in the other servers. Players get all their data from the RepLayer (they are no longer connected to the server at all) and the RepLayer knows what is happening to all entities in the shard (not just the once from specific Object-Containers).
The reason Replication on other servers nearby is needed is simply for the transitions between Servers. When one server needs to take authority over an entity it needs to be basically instant, otherwise we would have lag in player-movement, physics, ... if anything changes the Server. By already replicating all entities that are simulated by nearby servers the server now doesn't need to load those anymore (which is the most time-consuming part) but simply needs to start simulating it.
6:46 - The performance problems you have with your client aren't or at least have very little to do with the Server-Performance (unless SC still has some VERY BAD backend design). That is simply optimisation of the Rendering-Pipeline that still needs to happen. The bad physics and desync during an event like Daymar-Ralley is however often a result of bad Server-Performance which SM will/can help with
11:36 - From what I have heard (this is based on a podcast where one person said something he heard from some CIG Person): Pyro is currently too big for a single server to handle. There are too many entities in that system hence the reason why they can't test the entire system on a single server (so they only test parts on the Preview-Channel without SM). Unfortunately I can't say anymore which podcast I actually heard this so take this information (even if I am pretty sure I have heard it and that it was stated by somebody at CIG) with a grain of salt.
Thanks for the input! There are a lot of details in this tech!
If individual servers don't need replicated data then how will each server handle operations it has to perform that require cross-server information? For example, if server 1 has a heat seeking missile that is trying to hit something in server 2.
@@bobbycrosby9765 I expect you mean what I wrote in Point 3
Read it again. I didn't say it doesn't need to be replicated on multiple nearby servers which it obviously does for transition sake (as you said).
What I was trying to say with that point is that it is not necessary (like I heard so often) because if is it is not done clients wouldn't see the changes made in the other server. Clients have their own criteria that is replicated to them completely separate from what servers replicate so not replicating on nearby servers would obviously hurt transitions but not clients seeing what changes on the other servers
Even if star citizen doesnt make its goal, selling & licensing the tech out will make them 100 billion easy if they have a revenue share system. Allowing small studios to off shift thier success instead of leveraging gigantic budgets to build the tech themself. instead focusing on art & game loops and playability & storyline
Great video! I love your content.
Thank you!
Glad you enjoy it 🙂
Excellent summary, mr. Tomato! Good times ahead for SC technology wise
So i wonder, will server meshing increase the FPs in Hometowns and Spacestation and will it reduce the amount of crashing servers?
Damn it really took them almost a billion dollars to discover load balancers and database replication. Truly one of the innovators of all time, just like Tesla.
i like how eve online has this for 10 years
"It is the RAM to the entity graph's hard disc long storage drive"
I wonder how different this is to Server clusters- if not server meshing is server clusters but modified for Star Citizen and more expensive to run
This is it the moment we've all been waiting for
Does it mean we will finally see this game function properly? Insane.
Subs are so low… this channel is just amazing
The only way to mesh servers seamlessly is to keep hot backups online at all times whether used or not due to the time required to initialize a fresh server once a crash has happened assuming crash detection requires zero time. Best of luck in reducing the time required to recover from a 30k quickly.
🤓
@@kekkopf69 go play with rocks, adults are talking..
@@kekkopf69 Your assertion that server meshing will save money is false. Server meshing requires hot backups to be constantly available, or a finite amount of time will be needed to bring a new server online. CiG would be better off just using Redis for their server mesh, but this, too, would require some time when a new server is brought online. Server meshing will increase the cost for CiG, assuming players don't want to wait.
Or not.
Another option is to mesh with some overlap, allowing "nearby" servers to take over the charge (with some degradation?) while a new one in instantiated to replaced the broken one.
Like some kind of RAID 5, but for servers.
@@eldarionmarchombre4568 Ok. You assume the replacement server has few objects to load to be meshed. It would have zero objects to load until a player touches it. However, it will have to load objects to get meshed, and getting meshed will always require some time. Assuming we are talking about Cloud Servers... a replacement server would have to get provisioned and then started before it can get meshed; I am assuming the image for the replacement server has the current game installed and does not need to get updated before it can run, I am also assuming the OS does not require any updates. OS updates will be necessary, as will game updates, and this is what will eat time, assuming you have no hot backups. Players will not want to wait 5 minutes for a 30k recovery, and this supports my initial thesis because hot backups require server nodes to be made available in advance of being used, and doing so eats money, assuming your Cloud Provider will allow you to offline a server node without being billed however I am aware of one Cloud Provider who charges money for server nodes regardless of their use.
i expect them to start with something basic with server meshing to then ramp it up overtime
i mean we all know this game wont survive if the server performance stays shit.. we are only here because we are confident that these issues will be fixed
Good video tomato!!!
Thanks!
Asmongold actually reacted to your video. Congrats.
Awesome to see! Thank you!
@@SpaceTomato also your videos are not getting recommended even when the similar subjects are searched.
Top tier content, Tomato. Thank you much. ❤
Thank you for watching!
You see a lot of critics saying that the concept of server meshing exists in other games already. WoW has server phasing, ESO uses layering, SWTOR uses instances. Different executions of the same idea.
People who think this is the same thing as CIG taking 3000 concurrent players and serializing the client session to pass between servers seamlessly, do not know what Server Meshing is, or why this is big. No company in history has ever done that which is exactly why you see developers on the outside claim that CIG's dreams are unrealistic. Because they haven't done it.
Dual Universe did it. SpatialOS did it. This was new in 2015 but not anymore.
@@SETHthegodofchaos Read above.
@@Tazytots Well, maybe you should read mine? Did I mention WoW, ESO, SWTOR? No, I did not. These are clearly wrong statements made by people who have no idea what they are talking about. But if you look at the implementation of DU and SpatialOS, its pretty much the same as what CIG implement. If you dont see the similarity, then quite frankly are doing the same thing the critics are doing.
@@SETHthegodofchaos DU is not the same thing as Dynamic Server Meshing. They're categorically not the same thing. DU is the closest a GAME has ever gotten to what SC wants to reach, but no. They aren't the same.
I won't touch SpatialOS. It's effectively failed to produce any actual successful projects.
@@Tazytots Hm ok. Maybe I missed something then. How does the DU tech differ? They split the game world as well and put them the sections dynamically on different servers, no? They have systems in place to optimize loading and networking too. Like CIG has with OCS and PES.
And that SpatialOS doesnt have any games doesnt worry you? They too have a service inbetween game servers and clients, just like the Replication Layer. It too has a in-memory cache and database connected. It also uses gRPC for communication. It dynamically divides the game world into sections and hands of authority over them to multiple game servers. Where is the different there that will make CIG's version work (better)?
CIG is nothing like Tesla.
CIG is more like Scientology, but instead of paying money for the salvation that never comes, you're paying money for pretend spaceships in a broken and featureless "game" that will likely never fully release
Awesome content!
Can't wait to play this game in 2077
Thanks Tomato! Great info :)
Space tomatoe Asmongold reacted to a compilation of videos with you in it!
Just to clarify, is it going to be that servers are handling both players AND environments separately? So you could have a TON of people in game all be at the same place at the same time, but also have 50% of people in one environment "shooting cannons" at the other environment seamlessly? If im right that sounds pretty cool
Based upon the video and my rough knowledge of this subject, they are doing positional meshing. So no, you couldn't have a TON of people in the same place, because since they're in the same place at the same time they're all on the same server.
I think people are massively overselling the current state of this - don't get me wrong, it's still impressive! It's like most MMOs but no loading screen between "zones". Which is super cool because you can have a seamless world. But it doesn't really allow you to have more people to be in the same place like so many are letting on.
@@bobbycrosby9765 This is true for static server meshing, which is what CIG showed us at cit con, but the next step of this tech will be "dynamic server meshing", where the servers will be automatically alocated to different places based on the amount of entities in said places. Let's say for example a ton of players decide to hang out in the same place, with dynamic server meshing, there will basically be multiple servers assigned to that location to reduce stress. But the most important step was this one I'm pretty sure, the transition from static server meshing to dynamic will take some time and effort but not nearly as much.
@@PikChouk it's still not necessarily going to help. And it may actually hurt.
If a bunch of people in 1 spot interact with 1 thing, that interaction is still going to have to go through the server with that 1 thing, even if all those players are on different servers.
Then you're going to have to replicate data across those multiple servers because they're close enough to eachother that players can interact with them.
It's not a silver bullet. It might help in some scenarios. But there definitely will be others where it all falls down. E.g. all the players shooting a missile at 1 entity. They will all converge on the 1 entity, and they will necessarily end up all on the same server, and all the surrounding servers may need to be informed of the movements for any processing they do.
It's not anywhere near as much of a sure thing as people are making it out to be.
@@bobbycrosby9765 but a server could perfectly handle this scenario tho, the "server with that 1 thing" is the replication layer, what it does is that if a server crashes, it is automatically replaced by another one preventing players from totally crashing, in effect if a server still crashes which will be unlikely, you'll just a wait a few seconds and the servers will get replaced so you can get back into the action. At the moment, servers can handle the load of the entire stanton system, even if 1000 missiles are shot at 1 entity, if the server is limited to a very small space around this entity, the load will be incredibly lighter compared to what it can handle right now.
Awesome content.
Great Video love it🙏🏽
we know, we heard about it like..7 years ago? something like that
Can confirm, it's in the video! 😊
I feel like a shareholder instead of a gamer at this point, they are pushing the tech and innovating the industry. At this point it would be better for me if they were to convert our ships to equivalent shares and went public.
Server meshing aint that special Ultima Online had a version of server meshing in 1997. Yes dynamic server meshing would be amazing but there is no set date for this technology. 4.0 is supposed to come out later this year but that only has static server meshing. So really there is all this hype for a technology thats still just a proof on concept, but hey maybe if we give star citizen another billion dollars they can give us another presentation in the next 5 years.
Dynamic server meshing is right around the corner... not this year....... probably not next year either, but 2026 is gonna be a game changer.
I bet they don't even know when it might be. But I hope it's fairly soon!
when starft that server meshing?
Testing has started and currently we are seeing 700+ players in a single star system.
perhaps a new industry norm
so they have basically changed the multiplayer from being monolithic architecture to microservices.. that's awesome
That's how I understand it. The actual professionals explained it better in the podcast I noted on this video, but micro services are the key!
Wow they built all of this amazing tech and didn't consider server cold starts? Sounds too stupid to be true.... interested to know why not keep a warm pool of servers ready to replace a down one or maybe the latency is due to the syncing time it's taking with the "Replication Layer" to get to the current world state.
I want to know so much mooooree about this.
Thank you ^_^
Thank you for simplifying an often poorly explained concept as well as all the phases of its development!
Not sure if I will survive to see a tech where we can play games on single digit ms in a game with everyone around the world
When an MMO finally figures out instancing.
Lol this game has had instancing since 2013. This is more along the lines of what Spatial OS or EVE does.
You made a virtual network.😊😅
SERVERY MEESSSHHHHIIINNNGGGG *woot woot*
This is still Static servermeshing without the hope of ACTUAL Dynamic servermeshing for quite a few more years. I've covered this at such a granular level over an entire series of videos over 3 years ago already. The server technology is the main thing holding this entire development back. Current server technology isn't quite literally designed to handle ramp up and ramp down in such concurrent cycles. It will literally kill servers physically. Current server technology is only file based technology and not virtualisation based like quantum computing is. We have to wait for quantum computing servers to become the norm still. As I said I've covered this granularly over the years already. The overhead costs of current server technology also would lead to prohibitevly expensive running costs and failure rates. Just like you're seeing now with 5 min down to 2 min ramp up and ramp down issues. That's a tiny sample size that is still firmly based on static meshing technology otherwise know as file system based server technology. That's a best possible case for current hardware technology.
It seems they finally have the net code 2/3 sorted out with the replication layering code but there is a huge way to go still. Don't get your hopes up at all. This is not going to be a real solution for years to come. CIG at Citizencon 2 years ago also publically admitted that as well. Over a year after I already proved that with industry evidence and so on. Spectrum was alive with comments and questions for that year or so prior asking the same thing based on what I had already granularly proved as well. Hence why CIG finally came clean about it!
We are no where near REAL Dynamic servermeshing. Please don't get peoples hopes up. We don't need another community war breaking out again! Otherwise wonderful graphics in your presentation and I can see your hearts in the right place.
But please let this sleeping dog conversation lay. This is something people get very "excitable" about as if you're an OG backer like myself and a massive amount of other people are, it can seriously set off some major arguments and resentment from the community. Trust me you don't want that. Been there done that and proved my self right 100% and took down the psychos in the community as well including "The Organisation ORG" CIG protagonist Meta staff! Yes got them fired with a formal legal letter with intention to sue from myself and a list of others in the community. The bad actors (entire framework) fell apart shortly after. All the bad actors began to interfight with each other taking each other down. The community has settled down completely since then.
Please do not start that all up again over the very thing that started the meta bullying along with control freaks running ORG's aone example Montoya from TEST!!!!!
Just stay away from this conversation until we have official CIG confirmation on Dynamic servermeshing specifically ok?
PLEASE!
Oh hey, it's you. How is it going? Long time no read.
I see you are still playing the know-it-all tech police. How is that going for you?
Hurting servers physically. Quantum computing to the rescue. File system based server technology. Man, you really have no clue about anything at all and need people to know, dont you? 😅
How about not catastrophizing about thinks you dont understand for attention, hm?
My friend what? Nowhere in here did I say Dynamic server meshing is near.
@SpaceTomato ignore him. He is just looking for attention. Nothing he says makes any sense.
@@SpaceTomato lol no I didn't say you did. I'm saying that it's now truly possible with new NVIDIA server technology and soon (about 5-7 years from now) to finally be fully realised quantum internet. This will allow for actual world wide close to 0 ping Dynamic servermeshing to be achieved! No more regional servers and true dynamic meshing instantaneously through NVIDIA's Hoppa 100 (by then probably 1000) dynamic GPU server technology replacing the current file system sever technology we currently have now. I'm covering this very soon myself. It's ground breaking news for Star Citizen mate! The actual realisation of 100+ solar systems and international dynamic servermeshing world wide simultaneously without lag or anything to hold the true vision of what Star Citizen was always meant to be finally! This will not end up like freelancer did now!!! Mind blown yet!?
I've just finished a series on gaming in general and it's all bad news. BUT for Star Citizen it's the dawn of its true realisation now at the same time!! Star Citizen will set a completely new president in what true future gaming will be. A new standard that if any other developers are left by 2030 will have to start way behind the 8 ball on to catch up! Star Citizen will be truly in a league of its own as the new AAA standard in gaming. What I never thought was possible with current technology was right all along but with Hoppa 100 server GPU technology and soon quantum internet, Chris Roberts extremely ambitious journey will literally come to full fruition now!!! Please check out what I've covered over the years to see why I'm so excited now!! My predictions were 100% correct when talking about Dynamic servermeshing but with these technology breakthroughs finally, Dynamic servermeshing will become a very real reality finally in the years to come now!! Finally what I'd hoped would one day happen has happened right inline with the dates I set years ago now. I know there are detractors and haters, but what I predicted was also admitted by CIG at Citizencon a couple of years back as well themselves. Now with these two major technology road blocks being removed soon Dynamic servermeshing will be technically and practically achievable finally. This is ALL good news for the developments success! And this is why Star Citizen 1.0 and Static servermeshing 1 has been implemented. Keep in mind they are doing 3-4 Static servermeshing roleouts prior to finally launching Dynamic servermeshing already. This was already told to us years ago as part of the process for this tech. I've been quietly waiting for this to happen for 5 years now! I said 7-10 years and this will fall precisely inline with my industry based facts and proof in past. Personally I'm truly excited for the developments future for the first time ever! 2015 OG backer and I.T professional by the way. The haters don't want this game to succeed by the way. Be wary of who's in your comments section spreading hate and lies. Yet another reason why I moved away when I did as there was nothing to talk about until now. And holy hell what happened to Mike? He's so angry all the time now? Oh well like you'll I'll be bringing some love back to the development for now at least. And some really needed hope it seems. So many angry TH-camrs in the Star Citizen scene right now! Hopefully what I have to share might make them happier/optimistic again? Let's see yeah?
@@SETHthegodofchaos funny thing is kid, is that you're part of the finished with meta. What I'm talking about is so far beyond your squirrel brain that when it actually happens you'll be even more irrelevant than you already are. You and the skitziens are a faded memory. Don't try to resurface, it'll be the end of your time in the community this time. You're also lurking around TH-cam channels that you have no relevance in either, sad! How's your Aurora going? 🤣🤣🤣 You are not important kid.
As always a great video, keep it up man!
just so you know homie. tomatoes are my favorite food lol. thanks for the video
don't worry, its only 2 more years away™
yes and what happens when it brakes down ? this is the reason is not use in video games cause at one point it will brake and recovering will effect all entities related to the problem, so u will slow down allot of player , eve online has a time dilatation so shit will not hit the fan, SC is a online first person shooter it will not work , online first person shooters need stability first. Any way i wish them good luck and god speed.
There is meant to be refinance built into the system. If a server breaks, the replication layer takes over, if that breaks, the entity graph storage still exists. You would just experience a true crash from the game, at that point.
Star citizen is literally an animation waiting to become a game that doesn’t run like shit and that can be the 24 hour online universe serving tens of thousands of players flawlessly. Just because Chris Roberts writes a flowery letter and they are doing basic testing doesn’t mean it’s time for a new wave of content creation to hype up the game as something that it is currently not and will probably not be for still many many more years. The notion of having multiple servers for various parts of the game is as old as dirt so to be presented as if it was just thought of the other day is disingenuous. Remember these are content creators it is their business and their interest to hype up any bit of news they can out of star citizen
You just missed the part where I said server meshing exists in other games?
Running and running on a single server… trying to run
chairmans letter so chonky.
👍
how about fixing the fact that 95% of missions don't work
Server meshing is part of the stability fix that should help there. But there is also a missions team who is working on that! You can find more info about that here: th-cam.com/video/xMN2xLmvJzc/w-d-xo.html
call me stupid
but didnt blizzard already perfected this kind of tech for WoW ?
They’re definitely making this technology to sell it. They’re not making the game because they’re gamers they’re making the game so they can sell the servers.😂
They don't own the servers.
can't wait to see this actually implemented in game... in 2050.
It's in the testing branch...
asmongolds video is abosultely swamping yours, give him an invoice
???