SaltEMike Reacts to Server Meshing PROGRESS! |

The Video - • Server Meshing PROGRESS!
Check out my main channel - / @saltemike
Check out my clips channel -
/ @saltemikeclips
Check Out My Twitch Stream to see this live: / saltemike

[@Swatmat]

1042 is such a hero for the performance testing he does, no one has run round lorville as much as him

[@johnzoidberg1160]

@tenpoundfortytwo‬ is a hero. Him and his discord are some of the most knowledgeable and helpful computer builders and testers i've ever seen. Dude is legit one of the purest gems of the Star Citizen community.

[@HarperLylia]

I'm actually in the camp of having smaller server sizes in the beginning and bumping up as the tech improves. I just reeeeally want a (somewhat) stable 4.0 launch.

[@absenth2007]

agreed, more than anything i want a smooth, responsive experience.

[@jamesalexander1647]

He'll even if they allowed an option at the main menu for either Stanton/ pyro

[@jebidyah]

sure would make testing a lot more efficient and reliable wouldn't it....hard to properly test a game when 80% of the time you are just dealing with server and optimization jank....

[@Bronwyn031]

@@jamesalexander1647 I'm with you. Enter the quantum gate and literally the game logs me out of Stanton and logs me into Pyro. I'm totally fine with that. These server boundaries sound like they are behaving awfully. Players suddenly being dumped into open space and ships disappearing. The game is already buggy as heck, imagine losing all of your cargo or mining haul just from quantum'ng to your home base.

[@CitizenScott]

Why? Genuine question.

[@bwdufferco]

I think 200 players total (100 Stanton, 100 Pyro) would be a failure if the server performance is the same or barely better, but if they deliver 20-30 server fps and it's 200 players total I wouldn't consider that a failure personally.

[@TheInvisibleOne]

Server fps was over 20 with 350 players. It will be even better with 200 players per system.

[@123TheCloop]

@@TheInvisibleOne server fps means nothing........ what matters is client to server latency and tick rate.

[@robertchandler2063]

Sever performance has been the excuse since 2018 for why AI doesn’t function and several other hundred bugs (the old works on my computer) nonsense needs if sever performance is above 30fps. Then they can find all the real problems.

[@Blkstang82]

​@@123TheClooptick rate is server fps Server fps is just what people have come to commonly call what is really tick rate

[@TheInvisibleOne]

@@123TheCloop correct. The interaction delay was also okay (between 1 and 5 seconds delay)

[@OddJobEntertainment]

8:10 Jared gets to pick the config for Mike's server confirmed.

[@thediadect8914]

It went so well that we will not have server meshing for another year :) minimum

[@natewagstaff2516]

Not so sure about that. Maybe. But maybe not.

[@mariopenulli1395]

Its just two years away, like always.

[@thecancermen245]

I anticipate a botched bare minimum version on PTU by Christmas.

[@duramirez]

@@thecancermen245 CHRIStmas hehehehehe

[@SETHthegodofchaos]

@@mariopenulli1395 It was always two years away, now thanks to the tests its always one year away. I see this as an absolute win 😆

[@justinhausaman2867]

I think it is painfully obvious that the entire legacy ATC and Transit need to be totally binned and started from scratch to work in a meshed environment, along with the entire backend database functionality need to all be transitioned to microservices to function properly when scaled

[@robertchandler2063]

Luckily that’s what they are doing basically. 4.0 is supposed to fix transit with meshing. That’s why you can safely ignore any talk of transit errors because it’s “supposed to be fixed in 4.0” this test was running 3.24 tech and code

[@aguy446]

And this is why there isn't more stuff in the game because they didn't want to rework everything. Makes more sense that they decided to focus on squadron

[@robertchandler2063]

@@aguy446 yeah the devs in mikes chat have saiid that for years

[@comniverse_0122]

That is the plan

[@AlexanderLeister]

Year 13 of development and they still struggle to create the main feature of this game to even make it a real mmo. I'm flabbergasted too, Mike. Let's hope they pick up the pace and solve it or this game will take another 13 years to be even in beta.

[@thecancermen245]

And just a couple months ago they added more shit to rework on top of already messy systems, i.e ATC and transit, instead of reworking them with the new architecture in mind. This is what happens when the marketing team, a.k.a release gods, call the shots instead of the tech departments.

[@senn4237]

I've been trying really hard not to be so pessimistic lately with SC. But it is really starting to look to me that they are so far in tech debt it is going to be impossible to get this to a release state within 10 more years or before all the backers quit and stop spending money.

[@TheInvisibleOne]

I have some free time and nothing to do atm so I'm gonna argue with you lol 1. 13 years of development is technically correct but also not correct. The collection of ideas and the creation of rough concepts started in 2011, crowdfunding and actual coding started in 2012 but development kind of started again in 2014/2015 with the introduction of procedurally generated planets (and other stuff). You probably know that and just wanted to vent with your comment. 2. Even if it's been 13 years of development, it does not mean "finished product has to exist". It's an expectation (rightfully so) but those two statements are not connected. 3. 13 years of development does not mean that they can develop server meshing. Again: time spent developing does not equal "finished product". Server Meshing is an extremely complicated idea that takes lots of trial and error to achieve - as we can tell. As you probably also know: server meshing has been worked towards for at least 6 years now. But they have been officially working on the actual server meshing tech for the last two to three years now. So it's been 3 years of trial and error. You can be angry at that and anyone else can and that's fine. But just don't state "X years of development time = finished product / feature". It's simply wrong. (I'm not ignoring all the bad stuff CIG has and is still pulling off like fomo tactics, management and marketing problems and so on. CIG should be criticized.) Alright I'm done. Have a good day

[@mikelelele7112]

​@@TheInvisibleOne But who even said "X years of development time = finished product / feature"? And he said 13 years of development, which is technically correct, but even ~10 years is bad enough, considering the state this game is in imo.

[@AlexanderLeister]

@@TheInvisibleOne maybe they shouldn't have wasted time and money on SQ42, expensive office chairs and artwork. Roberts' management is a disgrace.

[@ScottZupek]

They aren't skilled enough and there aren't enough compute resources to make their vision a reality. There is a reason it's NEVER been done.

[@anthonystrickland7049]

Yeah, I'm betting they're firing waaaaaay more data than absolutely necessary over the network. I'm tempted to try joining a testing run and monitoring data traffic to see just how much of a hose they have blasting up and down the pipeline.

[@alexpetrov8871]

@@anthonystrickland7049 you can see amount of client in/out traffic in r_displayinfo in real time, it is not that much ~0.2MBp/s

[@SETHthegodofchaos]

Yeah, I am afraid there might be something to this. We had basic Static Meshing since the early 2000 with games like Second Life. One has to wonder why nobody has done a dynamic version it in the last 20 years. Besides, the cloud has existed now for over 10 years too. Nowdays we have lots of insights into how to create highly scalable distributed systems. That was kind of the cornerstone that pushed CIG toward this type of implementation back in 2015. And yet, it still took 10 years to get to a static version (albeit setup in a way to also work dynamically, but nevertheless thats a lot of time). Well, to answer the first question: There have been dynamic implementations. SpatialOS might have been the first, besides Dual Universe. There have been plenty since then, even for Minecraft. Does it solve the issues? Kinda, but not really. If 10 servers are going to simulate sections of a small area, then those servers have to exchange pretty much all data with each other so that they are ready. Especially in a game with fast moving objects. The costs are going to be insane. And the player clients will have to utilize insane culling to perform well. Enjoy being able to only view max 100m into the distance and objects fading in and out. Great stuff! And those are the reason why we still have instancing in most MMOs since the 2000s. They are cheaper and you can control the player experience much better than a free-for-all universe. Therefore, it is a solved problem: You dont do it!

[@MikeMcCarty42]

Start at 200, with multiple systems. Go from there. And the transit system and ATC need full reworks.

[@1aatlas]

Theyre literally being reworked right now for 4.0 along with the mission system refactor, CIG has stated this many times i believe.

[@MikeMcCarty42]

@@1aatlas I understand that. Just saying that it will go a ways to making things better for players.

[@jenathan76]

From everything I've seen and heard, this was a stress test. It's supposed to break and be terrible for the testers, because it is pushing the limits to see what settings work best, where the stress points are, and what causes the issues, so they can fix/adjust them. From the sound of it, they both identified some issues that need to be fixed, and the RMQ solution worked as well as they hoped it would. That means the test was 100% successful, because the goals were met. Achieving smooth gameplay is not always the goal of testing, and after reading the comments here, it seems to me like a whole lot of people don't understand that.

[@XeroJin84]

No. That’s not what a stress test is for. Stress test is to test the STABILITY of something. Stress testing doesn’t mean it’s supposed to break - it’s not supposed to break. You also don’t stress test until the function is working; it’s one of the last tests that’s done. This is straight up functional testing. It’s basic QA you do in the beginning to determine if it works per the requirements.

[@VerusKrutch]

Delays were at least 15 seconds for me. I was at lorville as well and magically got swallowed into a train. It was off the rails most of the time and the door kept opening randomly to where after a few minutes i just walked out of it. I landed in the middle of the city somewhere and then had a black screen and appeared in the spaceport. I was able to call a ship and then call for a gateway and leave. I didn't exit the doors until they were opened for a bit to be safe. The 15-30 second delays continued until the server finally had a 30k.

[@sirbonobo3907]

Sm in 2030

[@mariopenulli1395]

If we are very lucky. Imagine if you said that on Reddit in 2015.

[@MagnaRads]

This game is so far from complete its scary

[@commanderjericho]

I mean it's really not that scary. Play when it's fun, and if the company goes under oh well.

[@senn4237]

@@commanderjericho scary, disappointing, aggravating, sad, pathetic, miserable, despairing, etc. I don't think the word he used is the point.

[@MagnaRads]

@commanderjericho sure you can be positive but their event right now runs at about 5 fps with just 100 player servers total dumpster fire

[@MagnaRads]

@@senn4237 💯

[@mastrtonberry2]

​@@commanderjericho This is my attitude. I've gotten a lot of hours of entertainment from the game so far

[@Jeremy_Walker]

The whole premise of a huge world with super high fidelity and a load of players with complex objects just seems dubious at face value.

[@1aatlas]

The trouble is none of those things have ever been specifically quantified by cig, so exactly how big/how many players that is intended to be is left open to interpretation.

[@SETHthegodofchaos]

yeah, exactly. as if CIG were the first ones to have that idea... that idea is as old as the first video game. the hubris to think they are the ones to pull it off.

[@VFW-Mayer]

When you start 4 servers. 1 will be the habs. 2 will be the tram. 3 will be the space port. 4th will be the planet space. As people get into space, the servers must be able to expand. After 25% of players are going for space. 1 will be the habs, tram, and spaceport. 2 will be the planet space. 3 will be the planets moons. 4 will be the solar system space with other planets able to load. The servers must be able to expand and contract based on population they can handle. If 1 server can handle 100 clients, then when 100 clients are in the habs, you must have the habs be 1 server. When they make 1 server a planet and everything near that planet. They might have 400 people trying to log into the habs on the planet. If that 1 server can only handle 100 players. Then the server will freeze and data will be lost as we saw happening here.

[@reamoinmcdonachadh9519]

Quantum Travel, and Mission Completion are things that need to keep going as you cross a server line. AND the crossing of servers or shards needs to be invisible, unnoticeable.

[@1aatlas]

It will never be totally seamless.. but even now the delay crossing the server line was only a fraction of a second hiccup, it will get even more efficient as time goes on.

[@SETHthegodofchaos]

@@1aatlas I dont know. It was pretty seamless already in some of the tests I have seen. With some additional smoothing out in the future, I think on the client it can be very seamless.

[@SETHthegodofchaos]

Crossing servers needs to be seamless indeed. But crossing shards will never be seamless because you cant switch to another shard while playing. You cant cross shards in the first place. You need to logout and join another shard from the main menu. They are separate game worlds, star citizen universes. Each with their own Stanton, Pyro, ect and each shard conisting of many game servers.

[@MrSmith123123]

I was in the 100 and 500 and 1000 person tests. What helped me the most was getting my bnd point out to the LEO station.. then I could fly around and in a different shard than the city shards.. it worked but still overall a poor experience.

[@odcon]

I think they need to try setting a higher limit, but letting people in gradually initially. Use queueing for login or something. When they’re doing these tests everyone is trying to join into the few spawn points at the same time. Unfortunately even if that worked it would be a matter of time before enough players came together in a meshed environment to crash a DGS though.

[@tomato6460]

Coming up to the Nov sale. Coincidentally signals coming out of CIG are that everything is going well and major progress is right around the corner.

[@JohnCr0w]

Weird...

[@UNBE4R4BLE]

Mike adds the salt to the analysis and the crowd find it delectable Hallmark of a good SC influencer, for better or worse

[@nebulajumper6216]

6:50 Mike, I think that they had an algorythme that distributed server power depending on how positive you are towards the game. So you got a few minutes delayes he .. ;)

[@JJS563]

We are still 5-6 years out from a 1.0 release huh

[@ajkulac9895]

10 years would be great at this point

[@Bronwyn031]

I'll say it once again. I truly do not believe ANY landing zone currently in Star Citizen can support hundreds of players logged in at once. We don't have enough habs, store vendors, ASOP terminals, hospital beds, et cetra. There would be queues everywhere you went. At elevators, on trams, getting ATC clearances to land and depart. It would be a nightmare. CIG would have to implement players caps for the landing zones just to keep the player experience tolerable. Meaning you might fly to a LZ and have to hover idly by till clearance could be granted for you to land. Old Levesky was like this. 200 players for Pyro and 200 players for Stanton seems doable. With hard player limits on the number of active players in each system. meaning you may have to wait at the Quantum Gate till you are granted access to enter Pyro or Stanton. And I also think every main planet should have more than one landing zone.

[@nydabeats]

I was encouraged to see 1000 players on server and the server fps was 30, that's huge. Now the interaction delay is game breaking of course and makes the server fps irrelevant until they can fix it but I'm surprised it already works this good.

[@yonoqueriacuenta]

It is just not working lol having 1k people connected means nothing if they become irresponsive

[@nydabeats]

​@@yonoqueriacuenta 1k people connected. that's a win in it self, especially when your testing tech that's never been done before.🤓

[@yonoqueriacuenta]

@@nydabeats I'm sorry, no it is not any win, 'testing tech that's never done before' by them, because this has been done before and it is working in other games, sorry to pop your bubble.

[@danielcojocariu8438]

@@yonoqueriacuenta it's not a win for you bcs you can't see it as a backer

[@yonoqueriacuenta]

@@danielcojocariu8438 want to bet who of us backed into this project first?

[@UmmerFarooq-wx4yo]

The train system turns into a massive white elephant designed for real life crash disasters. They should change it to individual taxis, so you could go anywhere in the city or even to space for where your large space ships are parked or, while upto small truck sized ships parked along side your tower. Whereever you like. Out side your home. Or apartment.

[@VonSpud]

Can only hope CIG devs get control of this Server meshing golden goose.

[@romanwiller2180]

Hopefully the next test will focus more on getting consistent results, and then work on improving performance. With luck we could see some good progress relatively quickly

[@Lestat070707]

Testing to see what the hardware can cope with based on software work upgrades/arounds. Then be able to work out optimal player per server until hardware catches up. Don't loose hope.

[@Maverrick2140]

lol .. that guy in the white astronaut suit was probably me .. i was stuck in that elevator and got out after 2 minutes .. then left the elevator and later got dropped under the map by the train rotating on arrival and hitting me in the head ^^ i was watching different streams throughout and i also saw that some had terrible outcomes and others had a coparatively smooth experience if you ignore the 30 second delay for any interactions. i myself did get to space once and then the server crashed and i fell through the floor of a station elevator .. a later iteration with 6 servers worked and i got to a jumppoint where people were grouping up but then that one also crashed ^^ .. but CiG said they had good results and identified multiple new bottlenecks they now want to address for the next test.

[@c-shepard]

I knew it was going to be a disaster. Current server meshing doesn't work and the number one issue holding back the game.

[@icewyrm.]

From the sounds of it, a lot of the delays and interface issues are related to the udp message queue that so much of the game seems to rely on in the background for communicating changes in the game world. I recall one of the devs spoke about that process in a recent interview. I wonder if congestion in that queue is a problem separate from those server meshing is trying to solve?

[@Goyahkla100]

It just showed how it shouldn't work!! But yet, some youtubers are sooo positive that it is unreal. That's when I lose Interest....

[@widt94]

The hectic thing I experienced was when you click on a interaction or try to interact with something the ping will shoot up into the 1000-5000ms which was really bad but I dont know if that might be another network issue which might be causing it. Alternatively it could be because of the amount of network traffic

[@ammax9155]

on the 3DGS 500 i had 1-3min loading time and the langer test 15-20min, total experience was not great there where multiple DGS clusters and some worked very well and other hell on earth

[@saberkiller6781]

Gotta keep saltemike salty lol

[@Silverhawk-u2f]

Going from 100 players per servers to 350 in 2 star systems is still almost double of what we have now, so that would be pretty acceptable IMO. And big bonus if we have constant 30FPS servers as well. Edit: math is hard

[@Charlouf_]

for those thinking 4.0 will be 100 stanton + 100pyro, have absolutly did not understood the point of SM.

[@SpearHead1011]

@saltemike…. Don’t cry this year, they will try to draw you in again!

[@Lestat070707]

Less players per server costs more. But if game can run smooth on 100 players per shard. Then it's a win until hardware can catch up. Keep hope up peeps.

[@Vioblight]

I see these thumbnails and just roll my eyes… “progress”

[@project.jericho]

CIG's optic fetish extends to their hiring practices. It's been years since they've hired due to merit and creativity. They hire to check the correct ideological mandates and it is painfully apparent - particularly at locations like Bee Cave.

[@zeyosin1771]

It worked so well that they let us know, that we wouldn't see much change on our end.

[@falcon758]

I would be more than happy with 200 player shards that are smooth and performant. AI needs to always spawn and work properly. I do not actually care about any of the other features coming. All we ACTUALLY NEED is working smooth servers with high server FPS

[@kennethjensen730]

"went well" means 5+ years until implement. Atleast.

[@iamthesentinel584]

I love tenpoundfortytwo, it's good to see your thoughts on his videos.

[@jebidyah]

Defining what an unequivocal failure is by any measure, that is exactly what CiG will inevitability deliver........without fail, every time, since 2012..... Yet here we all are still....

[@thecancermen245]

For me the success/failure line is drawn at whether we get multiple DGSes per system regardless of playercount. Even a 100/100 player split but with 2/3 servers per system locked at 30 tick would be a win cause it shows the base is working and we can finally play with decent NPCs.

[@anthonystrickland7049]

So, I've not been keeping track of testing. Have they done any testing with current capped server levels? Can meshing even handle maintaining the quality of existing tech? If adding meshing isn't a boost to current levels, it's only going to fall apart harder with higher. I think they are jumping the gun with this testing. It's pretending to get progress with no evidence beyond lucky outliers in small bits.

[@518UN4]

I had the same experience as tenpoundfortytwo. But also I went to grimhex at the 100 player test and then logged in there the following tests.

[@UmmerFarooq-wx4yo]

This is why $20-$50 a month for good servers for new tech mmo.

[@Spike.SpiegeI]

Even 200 is fine with me if my server FPS is consistently 20+ and the AI is working. I'd rather have that experience than 300+ players and things are as they are today in game.

[@reamoinmcdonachadh9519]

I think it will be ok, when it is 'eventually' released but I wont be surprised if there is a bit of "transitional" problem solving by CIG

[@cactusmamelu313]

My experience was similar on EUW servers. A bit better than previous tests but not by a lot. On 500, I managed to use QT, change server and tried a bunker mission but it was pretty much unplayable. On 1000 it took 10 to 15 min to log in and it was a mess I was stuck at Tressler screens. 750 was almost like 500 and 350 was I bit better than 500 but it was not that great either, I managed to do a bounty at Yela. Like you I was very frustrated at the end of the session (it was late at night for us). I don't understand how they ain't able to test the build with bots and tell right away what's. They can record previous tests and replay them to try to see how the new code preform now and compare (Im a software engineer and this is common practise in the field)... Either they're lying to us either they are quite helpless in term of those techs. Bottlenecks sure, but come on CIG it were only couple percent better than the previous one. We were not 2000 players and suddenly a new bottleneck appears and needs extra work to be tackle... Im sure they should been seing that it was only small improvment with their very precise "microscope" and not a giant leap as they let us thought...

[@ArcticGamez]

500 / 1000 players had massive delays and crazy desync for every damn thing i did 500 player had like 5-10min delay ping into the 300-700 range as well 1000 player had over 15+min delays had pings up into the 1-3k with spikes up to 9k for w/e reason

[@darqshark3506]

If there's 90-100 players per server, that won't be the same amount of players in a dead universe we have right now. There will be multiple servers. So if there's 6 100 player servers there will be 600 players in the same play space + pyro. That is 6 time the amount of people. Unless I'm confused

[@1aatlas]

one thing i would like someone to talk about is how this is going to work. For example: A 400 player "shard" would consist of 2 x 200 player 3dgs "meshes". 200 players, 3 dgs mesh for stanton - 200 players, 3 dgs mesh for pyro. In that fictional scenario surely each server mesh will have to be able to handle the potential for all players from the shard to be in one system at the same time... so with this example any 3 dgs server mesh (stanton or pyro mesh) will have to be able to handle all 400 players being in stanton or pyro at once. does that even make sense to you? im having trouble phrasing it.

[@pebegou]

WHO gives a shit about server meshing if we have 100 ppl online with no physics whatsoever and all this insane delay?! Just make a normal server at this point.

[@alexpetrov8871]

This is exactly the problem that a server mesh should solve. The game has physics, it is not the best, but it is there and is calculated on the server. Creating a server mesh means more server frames per second and more resources for better physics calculations.

[@1aatlas]

We already have a "normal" server setup... 1 dedicated server running everything... To quote Barbossa "It be too late to alter course now mateys! Arghh ahahaarrr!"

[@SETHthegodofchaos]

@@1aatlas Pretty sure he means going back to instancing areas like it was supposed to be in 2012-2014. Like when we quantum from one place to another we could just be put into a different instance. No RL required. No real-time replication across servers required. Direct connection to game server. Easier to limit how many players are in the server for optimal server tickrate. Downside being that we wont get many players into one area or into a big fight. You wont be able to see players in other instances.

[@qutatron]

I didn't even be able to install test version, because unknown error appeared.

[@rickskynight98]

Mike stop with the BS that CIG doesn't care about coming out with the game. That's your opinion and not a fact.

[@floppalore]

Would it be better if they lowered the player count to maybe 75-ish players per server and then increased the amount of servers per shard?

[@vorpalrobot]

that would increase the fps for stuff like AI and physics, but then you have more servers sending more messages to more servers. The replication layer needs to talk to even more servers which could see more slowdown in certain aspects.

[@floppalore]

@@vorpalrobot ah ok

[@rhetor4mentor13]

Mike there's a simple explanation for why your experienced differed than 1042: the shards had different A/B test configurations at each player cap. It's usually how this kind of test is setup, there are control and test conditions and the of the data is that you can isolate away how specific settings affect performance

[@SETHthegodofchaos]

Yeah, 4.0 is only the beginning of the end, not the end itself. 4.0 marks the time where the engine is in a "technical feature complete" state. 4.X will then be used to get all the MVP gameplay and content in before all of that is polished in Beta. Thats when people should expect a good experience at the earliest. edit: I mean, it is a big milestones for CIG. Just look at how long it has taken. But yeah, people are tired of waiting and coping hard that 4.0 will finally deliver The Game. Although that is still coming.

[@robertchandler2063]

I also want to say it’s dumb as hell to critique this knowing they wanted to test certain metrics not in its current state in 4.0 branch.

[@Valium_x]

Even best case with 1000 across 6 servers how will this be an MMO when we have 3 or more star systems...the universe will just look empty.

[@jamenta2]

Weird how the Server technology is not layered far away enough from the code in the game where everything ends up being buggy as heck. Why wasn't the code layered correctly? I can change the hardware in my computer with very little impact on the applications I use.

[@Mysterious_RSA]

cheese

[@kennethjensen730]

I said 2 years ago, that this game is still 10 years away, People where screaming at me lol, its ready in 2-3 years lol, now im not even sure 8 years more will do haha. Fucking amazing. Biggest scam ever if this fails.

[@merlark]

yeah if it only supports 100-200 players that to me proves that their server meshing tech just isn't going to work at least not for the type of game they are trying to make. It will prove that what they are trying to develop isn't reasonable with the tech that they have available. At that point the only thing that will make the train change tracks is if the player base blows up the bridge and stops investing so much money in the game. at some point Star Citizen needs to accept its limits and make some tough choices on what they can deliver and what they can't deliver in a reasonable amount of time.

[@illgottengains1314]

I’d rather they had 10 player servers and proper AI NPCs

[@gam3rguy174]

mayebe he meant 100 players per location like planet servers, stations and what not

[@Fragnatix]

The backend of 4.0 needs to be rock solid, the front (client/game) would be ok with current WIP features, but the servers has to be much more robust. At this point, its not acceptable, because things will get much more ugly if its not near perfect. And thats not even resolving the cost of infrastructure it will require from 4.0 and over. Are they going to sell ships again to cover the cost? I don't think people will get into the band wagon like before. So how? Fake it until you make it?

[@Fragnatix]

Even if we reach 300 players but we are bottle necked with 300 players, it only tells us that server meshing can't scale, and if they can't scale, it will run like shit at 300 players.The only parameter to prevent meshes to be over 300 temporary is the client performance on the current mid range machines. But should def be able to reach 1000 easily. At this point, its either its working, or not, because on the backend side, this will bring a lot more discomfort for us players if its half working or not working at all.

[@robertchandler2063]

I hate no streamer really tested on 100 player multiple sever just to see the performance improvements. I want to know if sever meshing works just fine without increasing the load on a single server.

[@The_Real_bubbazaneti]

at the end of the day it will either be an Epic success or an Epic failure! to be honest, I think the max will be like 300 players when/if they launch 4.0 end of year or beginning of next.

[@absenth2007]

sorry I'm confused. if its 100 players per server shard or per server or per solar system? I was thinking each time they mentioned player counts it was per shard. so we can have 100 players per planet kind of thing depending on where the boundaries of each shard are? i thought the idea of a server being a solar system was the old/current live tech.

[@uther10]

Lol you mean The Road to Pyro? 350 people? Lol good call on everything still being barren.

[@youdotoo3]

I never made it to the hanger

[@Viktoria_Thaelin]

Nah, SaltE got worse delays cuz he said critical things about CIG and SC lol

[@alexanderdooley5833]

I had both his exp and Mike's. First 2-3 tries were good rest baaaaaad the 1k test was just all bad.

[@Terran0va_Plays]

Not only does server meshing not work. But CiG keeps pretending like they created this tech lol. This is not new and it's been successful a couple times at this point and way more now in 2024.

[@mane4582]

i personal think 350 player is more than enough. if we have 350 along with the 10 to 1 npc the game would feel alive. the game doesn't need 1000 players to feel alive. i am a bit confuse r they planning to have around 1000 player per system or around 1000 across both? Cause if it's per system i understand this test but if it's across both maybe having hundreds of people rushing to login in at basically the same location handle by a single server in that locate might not be a great test of a normal day. i wonder could they do something like have one server handle the area of the city (LZ) where u spawn in and it hands off to another while u on the trans system to another one handling the spaceport and the rest of the planet till it hand off to the one in space. just random things from watching all this

[@tossitupman]

Start with 100 and put the higher numbers on ptu or playtest. I just want to see what this server meshing ai is supposed to be

[@kitchenersown]

man...this is what 13 years and 700 million gets it. My last hope is on SQ42. I'm hoping that game will blow everything out of the water and make CIG a couple hundred millions USD in the long term. It should ideally come early to mid 2025...and after that only a skeleton crew will be left behind to do bug fixes while development of SC is sped up very rapidly leading to beta maybe 2027...but maybe thats too much. Im overdosing on copium...im prepared to be dissapointed as shit and just hop back to Eve and Elite.

[@UnrebornMortuus]

how many speeders can we fit in server meshing?

[@jebidyah]

1 server isn't enough for 100, what kind of numb nuts math is CiG doing thinking 6 servers would run 1000? Are they really trying to min max how much money they can save by running the bare minimum of servers to run the game smoothly!? 800million isnt quit enough? At its current state of optimization, 1 server can really only handle about 50 people. It should be 20 servers for 1000 if they had any common sense.

[@mayoluck]

Tube socks?

[@Dulc3B00kbyBrant0n]

how can servers mesh if they cant play functionally on a small scale

[@avatv4658]

Tests were done on servers closer to him physically than you maybe?

[@beavschannel5217]

If we have the same server performance but just more players, is it really a success? I know I will still wait and not play, I would much rather have the same amount of players we have now, but better performance so we can actual enjoy the game and see things like AI actually work., and then let them build up the player count.

[@Aeoxmusic]

CIG has a douche flag for IP, makes them hate their life playing the game

[@CFDepartment-su5fj]

UK vs US servers

[@saltybulldog3241]

starting to find your cocky attitude really grinding. Dude thinks he knows everything.

[@theyshouldhavenevergivenme5439]

When has anything ever worked. When has anything promised ever been delivered in a timely manner. NEVER. The joke is on all these 'testers' and shills actually. A couple of more years and the entire thing will be technically obsolete anyway. And CIG pockets will be bulging. All these people waiting in vain, following every announcement are in fact the actual business model. Always were. 1200 devs my ass!

[@bartasbxx]

That's mad. All the rage about the things they sell but this is like... If this doesn't work this whole game doesn't work. And this doesn't look like it can ever work. I mean nobody else has done it I don't think CiG will be the first to do this on old tech.

[@younboyce7068]

Server meshing is the biggest scam

[@xiuhtezcatl8161]

dead game

[@TheJack4000111]

This game in going nowhere.

[@palehorseriderx]

15:34 CIG developing new time-dilation tech. Is this what Soulsinger is? /s