Server Meshing PROGRESS!

[@tenpoundfortytwo]

Thoughts on the test? What do you think we'll get for 4.0?

[@LK-R]

200 player static server meshing (1 for each system) for 4.0 with dynamic server meshing coming in a later 4.x patch

[@LK-R]

@@CyborgeWashington yeah, you're right. I was thinking that as well. I just don't want to be too optimistic yet...

[@The_Fallen_1]

Given how they keep testing multiple servers per system (the 3 and 6 server setup seem to be popular), I'm somewhat inclined to think they're going to try a 6 server setup, with one server for space, moons, and planets in both systems. As for player numbers, I think they're going to end up in the 200-400 range based on the results of this test, though I think they'll try to push it higher if they can fix most of the bigger issues they just found for the subsequent tests in the near future.

[@echo5827]

Reminded me of ~2018 SC, people frozen in place teleporting and trams killing you. Item interacts not working at all The 4 server 350 was very smooth, so I expect similar ratio but for player counts I'm assuming over 500 but less than 1000. 8:700 or 6:500 minimum for 4.0 unless they get 6:1000 working

[@TKanal3]

@@tenpoundfortytwo cant say its not a step in the right direction when there is shards with 1000 people and the server survive it somewhat. But it aint a big step. Gonna need MORE TESTS!

[@dawnphantom2439]

Test was pretty chaotic, but it seems like their goal was to stress the system to it's absolute limits which they did successfully. Everything else they will probably explain to us later, but it seems like on their end, they were able to discover new things to work on which will hopefully help push the tech forward. Looking forward to it.,

[@AlbertoMartinez765]

Yep stress it and see what breaks.

[@Crittek]

@@AlbertoMartinez765Everything

[@hiiiii895]

This is amazing! Incredible simulation. I can't get out of bed in real life either.

[@1aatlas]

They clear one network traffic bottleneck only to discover another no doubt. Its progress, progress is good, and like you said they can always fall back on 100 player caps for now to maintain stability if they really want to deliver 4.0. Thanks.

[@ImreBertalan86]

I love how you are the only one who is looking at StarCitizen more on the technical side. As an IT engineer, this is my bread and butter and I'm here to stay 🙂

[@tenpoundfortytwo]

Thanks! Also, I think as we ramp up towards Citizencon, it's nearly time to run the treasure hunt!

[@unobtanium]

Sadge :c

[@Sinsanatis]

for a sec i thought that was beard and butter

[@fnunez]

The 30Ks and Bault's optimism make sense. The DGS code was never able to handle requests coming in that fast, but before RMQ it didn't have to because NMQ wasn't fast enough to deliver them. Now that RMQ can deliver fast enough, the performance problems that were always in the DGS have come to the surface and can be fixed. I agree, huge shards are still months away but I expect we'll get RMQ really soon for its other benefits (lower button lag etc).

[@tenpoundfortytwo]

Yep that's a good summary!

[@aguy446]

RMQ is going live tonight

[@gankstrap]

@@aguy446 RMQ been live since 3.24 on some servers

[@mabutoo]

This is great. This should give CIG threshold data and scaling data. Maybe start with 300 then when they hit 600; add servers and vice versa when player count drops. The rendering issues will tell them how much to rely on the client side over server side. Point is the game ran great and the problems seem to be in server client communication. These are very positive results.

[@filipetavares1345]

its a sucesfull failed

[@Franchisco7]

Just as a helpful piece for clarity for viewers. Please only press the interact once. When you spam it over and over it is hard to tell how long the delay is and if the errors are from spamming the button. Plus if every player is spamming the buttons it will only overload the requests more.

[@bouck06]

Thank you so much for this and the gameplay footage. So happy to finally see this happening.

[@tenpoundfortytwo]

No problem!

[@cpt.tombstone]

"Plenty of train-related injuries" LoL :D

[@neocaron87]

The mesh test reminded me of the Weeping Angel in Doctor who XD. The Test server felt like a horror game XD

[@tenpoundfortytwo]

Yep it was closer to horror than any other genre!

[@BboyCorrosive]

It does finally give me some (c)hopium that the next 1000 player test will be better and kinda cool.

[@OneTwoMark]

The main game is already broken enough to cause me to RQ, the people who step into the technical tests have nerves of steel.

[@harabecswurveweathers]

My prediction for the game up to the release of 4.0, is that CIG is going to have Evocati builds up in a couple of weeks testing general Stability & features, while CIG runs the SM/RMQ tests on the Tech Preview channel in parallel to stress the systems. They're going to keep applying fixes after the tests and run them again and again leading up the the LIVE release of 4.0, possibly around IAE in November. That way they can have the numbers they need to push the scalability of these systems, and the laser focus of Evocati to make sure the features and refactors that interact with these new syatems are working.

[@cyvan1750]

He said later that RMQ worked but they found bottlenecks further down the network so that's what they're working on now.

[@tenpoundfortytwo]

Great info, thanks!

[@tbrws]

I’m unsure if the server was full, had no one to confirm. But the 350 player tests at the end of the testing sessions went great on my shard. Was very smooth and no real interaction delay. But I also guess 100 per system, they can collect data and have time to mature their their tech. People will be happy with pyro and seeing npcs somewhat work.

[@UtahSuka]

RMQ needs a lot of work but is expected it's a brand new system. Notice that server 30 FPS now that is impressive. Once the RMQ issues are resolved, hold on to your panties, people will be asking for AI to be tuned down a nudge.

[@minecraftprovie5076]

are those 30 SFPS real? because it's a bit weird that average is locked at 30 all the time and min is always like 1.5

[@VFW-Mayer]

Movement needs to be client streaming with a server check of proper time stamps. The less broadcast domains the better. My worry is the overhead for areas. The more complex the zone, the more checks the server needs to do. With higher pings between clients given players pings to servers adds rubberbanding and lag skips. Like 3 people in combat with pings of 100 would be a stuttered mess.

[@Captain-Comedy]

I nearly got that ikea light in the background today. funny that I see it now on your vid

[@refisherdesktop]

I usually play solo, for various reasons. Most of the time when I'm playing w/ another player I'm helping some newb along to toddlerhood. Seeing that many players in the core areas all the time is just gonna be weird. But that's what they've always intended. "Get off of my asteroid!! "

[@starzgarden5553]

When we get to 100 star systems or near it, 1000 players or whatever will disappear so won’t imagine crowded train stations in future

[@Voodoo-b4f]

Great vid, as always.

[@tenpoundfortytwo]

Thanks!

[@drakonhawk]

We will see if there is a new test next week, Bonoit wrote it this afternoon. 4 Server with 350 Players works a lot better. Never got into 1000 player server. RL crashes before. In space it runs much better. The cities was overrun and really laggy with much of desync.

[@GothyCakesofficial]

I love how the bugs usually consist of either the train taking a dump or lifts killing you 😂

[@4hire565]

Came from Salty channel. Rather hear the stuff directly from you my dude.

[@tenpoundfortytwo]

Hello, thanks!

[@Tartiflette-py5zh]

Good job ! 😉 👍

[@Jon.W.1337]

I think the problems are a non-issue. I would much rather have a meshing test every month and have it have issues; as opposed to a test every 4 months.

[@tenpoundfortytwo]

Yep would definitely prefer more frequent tests- there has been some talk of another test next week, which would be a good sign!

[@Silverhawk-u2f]

From what Benoit said on SCL, my understanding is that they didn't have much to test before implementing RMQ

[@quarkedupphoton236]

Firmly agree. SC is a bugged mess in year 12 and most issues are years old. If they can't resolve the player count as core feature of the game, push the meshing envelop and the MMO universe where fleets and player interactivity meet, the game becomes null imo. Only Sqn42 will be workable and relevant.

[@Jon.W.1337]

@@quarkedupphoton236 YES, I have been commenting this for years. Usually get down voted for it. I'm happy to see someone else sees this too. Server meshing is the only tech that makes any other piece of SC viable. (IE the real reason capital ships aren't in game). Without it there is no SC. I realized this when the Cutty Black was released, which was my last SC purchase.

[@Zeeves]

Can’t wait for this, likely it’ll release with 1000 player cap right off the bat, it’s going to truly change everything

[@tenpoundfortytwo]

It'll be interesting to see if they can push it that far at the start- I'm not sure!

[@KA317S]

I was apart of the 1000 player test and I wouldn't hold my breath on 1000 player shards .. it was a disaster.

[@Zeeves]

@@KA317S yeah, it’s a test, once it’s working I expect they maybe even up the count to 2000

[@BigBobBlazer]

With CIG, I’ve learned to believe it when I see it and play it. Speculation is worthless. “4.0 this year!” has been a mantra of content creators and backers alike for 5 years now. Every year there is a new reason for backers to believe that this is the year. The year comes and goes and all we get is a little starter ship here or a new flight model there. But the system and game feels relatively the same. Just some new dressings or different ways of interaction. I wouldn’t be surprised if the Polaris gets pushed to next year.

[@Jody_the_real_cool_cat]

Definitely is NOT ready, lol hope they don’t rush it and just work it out and get our data and make a good release to live.

[@jacklapham3097]

I'm not limited by hardware or networking, and still it takes me longer to get out of bed than that

[@supersasquatch]

im a big fan of the concept of much larger pop servers... i wonder the impact it will have on currently relatively safe non combat activities

[@tenpoundfortytwo]

Yeah so much stuff like that for CIG to work out!

[@Cmdr_V]

Hoping all those big orgs that been practicing med running are ready haha. Definitely going to be a blood bath lol

[@MersageSW]

Thanks for the video. As I understand, server meshing cannot limit players per star system. There's no 100 players in Stanton and 100 players in Pyro. Server meshing can only limit the total number of players across both Stanton & Pyro as one server mesh. Meaning, 190/200 players can all go to Pyro and 10/200 players can go to Stanton or vice versa.

[@cryshot8071]

I believe the underlying problem here is simple: Static Server meshing just doesn't cut it. Once the players spread to unevenly, limits are hit. And that happens quite often, since you have just a handful of major landing zones. Even if you have a static mesh with 6DGS; if 200 players are on Lorville, they will still all be handled by a single DGS, so a single server would need process double the amount of data as they do now. Can't speak of an advantage here, in certain scenarios it's even worse at best.

[@gankstrap]

200 players in Lorville would only handle those 200 player entities and the ones in Lorville itself, meanwhile now if youre alone in Lorville your're handling 100 players and the whole star system entieies.

[@cryshot8071]

@@gankstrap That is a valid point, but my arguments still holds true. That’s the reason why they want to establish dynamic server meshing in the long term.

[@gankstrap]

@@cryshot8071 Of course the game vision is not possible with SSM, but they need to implement that to scale on a dynamic mesh. Keep in mind Static SM it's like T0, the last test it's 4 DGS managing 350 people son that's less than 100 in each DGS, if they can make that stble you can end with 100 people on each main city and 100 for the rest of the space and that will give insane performance because I doubt you can fill the 100 population in 1 city without trying to get all the people there on purpose.

[@Crittek]

Things need to get a lot better very quickly. Let’s say we get 1 million players even split between 4 regions. 250k players, across 5-10 systems? Pffft no shot. So minimum each system needs to support 2500 players at launch or else we’re sticking with the ever present 2001 server technology.

[@InfiniteWatermelon]

Nice update. Subscribed for more! After they release this first version of server meshing, which will be only 1 server per system, how long do you think it could take until they do multiple servers within the same system? Is there much more code involved or is it just a matter of sectionalizing the space?

[@chaoscorgi9348]

I think that 250-350 is realistic, so CIG is going to try and tweak it according to their info, throw us into servers with 400-500 and brace for Spectrum to be a roiling cacophony of weeping, wailing and gnashing of keyboards. They will also have a ship-ton of new shiny things to distract the most vitriolic so 4.0 will be largely considered a success.

[@DaringDan]

Good stuff. I think your prediction about 100 player 4.0 servers is spot on. The way I look at it they have almost exactly 3 months to get this on the PU before their winter break. Then we'll have the usual month and a half of nothing. I just can't see it happening and hope I'm wrong.

[@BboyCorrosive]

I'm just thankful I set my spawn as Baijini during the 100 player single dgs period, avoided all the train stuff, so it was just hab and elevator door/asop issues

[@gankstrap]

They said stations had a bug and told people on the patch notes to avoid them. Also the objective of this tests is to stress the system, avoiding all those train/hab/door/asop issues makes no sense since it's what they need to test.

[@Billy-bc8pk]

Based on the tests just before the servers were brought down, the 350 pop shards seemed to work fine (for the most part). That seemed to be the sweet spot that CIG was somewhat gunning for. So if I had to guess, they will go for 150 players in Stanton and 200 players in Pyro for 4.0, or something thereabout, given how big Pyro is. Already with 100 people or 120 people in Stanton, it still comes across as fairly barren. 150 in Stanton should make it feel a little more alive. And given how big Pyro is, anything less than 200 and it's likely going to seem like a ghost town in most parts of the system.

[@dtrjones]

I don't want to be the bringer of doom, but it sounds like RMQ or there queue replication tech still isn't fit for purpose. We still don't have a server mesh which performs adequately with more than 100 players per Game Server. As this point, if we move to 100 players in Stanton and 100 players in Pyro, it will be enough to role out the release but will it apease those who think the server meshing tech is Star Citizen's holy grail...

[@NoQuo]

Well, at least you stayed in your ship while it continued QTing. My Gladius kept going without me after I hit the server borders 😂.

[@FlesHBoX]

Yeah, all these people thinking that since one node can handle X means two nodes can handle 2x... Aside from the overhead of communicating between nodes, CIG also has to account for the what-if scenario of when for whatever reason everyone on that server decides to all go to the same place, which can't truly be answered until dynamic meshing happens.

[@Funkm4ster96]

Honestly, I felt like this test was super promising. I would totally be ok with 200-400 player shards if we can get actual SM into the game.

[@tenpoundfortytwo]

Yeah I would be happy with smaller shards to start with. I know a lot of people really want to see the game actually become an MMO quickly, but I think they should start slowly!

[@pawe6199]

I don't care what they say. They've been saying stuff for 12 years. All I saw is giant lag, delays, desync, fucked train and 30k. Looks like a fail.

[@vulcan4d]

Boys, in 10years we will have a solid launch.

[@unicaller1]

Shard player limits will be per shard not per solar system. I expect 350-500 range for 4.0. I think it will depend greatly on how players distribute during 4.0 PTU.

[@willianbill1510]

I have a Dell G15 5530, with a 3050 6VRAM, i5 13th and 16GB Ram, what average FPS should I be able to run SC?

[@HarperLylia]

Maybe a hot take, but i really hope they keep the player cap at 100 for the release of 4.0 and slowly work on getting it up as the tech improves.

[@Tainted-Soul]

Its so different experience between different youtubers Ollie looks like he had the best time LOL Saltemike the worst makes you think was they running different server set ups on different shards?

[@Merkabuy]

BIG suggestion, please put in the dates of what youre reporting on.

[@tenpoundfortytwo]

Good call, thanks!

[@7thSmurf]

"the TRAINS were absolutely deadly !! " .. thats how u know its a quality product worth 700 Mn dollars.. at least .. ^^

[@SC-monzo-nt2lo]

If I remember correctly they said at one point they had 1700 players on tech-preview testing, which is much less that I thought tbh!

[@echo5827]

1000 in stanton still has me worrying about if the balance can handle that more than the servers. Imagine a 10x in players everywhere, so the two at breakers are now 20. The 1 guy in the shop is now 10 people standing in line for a terminal. 10x the amount of player bounties or people at spk or taking missions I think the max cap for 4.0 couldn't be higher than 3-400 in stanton, but 500 each would be fun

[@jokesy6112]

i managed to get to space port in NB. called my ship. got in my elevator. saw my hangar name on the panel. went to click it and some guy spawn out of thin air, the hangar name changed to his, elevator started moving and i got thrown under the map since it didnt teleportt me to the lobby but i cant entter his hangar

[@littleturtle8264]

thanks

[@tenpoundfortytwo]

You're welcome!

[@PbPomper]

Difficult to say. The RMQ implementation might have exposed some bottlenecks/bugs they can iron out which could totally transform server meshing performance.

[@Stormyy6310]

Real quality SC content

[@tenpoundfortytwo]

Thanks!

[@ericwollaston5654]

So, that same RMQ is now running 3.24.1, right?

[@tenpoundfortytwo]

I believe so yep!

[@sumiral6438]

I wonder if the amount of dgs in a shard are effecting the rmq delay negatively or are just the player the problem for the rmq

[@saleseng]

In my opinion, the train system is cute because we know Chris Roberts is trying to emulate half-life along with the gravity gun/tractor beam motif as well..BUT… when is the last time you were on a server in the past few years where every single train seat was taken up? In my opinion, the trains are once again aesthetically appealing to Chris Roberts, but they really don’t carry the practical function to justify what a strain they are on entities and performance of landing zones.

[@15ewolsey]

Anyone have the link to that tech preview on spectrum? I can't find it

[@okami803]

Was looking forward to your review. I’d love to participate but sadly I am bricked from updating. I deleted the game hoping I could with no luck. Any pointers?

[@okami803]

Sorry, I know that is vague. I think it was a 2000 error

[@tenpoundfortytwo]

Ah that sucks sorry! Best bet is probably a support ticket, which you've probably already done. Googling the error code might give some Reddit results that might help.

They will allow 150 players max! Thats what I would do. If you assume n*n style network traffic situation, its around 200% the load of 100 players which should be doable.

[@AuricNova]

I'm not expecting, nor do I really want, a huge player cap increase until CIG is at the point of developing and testing dynamic server meshing rather than static. I don't really know how static could really increase the cap all that much since players will be free to congregate in one star system, planet, city, or ship. Unless that is somehow helpful for data collection purposes, I don't see the point.

[@nachoilf8226]

They need to bite the bullet and add more DGS and no more than 250 per system and start polishing net code for optimization for future releases.

[@jossdoggett7213]

So say you have three or six DGS committed to a session/server. Is Lorville still hosted on only one of those servers? Like each server has specific areas of the game committed to it, and when you cross a border you hop between servers. If this is the case I'm not surprised Lorville is still lag city. Its essentially still only one server managing a massive portion of 500 or 1000 players all spawning at a similar time. So we need to see single server performance optimised, right? Hopefully RMQ is the next step on this.

[@jossdoggett7213]

Another thought: while I could see the wider universe being hosted successfully as above, could it make sense for the major cities, spawn points, and popular activities to be instanced? Assign a server per instance and boom, stability. Sure it's not the persistent, everyone everything all at once that is the dream. But maybe a middle ground?

[@arctic4291]

​@@jossdoggett7213The answer to this Is : Dynamic server meshing. With the static Is a problem

[@cmdr_stretchedguy]

for me the tech preview just crashed before game even loaded, never got into game, never got beyond the initial pre-game loading screen

[@TKanal3]

they say it was meant to burn the servers down to see where the cracks happen. Lets hope that is true and not an excuse

[@alejandrotroche6381]

Doing an AMD build sith 96mhz (2 48gb sticks 7000mhz) for content creation SC. Should i get a 4070 super ti? Is it good for 4k play? Ryzen 9 9900x any good or last generation? Will use da vinci resolve for editing. Not if is good for both Star Citizen and machinima recording. Thnx😊

[@tenpoundfortytwo]

Not sure 7000Mhz would run, 6000 cl30 (or maybe 6400) tends to be the sweet spot for am5. I reckon the 4070 super ti should be enough for 4k (as long as you're happy to use DLSS). I'd still go with the 7800X3D, I don't have any issues on the content creation side with it and its the best option for SC performance!

[@rapamune]

zen 5 has a massive avx512 capacity

[@alejandrotroche6381]

@@tenpoundfortytwo i would do live stream can i do a 9 7900x3d instead of 7 7800 looking not to bottle cpu while doing 4k streaming

[@TruthIsKey369]

Only on 3 servers? What is the server specs on one server?

[@captainefox8337]

if they make servers bigger and stable it's fine if they make it more than 100 players it's good just like they did when it was 50 players than 100

[@starzgarden5553]

Looks like they are a loooong way from getting this tech to work. No way ready for 4.0. They’ll be working on this all through next year and beyond. 100 per system most realistic

[@Fayker]

my take on the rmq and sm tests that, there is too much servesided dependacys ... like doors trains and elevators... if they keep this trend on everything is server side... server and bandwith costs gona be ludicrous

[@Diogenes76]

What I find curious about them spending so much time on the queue is that I thought the idea was to get meshing up so no single DGS has such a large queue in the first place?

[@gankstrap]

f course it will if you increase player count x10 times lol

[@CellDE]

It looks like the current performance thing is around 75-100 players per dgs. iam questioning if we are able to do an MMO without paying 20 € per month for the server costs ...

[@tenpoundfortytwo]

Yep, the next couple of years are going to be interesting from the MMO side of things!

[@Cmdr_V]

​@@tenpoundfortytwo monthly subscription would absolutely kill my whole vibe of playing this. I hope I am not alone lol

[@tenpoundfortytwo]

@@Cmdr_V Well they promised they wouldn't do it! I think there would be a lot of outrage if they did!

[@gankstrap]

@@tenpoundfortytwo They promise that to backers, anyone buying the game after 1.0 can be forced to pay a subscription

[@davidarhall]

50 people on Stanton 50 people piro. If it were 100 on each, everyone might move to piro and cripple it.

[@tenpoundfortytwo]

Yeah, that's a good point, going to be very interesting to see how they mitigate everyone moving to one dgs.

[@Valium_x]

Is 350 players across 4 servers better than 100 players across 1 server like we have now?

[@GW2Vids1]

I know it's a stress test but the systems are not big enough for 500 people. Not enough asop terminals, elevators etc

[@justsomewhitewolf]

Only video I've seen so far that's talking about the test with a little more information than "wow omg we're writing history". My thoughts... The game being an alpha with gameplay issues and with balancing issues aside, Star Citizen is getting a lot of shit for the following two reasons: A) there's not enough to do, and B) it performs like shit. They can fix the latter by focusing on delivering an experience with as little latency and issues on the live as possible. Zero Latency consistently, if possible. On the other hand, the former can't be fixed by increasing the player count alone, you can't rely on players to be your content for you, especially not if performance suffers as a result. So their focus should really just be to increase the player count ONLY to the point where they can be sure the game will run smoothly. If it means 50 players max across both systems, then that's how it has to be done. If 100 players works, great. If 500 works, even better. In the meantime, they can keep working on the tech, push out as many of those server meshing tests as they can, but leave the Live build alone. Eventually they'll be able to increase the player count if their tech can support it. Of course I'm not speaking for CIG here, maybe they really really need data from the Live build to the point they can't avoid harming player experience... but I highly doubt that. Aaaand of course CIG wouldn't listen to me on that topic anyway since they already have 100 player servers that don't perform well north of the 30 player mark. Their data probably supports that star citizen players will play the game no matter how awful the servers perform so they can save themselves some money and cram as many people onto a single server as they can without the thing crashing. So my own thoughts on how things should be handled aside... I fully expect them to do exactly that. Server performance aside, they can just promise players they will work on it, and then cram as many people onto a single server as they can, even if performance suffers. Saves them money.

[@bangyahead1]

It was terrible but thats the whole point of testing.

[@beeleaveme]

not gonna lie, the "server meshing" in Ashes of Creation looks more promising

[@tenpoundfortytwo]

Yeah it's going to be interesting to watch them develop side by side!

[@gankstrap]

How can it look more promising if you only watched a 10 min demo about it and a closed alpha test with no real players just devs, it could be 1DGS, or 2, no one knows

[@marcuswynn89]

RMQ has a bunch of memory leaks

[@yous2244]

Can you make a video on how Anti Cheat is lowering fps

[@tenpoundfortytwo]

Hi, sorry I didn't reply to the other message! I'm a bit split on whether to look into this or not! I'm not sure testing with anti cheat off is really going to be much use as we are kind of required to use it. But I can see how it's an interesting topic. SC also wasn't great before they added it! I'll have a think, thanks!

[@yous2244]

@@tenpoundfortytwo 👍

[@EXRAY479]

Watching this performance of server meshing is making me lose hope, at this point I think we will see dynamic server meshing in at least another 5 years, I think we will see 500 player servers by the end of 2025 maybe and 4.0 will just be 2 servers 1 pyro and 1 stanton with pyro being super laggy with nearly everyone playing there and stanton being empty

[@DustinHarms]

Yeah...I dunno, man. Your footage of getting out of bed in the 500 player test must have been on the best shard or something, because me and everyone I knew testing didn't see less than like 10s delays, at best, until way later when they dropped the caps back down and most everyone stopped testing. None of us even got to space. So many deaths and issues just getting to the spaceports. If the goal was less delays, I don't think I'd see this as a success in any way, unless the last time you test you got a super bad shard, and this time got lucky and got the best shard. Far as A-B comparison to the last test with my whole crew, everyone felt this one "felt" worse. Plus the players not even moving, the transit still being broken, just so many issues. The only improvement that seems universal is not getting pulled out of QT when crossing boundaries....but everything else really seems to come down to getting lucky or not. So, "definitely improved" is certainly not how I would put it. I'm sure I'll get blasted just like everywhere else for sharing the experience me and ~12 other people had over like 6 hours as being "too negative" or some BS regurgitation of CIG's "It was great data" or "it was a success to us" but considering the time between the tests we've had this year and how they performed for what seems like a majority of testers, I just don't get the optimism. There's simply no way we're getting 4.0 this year, unless it's like a 2 server mesh with 150 players.

[@DustinHarms]

May as well add - for the last 4 years I've refrained the words, "I'm disappointed, but I hope I eat my words" and unfortunately, I just haven't had to, yet. Benoit can say whatever, that it's new issues causing the same problems, etc. but that doesn't change that we've been expecting 4.0/SM for 4 years, and were led to believe "for real this time" and even pushing back against the community by saying it was still planned for Q3. But here we are, and the state of the tech is so far from ready, it boggles the mind to think how on earth they thought that kind of prediction was remotely feasible. I work in software development, I know timelines are hard. I have a lot of respect for great project managers. But the stuff CIG pulls is so off the wall there's just no way to reasonably defend how they communicate their progress on the project. At this point, the reality of the situation is this: - Server Meshing as a technology is fairly close to ready (as in, connecting 2 servers, with no player increase) - Star Citizen becoming an MMO is still very far off, and the technology they have in place to do that has NOT been proven to scale yet - CIG would rather string everyone along indefinitely until, someday, they manage to get more than a few hundred players together in a reasonably stable environment And to that last point, I wouldn't be surprised if they launch the game there, as an "MMO" with less than a thousand players together. Anyway, already putting too much energy into wasting my time about this. Surely the *next* SM test will really blow my mind!

[@tenpoundfortytwo]

Yep good point- this was only one perspective on the test. Competely agree about their approach to timelines and this: - Server Meshing as a technology is fairly close to ready (as in, connecting 2 servers, with no player increase) - Star Citizen becoming an MMO is still very far off, and the technology they have in place to do that has NOT been proven to scale yet - CIG would rather string everyone along indefinitely until, someday, they manage to get more than a few hundred players together in a reasonably stable environment I think is a great summary of where they are at in reality. The massively part of MMO is still up in the air for me! I do hope the next SM test will blow your mind, but you're probably right!

[@UmmerFarooq-wx4yo]

The test shows that the train system was nonsense. And it should be individual taxi pods.

[@cmdrls212]

RabbitMQ FTW!

[@BestShifty]

The transit system has been hot garbage for a while now and its no wonder that it didn´t go well with meshing. A transit system rework for server meshing is on the 4.0 roadmap for a reason. So I´m not that worried about those not working.

[@tenpoundfortytwo]

Yep very good point!

[@7gpanda744]

CIG really needs those really strong servers.... like those monster servers the size of a building.

[@tenpoundfortytwo]

Yep! I'd love to know what version of AWS they use!

[@Franziskaner21]

Amalice throne and liberty test

[@kDenos]

thanks the omnissiah

[@Anomyos]

No time stamps? Me no watch.

[@Steven-t8h]

Never mind server meshing what about server working.

[@Volf1916]

I heard in this the statement that CIG's messaging system, the one everyone was hyped over and praising from the last citizencon was a fail. So all the empty hype again on CIG who marketed it as the backbone of the servermeshing tech that was so ground breaking is now thrown in the bin as worthless. Glad that is cleared up. interaction delay at 500 "better" is not viable. so this is NOT functional. And this is with MAJOR functions and components on the game disabled. Vary telling that reaching a hanger constitutes surviving to make "progress" and not "first minor issue". interaction at 1000 players, "not great" being a more accurate "dumpster fire worthless" Last years test and hyped expectations to the fact they threw out the system used on that test, to the point of this failure, I would say expect 500 players on a server around about 2035, when everything CIG has done is completely obsolete and can't even be used with the base line hardware gamers will be using as it will be archaic and too clunky. Sad seeing such a massive fail get hyped as progress. There is no progress here, just hype on "new" backend programming because what was hyped last year is gone as that hyped as ground breaking was garbage and has gone in the bin. Going to continue to tell others not to waste money on this project because that is what it is now, and I don't see anything as far as progress, just add on broken features that have more jank than substance.

[@indivestor]

disappointed. 4.0 rip

[@tenpoundfortytwo]

A lot of people feel that way! I'm still hopeful!

[@MrDazTroyer]

So still unplayable 😂

[@TKanal3]

your performance was not a ptu standard and people had worse experience than yours!

[@tenpoundfortytwo]

Yeah fair to say that this is only one point of view!

[@kVidStream]

I don't have faith in the way CIG knows how to scale their project that advances in multiple fronts and expect things to work out harmoniously.

[@Biggerfoot]

This is such a lie they pay for severs for one day to give a illusion of progress to new players and get a return on there investment it’s a completely lie, server meshing isn’t possible and the game wasn’t playable with 500 players and it was no different then the last 3 1000 player test and cig is paying you to make these videos which is against the law by the way

[@tyronedejager2202]

Nah game broken still, they won't get this right I am sure of it now . . . :( It is honestly a completely failed test.

[@cabritsanscorgaming]

What do you mean? Things broke (as it should be in a test), they gathered data, they now have a clearer view of the actual state, how is it failed?