this is a good explanation of what they plan is in implementation. good job. yes once the work load comes off the prime game sim there should be a boost to productivity and performance for the game is self. which also means population increases as well, so like instead of 100 people you would be able to get 600 people and so on, well that is the working theory anyways. will have to see what your next video investigation on this reveals.
Yeah, but when do they decide WHEN to switch over, because, when there are no 30k's currently, the server degrades so much to a point that it becomes unstable and flakey, especially over the IAE/Freefly. They need to have some cutover prior to the server degrading to such an instable state.
So that happens when server meshing is in and maybe a bit after the first implementation. The load balancing should actually prevent 30k issues a great deal altogether. You've given me a great video idea too with this question, thanks! I'll discuss how this will/could work
Servers hanging for 5-15 minutes instead of 30k is probably the lowest bar i've ever heard of. Ya 5-15 minutes of non-functional servers vs oh no the server crashed... Is there even a difference?
Awesome stuff man! And for performance it is bound to happen when only the single process is run (simulation in dgs and replication layer + smaller services in the other). Now the simulation only has to transfer some data to another machine running the replication layer, which transform to simple data send and recieved when it is needed to recreate a server state after a 30k. I do expect to see servers at least at 10-20fps to be the new norm as a result of this, so not a huge improvement but enough to notice
Thank you! I agree that we may see a VERY slight improvement, but once static server meshing comes online, if CIG balances out the load on each server, perf improvements could go through the roof
@@DigitalMaster_Gaming ah yes static server meshing will improve a LOT for sure. Why? Simple to explain as in we can have 2 servers run in say Stanton. Half the ressources or a bit less than half (for the overlapping zones) that are simulated for stanton means lots more computation ressources for said server for npcs and other problems, and that would mean 30 or more fps server side to be the norm imo
Could you explain one burning question I have? If the rep layer syncs multiple DGS, does that mean if someone that's standing next to me on a different DGS (and I can't see them) puts a bottle on the ground, I'll just see a bottle pop on the ground without seeing the player?
You'll see them too. Just as the rep layer replicates objects, it does the same with players. We'll all see each other as well across servers, it's just that only one will have authority over your characters simulation.
EVERY SINGLE TIME they find a way to improve server performance. Chris steps in with his fat ass and forces the server player cap to be raised. We will never be able to enjoy decent server performance while Chris is involved in development. We are here just to be unpaid stress testers.
this is a good explanation of what they plan is in implementation. good job. yes once the work load comes off the prime game sim there should be a boost to productivity and performance for the game is self. which also means population increases as well, so like instead of 100 people you would be able to get 600 people and so on, well that is the working theory anyways. will have to see what your next video investigation on this reveals.
The best explanation of the current and planned architecture I've seen. It feels like you've been reading some of my posts on Spectrum! Very good.
😁🙂 thanks!!
✊Well done DMG, Thanks!
THANK you so very much! I am very happy you enjoyed!
Yeah, but when do they decide WHEN to switch over, because, when there are no 30k's currently, the server degrades so much to a point that it becomes unstable and flakey, especially over the IAE/Freefly. They need to have some cutover prior to the server degrading to such an instable state.
So that happens when server meshing is in and maybe a bit after the first implementation. The load balancing should actually prevent 30k issues a great deal altogether.
You've given me a great video idea too with this question, thanks! I'll discuss how this will/could work
Servers hanging for 5-15 minutes instead of 30k is probably the lowest bar i've ever heard of. Ya 5-15 minutes of non-functional servers vs oh no the server crashed... Is there even a difference?
5-15 isnt the finished product..theres tons of testing and tweaking and dialing things in they have to do. Its not plug n play -
Massive difference yes. You won't lose any progress. Everything will be where it should be. You won't be spawning back in millions of KM away
I'm loving these videos explaining the backend work that CIG is doing to make this game a reality. 2024 is going to be a HUGE year for SC
Thanks! it is indeed going to be an impressive 2024 for CIG and SC
if they keep this pace, we might see the full game in 2050, so exciting
Awesome stuff man! And for performance it is bound to happen when only the single process is run (simulation in dgs and replication layer + smaller services in the other). Now the simulation only has to transfer some data to another machine running the replication layer, which transform to simple data send and recieved when it is needed to recreate a server state after a 30k.
I do expect to see servers at least at 10-20fps to be the new norm as a result of this, so not a huge improvement but enough to notice
Thank you!
I agree that we may see a VERY slight improvement, but once static server meshing comes online, if CIG balances out the load on each server, perf improvements could go through the roof
@@DigitalMaster_Gaming ah yes static server meshing will improve a LOT for sure. Why? Simple to explain as in we can have 2 servers run in say Stanton. Half the ressources or a bit less than half (for the overlapping zones) that are simulated for stanton means lots more computation ressources for said server for npcs and other problems, and that would mean 30 or more fps server side to be the norm imo
Exactly, and its difficult to avoid getting super excited about it. Good times are ahead and not too far off at all imo.
@@DigitalMaster_Gaming hopefully soon, i cant wait to see how good the game gets and how many bugs are actually solved from this
Awesome content!
Thank you!
Could you explain one burning question I have? If the rep layer syncs multiple DGS, does that mean if someone that's standing next to me on a different DGS (and I can't see them) puts a bottle on the ground, I'll just see a bottle pop on the ground without seeing the player?
You'll see them too. Just as the rep layer replicates objects, it does the same with players. We'll all see each other as well across servers, it's just that only one will have authority over your characters simulation.
@@DigitalMaster_Gaming That's awesome! Thank you!
EVERY SINGLE TIME they find a way to improve server performance. Chris steps in with his fat ass and forces the server player cap to be raised. We will never be able to enjoy decent server performance while Chris is involved in development. We are here just to be unpaid stress testers.
It won't be like that I don't think. The idea of server meshing is to alleviate server load
Nice vid
Thank you!
*Promo SM* 😅
Thanks for explaining this ! i had some confusion o7
SF_SpaceBARD
You're welcome!