Welcome to a Star Citizen Short explaining Server Player Caps, Object Container Streaming, Server Meshing, how it works and plans for the future.
The Aim with Server Populations or Player Caps is to eventually have everyone be able to play in a single shard or megaserver that everyone can join, though obviously this might not be possible. So as many as possible is the goal with no loading screens beyond you logging in.
Server Populations are partly tied to Client & Server Performance, Efficiency & Technology which is currently a work in progress, Object Container Streaming, Parallel Network Jobs & Server Meshing are all part of this.
Parallel Network Jobs which is currently pegged for Alpha 3.5 is a Server Thread Optimization that will allow Servers to handle a greater number of connections and should see server populations rise because of that or at least should help allow for it.
The limiting factor with the technology is going to the maximum connections a single server can take before dropping below performance thresholds (this will mean that even a small room will have a max population)
And latency of clients connecting to servers, if pings and latency are too high for USA and Europe to play in the same Server then they may have to be split Regionally.
Object Container Streaming is split into 2 parts Client & Server which required a re-engineering of the game engine and it’s something that has caused delays to both Star Citizen and SQ42 originally planned to take 6 months to get to this stage but actually taking 18.
Client Side Object Container Streaming allows for the dynamic seamless loading in or out of assets, textures, locations, ships and entities that are proximate to a player. The System has areas divided into nested containers that every object will sit inside and containers can sit within each other, they contain the information of what to “stream” in basically, this can be the stanton system, planet hurston, lorville and a room in the Hab Block there all being separate containers. CSOCS which premiered in Alpha 3.3 branch has seen a massive framerate rise and performance improvement for the game.
Server Side Object Container Streaming allows the Server to save and loadout areas when not being used, then load them back in when needed. This is for server efficiency but in the case of SQ42 the Client and Server are effectively both on the client PC so it’s still essential.
Long distance Objects can be partly loaded in if there is a reason to, like a ship scans them, this means that some of the entities data that is required can be loaded in BUT only what is required.
Server Meshing takes this one step further by allowing each of these Object Containers to run by an individual server, so this could be individual rooms on a ship BUT the system will dynamically have servers spin up or down based on the population needs for each object container then stitch them together so you can seamlessly move between them.
They want the tech to support 1000s of players on planets at once, though there is no way they could render 1000s of players in a single small dogfight where they all have their own ships.
Some Solutions to problems they foresee encountering they could potentially make high traffic areas have queues or be harder to access when large amount of players are there, this could even be done by increasing docking fees or taxes in an area so players don’t want to do the less profitable missions or land at expensive stations.
If the servers remain segregated in the short term then may still allow players the choice to move to other regional servers so you could play with your USA mates in the UK like we do now. However with a good enough server infrastructure I would hope that pings between these areas wouldn’t be a problem.
They have talked about having servers that share the same areas but are effectively instanced from other players too.
Server Side Object Container Streaming, Various Optimizations and the Groundwork for
Server Meshing are all pegged for 2019, server meshing itself is likely to be here in 2020.
Server Meshing is the final “pillar” they have outlined to be completed for a more full gameplay experience.
All of these features will continue to receive updates, bug fixes and iterative improvements throughout the project and in no way should they be seen as final.
It was stated at CitizenCon’s 2018 Road to Release that the completion of Server Meshing would herald No more Resets for the game, meaning that progress would be permanent.