Video - CIG QnA on Server Meshing, taking questions now

Lorddarthvik

Space Marshal
Donor
Feb 22, 2016
2,750
9,511
2,860
RSI Handle
Lorddarthvik
I know we won't get an answer to When, so I'd like to know the Why not right now. What are the blockers? Is it possible to throw more resources at it (like needing to refactor loads of stuff again and it's just tons of mindless drone work), or is it fundamental issues that are yet to be figured out, in which case more resources means nothing.

Another reason why they might not want to release this or throw more resources at it is that everything is else is not ready for it yet. They don't want to throw more resources at it cos even if it's ready sooner, it could make no use of the current state of the game. I always assume they are way further ahead but with CitCon I'm starting to doubt that. We might be really seeing the exact state they are at which is like WTF, what were they doing for the last 2 year if not concentrating on bringin Server Meshing online and everything connected to that??
 

maynard

Space Marshal
May 20, 2014
5,124
20,290
2,995
RSI Handle
mgk
[RANT]
the holdup is and has always been 'the combinatorial explosion'

as the number of ships/players on a shard increases, the number of objects that need to be tracked increases exponentially

Eve Online
, with its ultra-simplistic ships and flight physics, can now support lag-free fights with hundreds of ships in the same system

but it took years of development to get to that point

when the fights escalate into the thousands the lag is epic

I will be thrilled when Star Citizen can support fights between a couple of mixed 25-ship fleets in one instance

progress beyond that will be slow

if CIG waits until Object Streaming is perfect to release Star Citizen I won't live long enough to ever play
[/RANT]
 

Bambooza

Space Marshal
Donor
Sep 25, 2017
5,689
17,911
2,875
RSI Handle
MrBambooza
I know we won't get an answer to When, so I'd like to know the Why not right now. What are the blockers? Is it possible to throw more resources at it (like needing to refactor loads of stuff again and it's just tons of mindless drone work), or is it fundamental issues that are yet to be figured out, in which case more resources means nothing.

Another reason why they might not want to release this or throw more resources at it is that everything is else is not ready for it yet. They don't want to throw more resources at it cos even if it's ready sooner, it could make no use of the current state of the game. I always assume they are way further ahead but with CitCon I'm starting to doubt that. We might be really seeing the exact state they are at which is like WTF, what were they doing for the last 2 year if not concentrating on bringin Server Meshing online and everything connected to that??
[RANT]
the holdup is and has always been 'the combinatorial explosion'

as the number of ships/players on a shard increases, the number of objects that need to be tracked increases exponentially

Eve Online
, with its ultra-simplistic ships and flight physics, can now support lag-free fights with hundreds of ships in the same system

but it took years of development to get to that point

when the fights escalate into the thousands the lag is epic

I will be thrilled when Star Citizen can support fights between a couple of mixed 25-ship fleets in one instance

progress beyond that will be slow

if CIG waits until Object Streaming is perfect to release Star Citizen I won't live long enough to ever play
[/RANT]

I have a feeling that most of CIG's resources are dedicated to completing SQ42. I also know that most computer game companies have very few programmers as most of their staff needs are with art assets and designers putting together the game world. Of the few programmers most will be tied up with crafting game logic as most are not skilled enough to make fundamental changes to the game engine.

The biggest issue has always been that the game engine while not bad for the initial concept of a light multiplayer mostly solo game as originally pitched lacks the means to support a large open world MMO like what StarCitizen has become. While the technology is nothing new ( Amazon AWS supports and provides nice services for massively distributed server clusters for microservices) but I can see it not easy to fundamentally change the server code that was originally designed to support loading a single level and supporting 20 players in it to one that dynamically loads an area linking all of the assets pruning the tree and participating with back end services. And this is not the only engine code change as they are also working towards planet tech, Gen12 graphics, lighting, cloud tech, and while only planet tech (like outpost ground deformation, rivers) impacts the server side it does impact the number of devs available for any given task. I am also sure while we are impacted by the lack of server scaling it's not as much of a priority within CIG as it doesn't impact SQ42.

Eve-online had its own issues and it too struggled with grasping the end result impact of using python and TCP packets for networking. They have since rewritten a lot of the back-end server code as well as allowed time dilation to help deal with server tick slowdowns.
 
  • o7
Reactions: Lorddarthvik
Forgot your password?