- Joined
- Feb 15, 2016
- Messages
- 239
- Reaction score
- 58
Request:
Add TPS as a performance metric; splitting server performance metrics away from network performance metrics.
Currently:
FPS = client-side rendering (e.g. GPU) performance
PING = combined network and server-side performance metrics
Result of proposed:
Provide players with a more clear illustration of lag causes: is it truly an issue server-side, or is the lag caused by a client's network connection (e.g. high latency or packet-loss)?
Suggested Method:
StarMade already has the concept of server ticks; extrapolate this value out, and make it visible client-side. For instance, how many ticks is a particular entity consuming?
Rationale:
Right now, ideal lag levels are a crap-shot: aside from acquired experience, admins don't have a good idea of what they're aiming for, what works OK for a busy server, etc. Simply working to avoid high millisecond related log entries, without a good grasp upon their overall impacts, can be quite a frustrating process.
With a TPS value visible client-side, everyone now has better visibility of possible lag causes, and most-importantly, if a server is overloaded/underpowered. It also provides for a metric upon which admins can begin targeting performance-wise.
Continuing to report the current co-mingled application/network metric "ping", would enable players and admins to contrast values against "TPS"; as a means of comparing server-side performance vs network connectivity.
TPS would also help reinforce a point, that when players connect from half-way across the globe, lag is at least in part due to their connection choices: this means players are less likely to blame "StarMade" for being awful, and instead to shift their attention towards the laws of physics...more readily encouraging the player to choose a closer server.
Add TPS as a performance metric; splitting server performance metrics away from network performance metrics.
Currently:
FPS = client-side rendering (e.g. GPU) performance
PING = combined network and server-side performance metrics
Result of proposed:
Provide players with a more clear illustration of lag causes: is it truly an issue server-side, or is the lag caused by a client's network connection (e.g. high latency or packet-loss)?
Suggested Method:
StarMade already has the concept of server ticks; extrapolate this value out, and make it visible client-side. For instance, how many ticks is a particular entity consuming?
Rationale:
Right now, ideal lag levels are a crap-shot: aside from acquired experience, admins don't have a good idea of what they're aiming for, what works OK for a busy server, etc. Simply working to avoid high millisecond related log entries, without a good grasp upon their overall impacts, can be quite a frustrating process.
With a TPS value visible client-side, everyone now has better visibility of possible lag causes, and most-importantly, if a server is overloaded/underpowered. It also provides for a metric upon which admins can begin targeting performance-wise.
Continuing to report the current co-mingled application/network metric "ping", would enable players and admins to contrast values against "TPS"; as a means of comparing server-side performance vs network connectivity.
TPS would also help reinforce a point, that when players connect from half-way across the globe, lag is at least in part due to their connection choices: this means players are less likely to blame "StarMade" for being awful, and instead to shift their attention towards the laws of physics...more readily encouraging the player to choose a closer server.
Last edited: