Neu
Hey everyone! I’ve been a long-time fan of The Forest, and more recently Sons of the Forest, and like many of you, I’ve spent way too many hours perfecting my defensive perimeters and trying to keep the local "neighbors" from knocking down my front door.
I’m currently running a dedicated server for a group of about six friends, and we’ve hit that inevitable wall that seems to plague every long-term save: the dreaded late-game desync. It’s a point I’ve seen mentioned quite a bit in the technical sub-forums here—the idea that once your log count reaches a certain threshold and you have thousands of individual physics-enabled assets scattered around the map, the world-state starts to crumble for everyone except the host.
I’m a bit of a hardware enthusiast by trade, so my instinct is always to throw more "plumbing" at the problem. I recently overhauled my entire home lab to support this server. I’ve moved the game files onto a high-IOPS array and, to ensure there was absolutely zero local bottleneck between the server and the rest of my network, I actually upgraded the link using some enterprise-grade network transceivers. I’m currently running a high-speed fiber backplane with 32-Gigabit transceivers just to make sure the data transfer between my primary switch and the server node is essentially instantaneous.
My personal insight from years of tinkering is that we often blame CPU cycles for lag when, in reality, it's frequently micro-bursts of network congestion or high-latency "handshakes" between the hardware layers that cause those weird hitches. I figured a 32-Gigabit pipe would be more than enough to handle anything the game could throw at it, even during a heavy raid with thirty cannibals and fire flying everywhere.
However, even with this massive amount of bandwidth and near-zero latency on the hardware side, my friends are still reporting that "ghosting" effect where logs disappear or defensive walls don't show up as built on their end. It’s incredibly frustrating because, on paper, my network infrastructure is basically a mini-data center. It leads me to wonder if the bottleneck isn't the physical hardware or the network devices at all, but rather how the game’s engine serializes that data.
Has anyone else here tried to "over-spec" their networking hardware to solve multiplayer stability? I’m curious if there’s a point of diminishing returns where the game’s netcode simply can’t utilize the throughput of something like a 32-Gigabit link, or if there’s a specific setting in the DedicatedServer.cfg that I’m overlooking to allow for better asset synchronization.
Do you think we’re just at the mercy of the engine's architectural limits once a base gets too large, or is there a way to actually leverage high-end hardware to keep the experience smooth for a larger group?
I’m currently running a dedicated server for a group of about six friends, and we’ve hit that inevitable wall that seems to plague every long-term save: the dreaded late-game desync. It’s a point I’ve seen mentioned quite a bit in the technical sub-forums here—the idea that once your log count reaches a certain threshold and you have thousands of individual physics-enabled assets scattered around the map, the world-state starts to crumble for everyone except the host.
I’m a bit of a hardware enthusiast by trade, so my instinct is always to throw more "plumbing" at the problem. I recently overhauled my entire home lab to support this server. I’ve moved the game files onto a high-IOPS array and, to ensure there was absolutely zero local bottleneck between the server and the rest of my network, I actually upgraded the link using some enterprise-grade network transceivers. I’m currently running a high-speed fiber backplane with 32-Gigabit transceivers just to make sure the data transfer between my primary switch and the server node is essentially instantaneous.
My personal insight from years of tinkering is that we often blame CPU cycles for lag when, in reality, it's frequently micro-bursts of network congestion or high-latency "handshakes" between the hardware layers that cause those weird hitches. I figured a 32-Gigabit pipe would be more than enough to handle anything the game could throw at it, even during a heavy raid with thirty cannibals and fire flying everywhere.
However, even with this massive amount of bandwidth and near-zero latency on the hardware side, my friends are still reporting that "ghosting" effect where logs disappear or defensive walls don't show up as built on their end. It’s incredibly frustrating because, on paper, my network infrastructure is basically a mini-data center. It leads me to wonder if the bottleneck isn't the physical hardware or the network devices at all, but rather how the game’s engine serializes that data.
Has anyone else here tried to "over-spec" their networking hardware to solve multiplayer stability? I’m curious if there’s a point of diminishing returns where the game’s netcode simply can’t utilize the throughput of something like a 32-Gigabit link, or if there’s a specific setting in the DedicatedServer.cfg that I’m overlooking to allow for better asset synchronization.
Do you think we’re just at the mercy of the engine's architectural limits once a base gets too large, or is there a way to actually leverage high-end hardware to keep the experience smooth for a larger group?