Multi-Cloud and Bare Metal Game Server Hosting for Launch Traffic
A multiplayer game's traffic has two shapes. There is the baseline, the number of servers the game needs every day once things settle, and there are the peaks, such as launch weekend, a new season or a free weekend, when demand can be many times higher for a short while. Hosting that suits one shape rarely suits the other, which is why many large games now run on a mix of bare metal and more than one cloud.
Why is hosting a game on one cloud expensive?
Public cloud is very good at giving you capacity quickly, and you pay for that flexibility. If every server a game runs, including the steady baseline it needs every hour of every day, sits on on-demand cloud instances, the studio pays a premium for flexibility it is not using most of the time. Network egress charges add to that, because game servers send data to players constantly.
A single provider also means a single point of failure, and a single pool of capacity in each region. During a big launch, the instance type you rely on can run short in the region where your players are.
What does bare metal do well?
Bare metal servers are physical machines dedicated to your game. They offer consistent performance with no noisy neighbors, a lower cost per core for capacity you use all the time, and predictable network costs. They take longer to add than cloud instances, so they suit the baseline: the servers you know you will need for months.
What does multi-cloud add?
Cloud capacity from several providers gives you a deep pool to draw on when demand rises, in more regions than any single provider covers well. If one provider runs short of capacity or has an outage, new matches can start somewhere else. Cloud instances can be added in minutes and released when the peak passes, so you only pay for the spike while it lasts.
How does a hybrid setup work at launch?
In a hybrid setup, the orchestration layer treats bare metal and cloud as one fleet. New matches fill bare metal first, because it is the cheapest capacity you already have. When demand rises past what bare metal can handle in a region, the orchestrator starts servers on cloud. When the spike passes, cloud servers are drained and released, and the game settles back onto bare metal.
For launch, studios usually size bare metal for the baseline they expect a few weeks after release and plan cloud for everything above it. As the game finds its level, more of the steady load can move onto bare metal to bring costs down.
What should you check before choosing a hybrid host?
- Whether game servers perform the same on bare metal and on each cloud, so players do not notice which one they are on
- How quickly the orchestrator can add cloud capacity in each region you need
- How builds are deployed and rolled back across every provider at once
- Whether DDoS protection covers every provider and region
- How costs are reported, so you can see the split between baseline and burst
- Who is watching the fleet around the clock, and how quickly they can act
How Multiplay helps
Multiplay hosts game servers on bare metal and multiple clouds, in more than 40 regions, managed as a single fleet. Studios get bare metal efficiency for their everyday load and cloud capacity for launches and events, without having to run the orchestration themselves. We have hosted more than 600 games, and our engineers support them around the clock. For planning a launch on this kind of setup, our guide on peak load testing for multiplayer game servers is a good next step.
Frequently asked questions
What is hybrid game server hosting?
Hybrid hosting runs game servers on a mix of bare metal and cloud, managed as one fleet. Bare metal covers the steady baseline and cloud covers spikes.
Is bare metal cheaper than cloud for game servers?
For capacity you use all the time, bare metal usually costs less per core and has more predictable network costs. Cloud is better value for short spikes because you only pay while you use it.
Why use more than one cloud provider for game servers?
More than one provider gives you more capacity in each region, wider regional coverage and a fallback if a provider has an outage or runs short of capacity.