In gaming, user experience is measured in milliseconds. If your server is abroad, your player feels it with every move, and in competitive titles that translates directly into lost players.
portvMind servers are located in Istanbul and Ankara. The physical distance to your players in Turkey shrinks, and latency drops noticeably compared with infrastructure serving from regions abroad. In-region latency is below 2 ms. When session load rises you grow the server pool in minutes, and when the peak passes you pull it back.
In competitive, real-time games a difference of a few milliseconds directly affects the player experience; the player leaves the game without knowing why.
On portvMind: Servers are located in Turkey, physically close to your players. In-region latency stays below 2 ms; traffic between servers flows at this speed and free of charge.
During launches and events the number of concurrent players rises unpredictably; the difference between evening and daytime can be many times over.
On portvMind: You grow the server pool the moment you need it. If you run on Kubernetes, the replica count scales automatically with session load.
Whether a new game will take off cannot be known in advance. Investing in hardware on that basis means idle capacity if it doesn’t, and insufficient capacity if it does.
On portvMind: With the pay-as-you-go model you adjust capacity to real demand. You don’t have to make a commitment, and you can change the scale freely after launch.
Profile, progress and matchmaking data must be read with low latency; a slowdown in this layer directly breaks the in-game flow.
On portvMind: With full-flash Block Storage there is no disk-induced latency in data access. Game assets and patch files sit in Object Storage.
Game servers, management interfaces and the data tier should not sit on the same flat network.
On portvMind: With Router you split the tiers into separate virtual network segments; management interfaces stay closed to the internet and access is controlled with security groups.
Shutting down a server during patches and version switches means dropping players from their sessions.
On portvMind: With Floating IP you prepare the new server and move the address to it when it is ready. You don’t wait for DNS propagation.
| Service | Role |
|---|---|
| Compute | Low-latency compute power for game server instances |
| Kubernetes | Container architecture that autoscales with session load |
| Load Balancer | Distributing player traffic across servers |
| Floating IP | Fixed connection address and uninterrupted version switches |
| Object Storage | Game assets, patch files and records |
| Router | Separating game, management and data tiers at the network level |
| Network Data Transfer | Free in-region traffic and flow visibility |
This combination is a starting point; we adapt it together to your game architecture.
Your game servers run on Compute behind a Load Balancer. The player connects to a single address; the load balancer decides which server to route to and automatically takes unresponsive servers out of the pool.
Router separates three tiers: the server tier open to players, the data tier reachable only from the internal network, and the management tier completely closed to the internet. Player profile and progress data sit on Block Storage, game assets and patch files in Object Storage.
During events or launches you add servers to the pool. If you run on Kubernetes this increase happens automatically. When the peak passes you scale down; because metering is based on the time resources are running, savings start within the same hour.
Traffic between servers is not charged as long as it stays within the same region; this removes cost concerns in architectures that need constant internal communication, such as matchmaking and session synchronization.
Most studios start with a single server: a Compute instance, a Floating IP and an Object Storage bucket for game assets. The first $45 of free usage is enough to test this setup with real player traffic.
As your player count grows you add a Load Balancer and grow the pool. You don’t have to build your architecture big from the start; the infrastructure grows with you.
If you want to measure latency yourself, you can open a test server and take measurements from the regions where your target audience is. You can ask our support team for help with the setup.
In-region latency is below 2 ms. Access time to your players in Turkey is noticeably lower than with infrastructure serving from regions abroad; the exact value varies with the player’s internet service provider and location.
Istanbul (TR-IST-01) and Ankara (TR-ANK-01). You can build a redundant setup by using both regions together.
Yes. You are not bound by a commitment; you can increase or decrease capacity at any time and metering is based only on the time a resource is running.
With Floating IP you prepare the new server in advance and move the address to it when it is ready. That way you switch over without waiting for DNS propagation.
In Object Storage. Your servers in the same region access these files free of charge, and you don’t need to do capacity planning as file volume grows.
Open your account in minutes and try the first $45 of usage for free. If you want to measure latency or plan your architecture, write to our team.