9min

Gameye Server Orchestration

Current Gameye regions and locations

A region is a collection of different geographical locations. Each location inside a region can consist of multiple providers. By default, Gameye automatically selects the best location inside a region to create a session based on several parameters such as availability, latency, and price.

With built-in redundancy by design, Gameye can switch off individual locations or providers if there is some kind of disruption. This ensures that no further sessions are spun up in a location that may be suffering from an outage like a denial of service (DDoS) attack. New sessions are automatically deployed on healthy providers or locations. When sending your API payload, you can choose a specific location(s) by using the location's name or have the system automatically select a location by using the region tag.

Below is a list of regions and locations that Gameye can currently deploy game server sessions into. The naming convention is based on the Gameye API. The region needs to be entered exactly as it is described here.

Region

Locations

eu-west

Amsterdam Frankfurt London Paris

eu-south

Madrid

na-east

Washington DC Ashburn New York Manassas Florida

na-central

Dallas Chicago

na-west

San Jose Los Angeles Seattle

sa-east

Sao Paulo

asia-east

Singapore Hong Kong

asia-northeast

Tokyo Seoul

oce-east

Sydney



During the integration phase, your account manager will discuss which regions you wish to deploy to and what kind of availability there is.

If there is a location that is not currently on this list, that is potentially one you would be interested in please let us know! Exotic locations such as Seoul and various locations in China can be made available if required.

Region-specific request example:

JSON
|

Location-specific request example:

JSON
|



Active Steering

When using regions for session creation, instead of specific locations, Gameye automatically redirects the match to the best possible location, provider and compute instance type (bare metal, cloud or edge).

For example, during the daytime, much of the load can be spread across bare metal servers. While in the evenings or during a special event, cloud capacity can be ramped up quickly to deal with a peak increase in concurrent sessions.

Automatic routing around network disruptions

In the case of a network disruption, the provider or location at which the disruption is detected will have its sessions drained and then will be removed from the machine pool. The selection process for placing new sessions will no longer use this provider or location. When a large amount of capacity is drained at once, extra capacity is automatically brought online at other providers.

Mitigate single provider capacity issues

When a location inside a region is out of capacity (bare metal or cloud) it will automatically be ignored in the selection process and sessions are created on the next best location.