AWS gameliftstreams documentation change
Summary
Added new Windows Server 2022 stream classes (gen6n_medium_win2022, gen6n_small_win2022) and updated quota calculation examples. Enhanced documentation about GPU sharing capabilities.
Security assessment
These changes add new service options and clarify resource allocation mechanics. While tenancy models have security implications, the modifications focus on feature expansion rather than addressing security vulnerabilities or describing security controls.
Diff
diff --git a/gameliftstreams/latest/developerguide/stream-groups.md b/gameliftstreams/latest/developerguide/stream-groups.md index 787e37898..6f114e76f 100644 --- a//gameliftstreams/latest/developerguide/stream-groups.md +++ b//gameliftstreams/latest/developerguide/stream-groups.md @@ -69 +69 @@ More specifically, the GPU service quotas specify the maximum number of GPUs of -When calculating the total stream capacity provided by these GPUs, it's important to remember that multi-tenant stream classes support streaming more than one session per GPU. Therefore, if you're using multi-tenant stream classes in your stream groups, such as `gen6n_high`, you will need to take this into account when determining how the capacity will count against your quota. Single-tenant stream classes, such as `gen6n_ultra` and `gen6n_ultra_win2022`, dedicate one GPU per stream session. +Amazon GameLift Streams measures your service quotas in terms of allocated GPU totals. It is important to remember that some stream classes (such as `gen6n_high` or `gen6n_small`) share a GPU across concurrent sessions. Other stream classes such as `gen6n_ultra` and `gen6n_ultra_win2022` use one full GPU per concurrent session. Therefore, 10 GPUs can be allocated as a MaximumCapacity of 10 on a `gen6n_ultra` stream group, or a MaximumCapacity of 40 on a `gen6n_medium` stream class. @@ -136,0 +137,2 @@ Stream class | Description +`gen6n_medium_win2022` | (NVIDIA, medium) Supports applications with low 3D scene complexity. Runs applications on Microsoft Windows Server 2022 Base. Uses NVIDIA L4 Tensor Core GPU. Resources per application: vCPUs: 8. RAM: 32 GB. VRAM: 6 GB. Tenancy: Supports up to one concurrent stream session. +`gen6n_small_win2022` | (NVIDIA, small) Supports applications with low 3D scene complexity. Runs applications on Microsoft Windows Server 2022 Base. Uses NVIDIA L4 Tensor Core GPU. Resources per application: vCPUs: 2. RAM: 8 GB. VRAM: 3 GB. Tenancy: Supports up to one concurrent stream session. @@ -225,0 +228 @@ Stream class | Description +`gen6n_medium_win2022` | (NVIDIA, medium) Supports applications with low 3D scene complexity. Runs applications on Microsoft Windows Server 2022 Base. Uses NVIDIA L4 Tensor Core GPU. Resources per application: vCPUs: 8. RAM: 32 GB. VRAM: 6 GB. Tenancy: Supports up to one concurrent stream session. @@ -227,0 +231 @@ Stream class | Description +`gen6n_small_win2022` | (NVIDIA, small) Supports applications with low 3D scene complexity. Runs applications on Microsoft Windows Server 2022 Base. Uses NVIDIA L4 Tensor Core GPU. Resources per application: vCPUs: 2. RAM: 8 GB. VRAM: 3 GB. Tenancy: Supports up to one concurrent stream session.