AWS gameliftstreams documentation change
Summary
Updated list of stream classes for non-Windows and Windows environments, removing older generation classes (gen4n_high, gen5n_ultra, etc.) and focusing on gen6n_ultra variants. Removed specific examples of multi-tenant stream classes.
Security assessment
The changes refine documentation about hardware tenancy models but provide no evidence of addressing a specific security vulnerability. The updates clarify which stream classes use containerization vs direct host execution, but this is operational documentation rather than security remediation.
Diff
diff --git a/gameliftstreams/latest/developerguide/infrastructure-security-tenancy.md b/gameliftstreams/latest/developerguide/infrastructure-security-tenancy.md index 40255d610..987d54444 100644 --- a//gameliftstreams/latest/developerguide/infrastructure-security-tenancy.md +++ b//gameliftstreams/latest/developerguide/infrastructure-security-tenancy.md @@ -15 +15 @@ Within a stream group, resources are reused over time to serve multiple sessions -Non-Windows stream groups with stream classes such as `gen4n_high`, `gen5n_ultra`, `gen6n_ultra` or `gen6n_pro` execute your applications inside of dedicated per-session containers. Each stream session begins with a copy of the application files and an empty user profile folder. When a session terminates, all file system modifications are discarded and all processes launched by your application are terminated as part of container cleanup. +Non-Windows stream groups with stream classes such as `gen6n_ultra` execute your applications inside of dedicated per-session containers. Each stream session begins with a copy of the application files and an empty user profile folder. When a session terminates, all file system modifications are discarded and all processes launched by your application are terminated as part of container cleanup. @@ -17 +17 @@ Non-Windows stream groups with stream classes such as `gen4n_high`, `gen5n_ultra -Windows-based stream groups with stream classes such as `gen4n_win2022`, `gen5n_win2022`, `gen6n_ultra_win2022`, or `gen6n_pro_win2022` execute your applications directly on the host operating system. Each stream session begins with a copy of the application files and an empty user profile folder. When a session terminates, the user profile folder and application folder are fully reset. Sub-processes launched by your application are terminated. If your application modifies files outside of the user profile folder and the application folder, or modifies the system registry, then those changes might persist across multiple sessions. +Windows-based stream groups with stream classes such as `gen6n_ultra_win2022` execute your applications directly on the host operating system. Each stream session begins with a copy of the application files and an empty user profile folder. When a session terminates, the user profile folder and application folder are fully reset. Sub-processes launched by your application are terminated. If your application modifies files outside of the user profile folder and the application folder, or modifies the system registry, then those changes might persist across multiple sessions. @@ -23 +23 @@ For any stream group configuration, the underlying compute resources and operati -Stream groups are either single-tenant or multi-tenant, depending on your selection of stream class. Multi-tenant stream classes such as `gen4n_high` or `gen5n_high` share one GPU across multiple simultaneous sessions. In this context, multi-tenancy refers to running more than one session at a time on the underlying hardware. The hardware is still dedicated to your stream group and is not shared across stream groups or with other AWS customers. +Stream groups are either single-tenant or multi-tenant, depending on your selection of stream class. Multi-tenant stream classes share one GPU across multiple simultaneous sessions. In this context, multi-tenancy refers to running more than one session at a time on the underlying hardware. The hardware is still dedicated to your stream group and is not shared across stream groups or with other AWS customers.