Abstract: A branded fleet server system includes a pre-assembled third-party computer system integrated into a chassis of the branded fleet server system. The pre-assembled third-party computer system is configured to execute proprietary software that is only licensed for use on branded hardware. A virtualization offloading component is included in the server chassis of the branded fleet server along with the pre-assembled third-party computer system. The virtualization offloading component acts as a bridge between the pre-assembled third-party computer system and a virtualized computing service. As such, the virtualization offloading component manages communications, security, metadata, etc. to allow the pre-assembled computer system to function as one of a fleet of virtualization hosts of the virtualized computing service.

Abstract: Techniques for failover management using availability groups are described. According to some embodiments, customers can define one or more availability groups within their infrastructure. Each availability group may be associated with one or more rules which describe how the availability group should fail over. The availability group definition may define the resources included in the availability group, a target region to fail over to, and capacity reservations for the availability group in the target region. In some embodiments, the definition may also include a replication frequency for storage resources which defines how often data is backed up. Once a customer as created an availability group the customer is then able to either create resources in that availability group or associate already existing resources with the availability group. An availability group management service can ensure that data, capacity reservations and network infrastructure are defined and/or replicated to the target location.

Abstract: A branded fleet server system includes a pre-assembled third-party computer system integrated into a chassis of the branded fleet server system. The pre-assembled third-party computer system is configured to execute proprietary software that is only licensed for use on branded hardware. A baseboard management controller (BMC) included in the server chassis couples with the pre-assembled computer system via one or more cables coupled to ports of the pre-assembled computer system. The BMC enables remote control of the pre-assembled computer system, such as remote power on and power off. An electro-mechanical device, such as a solenoid, presses a power button of the pre-assembled computer system based on control signals from the BMC to change a power state of the pre-assembled computer system.

Abstract: A branded fleet server system includes a pre-assembled third-party computer system integrated into a chassis of the branded fleet server system. The pre-assembled third-party computer system is configured to execute proprietary software that is only licensed for use on branded hardware. A baseboard management controller (BMC) included in the server chassis couples with the pre-assembled computer system via one or more cables coupled to ports of the pre-assembled computer system. The BMC enables remote control of the pre-assembled computer system, such as remote power on and power off. Also the BMC may enable automatic and remote software and/or firmware updates to be performed at the pre-assembled computer system.

Abstract: Techniques are described for enabling users to create reservations for compute capacity at a service provider network intended for use in disaster recovery scenarios. A service provider can use separate geographic areas (sometimes referred to as “regions” or “availability zones”) defined by the service provider network to enable users to share reservations of compute capacity intended for use in disaster recovery scenarios. A disaster recovery launch reservation (DRLR) is a reservation of a given unit of compute capacity that is unique to a source geographic area (that is, a defined area of the service provider network containing computing resource(s) for which a user desires to obtain backup compute capacity) and a destination area (that is, a defined area of the service provider network at which the reserved compute capacity is located), where other users can reserve the compute capacity relative to other source geographic area-to-destination geographic area mappings.

Abstract: Techniques for private network mirroring are described. Users can select one or more existing private networks to be mirrored in the same or different network area. Any network configuration changes made in the selected private network (e.g., “master” private network) can be propagated automatically to the mirror private network. This enables users to utilize resources in another network area for disaster recovery, ensuring that the network configuration of the mirror private network is consistent with the master private network through real-time updates. Additionally, users managing infrastructure that includes multiple private networks can select one master private network and propagate configuration changes to other private networks, reducing management overhead incurred by these multi-private network installations.

Abstract: Techniques for failover management using availability groups are described. According to some embodiments, customers can define one or more availability groups within their infrastructure. Each availability group may be associated with one or more rules which describe how the availability group should fail over. The availability group definition may define the resources included in the availability group, a target region to fail over to, and capacity reservations for the availability group in the target region. In some embodiments, the definition may also include a replication frequency for storage resources which defines how often data is backed up. Once a customer as created an availability group the customer is then able to either create resources in that availability group or associate already existing resources with the availability group. An availability group management service can ensure that data, capacity reservations and network infrastructure are defined and/or replicated to the target location.

Abstract: Techniques for private network mirroring are described. Users can select one or more existing private networks to be mirrored in the same or different network area. Any network configuration changes made in the selected private network (e.g., “master” private network) can be propagated automatically to the mirror private network. This enables users to utilize resources in another network area for disaster recovery, ensuring that the network configuration of the mirror private network is consistent with the master private network through real-time updates. Additionally, users managing infrastructure that includes multiple private networks can select one master private network and propagate configuration changes to other private networks, reducing management overhead incurred by these multi-private network installations.