PaaS High Availability

Azure provides app services and cloud services for hosting managed platforms. Services and applications can be deployed on top of them. They provide flexibility, elasticity, and economies to create and deploy applications. These platforms are managed by Azure and users do not interact with base infrastructure on which they are deployed. They bring in a higher level of abstraction compared to IaaS by letting developers concentrate on their business problem and using the platform to fast track their development and deployment process. This alleviates them to manage, operate, and monitor the base infrastructure. When an application is deployed in-app services or cloud services, Azure provisions virtual machines that are not visible to users. The applications are deployed on these virtual machines and Azure fabric controller is responsible for provisioning, managing, and monitoring them. The fabric controller monitors the status of the hardware and software of the host and guest machine instances. When it detects a failure, it maintains SLAs by automatically relocating the VM instances. When multiple cloud service role instances are deployed, Azure deploys these instances to different fault and update domains.

The previous diagram shows that PaaS services with multiple virtual machines instances deploy these web and worker roles on separate fault domains. Deploying on separate fault domains means deploying on separate racks within a data center. It also means that these services have separate network switches and power supply ensuring that even if one of the racks undergoes maintenance, there is disruption of power supply to the rack or failure of the network switch, there are other instances available to serve customer's requests.