Overview of Resilience in the Cloud Technology Stack

Pivoting away from the traditional on-prem infrastructure, technology stacks have become more and more homogenous. This has created interesting ways to support resilient postures in the cloud. Resiliency for an on-premise datacenter leverages both software and hardware architectures to achieve that posture. Cloud has shift that paradigm by creating resilient services,  effectively shifting the technical knowledge away from implementation towards capabilities. Services have been abstracted to provide hardware capabilities through the virtualization layers. The software riding on top of the virtualization layers can also leverage resiliency patterns.

 

The workloads riding on the cloud’s technical stack have several distinct capabilities that are not found in the traditional datacenter. Capabilities such as shifting workloads across the virtualized layer inherently affords an application that level of resiliency. In turn, if the on-prem infrastructure, less a catastrophic failure (i.e. US east coast brown out outage a few years ago) has an error and a node drops out, another node can assume that failed footprint.

 

As we move up the stack, the to the application layer resiliency becomes more of a story about managing the transaction. These transactions have need for persistence and while the load can be shift across the virtualization layer, the application needs to be explicitly setup to support the transaction. Again, cloud infrastructure providers have some services in this space like AWS’s Route53 that help to support the application layer, but software level resilient patterns would still need consideration.