Apr 19, 2023
Cloud Performance Engineering: Maximize Your Investment
Posted by Srivatsan PA
Cloud computing is disruptive, promising scalable computing power at lower costs. However, the complexity of cloud-based systems makes performance engineering crucial. Relying solely on elastic computing without understanding the application’s architecture can hide performance bottlenecks, increase costs, and defeat the purpose of cloud deployments.
Different patterns of cloud migrations, from re-hosting on Infrastructure as a Service (IaaS) platforms to rebuilding with significant changes, require adapted performance engineering strategies to ensure the migrated system can handle the anticipated use.
Performance test strategies to validate the responsiveness, reliability, and stability of the cloud infrastructure include tests such as
- Load Test (expected average & peak users)
- Stress Test (extreme load)
- Latency test (to measure the time taken for moving data messages between two points within a cloud network)
- Capacity Test (identifying and benchmarking the maximum amount of traffic or load that the cloud system can handle effectively)
- Targeted infrastructure test (component or layer of an application is isolated and tested for the ability to deliver the required performance)
- Soak test (exposed to heavy traffic for an extended duration to validate its behavior in the production environment)
- Failover test (can also be conducted to determine system’s ability to call in additional resources during heavy traffic and usage to ensure end users’ experience is not affected.)
When re-hosting an application, load testing should focus on the overall environment, checking for sufficient resources and adequate provisioning of VMs in terms of CPUs, memory, and disk space.
When transitioning from a server-based architecture to a server-less approach & rebuilding with changes, load testing needs to focus on application performance at access points, as there is no control over CPU or memory allocation in server-less functions, which is done by the service provider.
Performance engineers face challenges with cloud systems, such as complex troubleshooting across the delivery chain, difficulty in achieving test reproducibility, and understanding the behaviour of server-less functions for optimal performance.
Moving to the cloud has numerous advantages for businesses. It enables cost optimization, improved service quality, and expanded service offerings. However, comprehensive performance engineering & performance testing should be an integral part of the migration process to ensure optimal infrastructure costs and high-quality service delivery to users.