View a markdown version of this page

ADVPERF02-BP02 Consider containerization for scalability, low latency, and cost optimization - Video Streaming Advertising Lens
Implementation guidanceKey AWS servicesResources

ADVPERF02-BP02 Consider containerization for scalability, low latency, and cost optimization

Adopt containerization as a strategy to operate at scale with low latency and cost optimization. Evaluate the various options of running container workloads on AWS.

Implementation guidance

Consider containerization, which helps improve application performance and helps scaling needs for adtech workloads, due to the following benefits:

  • Faster startup times: Containers share the host OS kernel and start only the necessary processes, so they can start almost instantly compared to a full virtual machine (VM) startup. This makes scaling up and down faster.

  • Lower resource usage: Containers require fewer resources than VMs, as there is no guest OS overhead. More efficient resource usage leads to cost optimization and the ability to run more container instances per host.

  • Portability across environments: Container images can run on any infrastructure due to standardized runtime without need to re-optimize for different environments.

  • Scaling and availability: Container orchestrators (for example, Amazon EKS) help to scale containerized apps, provide high availability, and improve performance under heavy loads.

  • Isolation: Containers isolate processes and resources per application, reducing noisy neighbor issues on multi-tenant hosts for more predictable performance.

  • Utilization: Higher density of containers per host allows full utilization of available resources, especially with auto scaling.

  • Microservices: Decomposing monoliths into containerized microservices reduces interdependencies and allows independent scaling.

Key AWS services

Resources