Well Architected Framework — Cloud Platform Management vs. Cloud Data Platform Management

Recently, I was looking into AWS Well-Architected Framework for some understanding which I was proposing to the client and its importance in cloud architecture. Looking at the broadness of its pillars, not just architecture, but we must look at how these maps into various activities that you do in cloud. Be it, managing the entire cloud platform or managing only the data platform or managing an application on it.

1. If you are doing the Cloud Platform Management these pillars apply to it. Platform teams must watch out for them.
2. If you are building Cloud Data Platform and managing the same these pillars still apply. The cloud data architect must watch out for them.
3. Any other solutions / applications, so that application architects can look into the same.

Let us look at what these pillars are and quickly see the linkage between them with Overall cloud platform management and cloud data platform management. The terminology for these pillars slightly changes when you look into specific area.

AWS Well-Architected Pillars (icon source AWS)

1. Reliability — Deals in ensuring cloud is performing as intended and consistently. Monitoring & Observability practices for Cloud Platform and Highly Available Data Architectures with Data Lineage and DQ Monitoring in case of Cloud Data Platform will constitute to this pillar.
2. Security — Deals in protecting information and managing system integrity. Cloud Security in case of platform will deal with network security, firewalls and user access controls. Data Security in case of Cloud Data Platform will focus on Data Access controls, Data Classification, Data Privacy, Encryption and Data Retention.
3. Operational Excellence — Deals with running and monitoring systems. In case of Cloud platform, it will refine the processes and procedures and DevOps practices for automation. For Cloud Data Platform, it will use CI/CD for data pipelines and deployments and collect operational metadata to verify the health of analytics workloads.
4. Performance Efficiency — Efficiency is using IT and Computing resources. For Cloud Platform, it will ensure Cloud Utilization is done appropriately. For Cloud Data Platform, it will deal with optimal compute and storage solutions to efficiently access the data.
5. Cost Optimization — Helps in avoiding unnecessary costs by controlling resources. FinOps practices with automated solutions will help in Cloud Platform Management. For Cloud Data Platform, ensuring optimal TCO models are followed in choosing compute and storage solutions/tools will eventually contribute on cost optimization.
6. Sustainability — Helps in minimizing environmental impacts using cloud workloads. For Cloud Platform, verifying the Carbon footprint and encouraging to use cloud resources in more sustainable manner will help. For Cloud Data Platform, Data Minimization, Data Retention to get rid of redundant data, preventing unnecessary data movements and optimal data modelling for efficient data retrieval will contribute to sustainability goals.

Conclusion

In short, its good practice to ensure whatever we are doing aligns with Well-Architected pillars and identify which pillar the particular act is contributing to. Whether you are part of platform team or data team, this mapping between Cloud Platform and Cloud Data Platform will help you understand that and design better solutions in cloud which are well-architected!

On high level below diagram summarizes the same.

Cloud Platform Management vs Cloud Data Platform Management © Manish Kulkarni

More Read:

1. AWS Well-Architected — Build secure, efficient cloud applications (amazon.com)
2. The Five Pillars of the Framework — AWS Well-Architected Framework (amazon.com)
3. Data Analytics Lens — Data Analytics Lens (amazon.com)