AWS Series: Reference Architecture #1 — Web Application

Case Study: A small start-up firm is on a journey and got their Series A funding approved with minimal MVP based on applications running on the Laptop. Now they want to start building on the application, before doing so, they want to right away build the base properly with the ability to secure, scale, durable, fault tolerant, highly available, self-healing etc. and proceed to pitch to the Venture Capitalists within the next few months for Series B funding.

Assumptions:

Considering a startup firm’s data and customer base and to provide the ability to scale,

• Initial application stack can be anything LAMP / MERN etc. and this reference architecture is technology agnostic.
• Data Input for Phase 1 is from the Users only.
• Data Warehousing & Big Data is not in the scope for current phase, will be part of next phase.
• Migration from On-Premise to Cloud is not part of the scope.
• It is assumed that Business Functionality can be implemented in Micro Services and packaged as Container.
• Digitization of current On-Premise solution to Cloud Native solution is part of the current scope.

Key considerations:

• The solution to include all key layers of Security, Infrastructure & Network, Sourcing / Streaming & Ingestion (as applicable), Storage, Application, Database, Consumption, Monitoring & Messaging Layers and CI / CD (container / micro services design).
• As the Users increases gets spread-out, it must be a low latency and seamless scaling just with configurations.

High Level Technical Architecture:

Detailed Technical Architecture:

Solution Detail:

High Level Target Operating Model:

Solution Details for Target Operating Model:

New / Changed Assumptions: In future if Data Warehousing or Data Lake is required and if External Integration is sought, then the solution can be extended without major changes to Phase 1 implementation.

New / Changed Components:

Budget (Architecture based assumption):

Summary:

The proposed architecture is a paper architecture solution with a startup that starts small with the ability to grow with low-cost and easy-to use infrastructure that will enable you to securely build, scale, high performance, efficient, elastic, highly available, fault tolerant and recoverable architecture that will allow you to organically grow. The first reference architecture should be a target operating model and should be as best a blue-print and technology agnostic. Any technology that you fit in should work seamlessly using this architecture.

Alternative Reference architectures:

• Use of EKS instead of ECS.
• The above accepts only user data — improve it to include Data pipeline and the complete stack.
• Future possibilities — Recommendation Engine; Elastic Search etc.

For other articles refer to luxananda.medium.com.