This Guidance can be scripted using an AWS CloudFormation template. To integrate and deploy changes, use a source control system to manage code and other artifacts, such as version-controlled CloudFormation templates of your infrastructure. When responding to incidents and events, you can send logs directly from your application to Amazon CloudWatch using the CloudWatch Logs API, or send events using AWS SDK and Amazon EventBridge.
Overview
This Guidance shows how payment institutions can use new data types, streaming data, data lakes, and machine learning (ML) on AWS to make rapid credit decisions, reach new consumers, and deliver on customer expectations of multiple payment options.
How it works
Use this high level reference architecture to build a Buy Now Pay Later (BNPL) platform on AWS with near real-time lending decisions.
Step 1
Customer selects an item for purchase and prepares to check out.
Step 2
The request is routed to a separate page which takes the user through the BNPL flow. This web page can be built using AWS Amplify, which is tightly integrated with Amazon Cognito for new customers to sign up, and existing customers to sign in.
Step 3
New authenticated customers are verified by sending a notification to their devices using Amazon Pinpoint.
Step 4
Authenticated clients make API calls to AWS AppSync using valid JWT tokens generated by Amazon Cognito.
Step 5
AWS AppSync uses resolvers to make direct calls to different microservices. HTTP resolvers connect to REST endpoints of the user service. An AWS Lambda resolver directs calls to the credit service in a virtual private cloud (VPC).
Step 6
The communication between the resolvers and the HTTP endpoints are protected with temporary AWS Identity and Access Management (IAM) credentials based on assumed IAM roles. The JSON Web Token (JWT) specific to the authenticated user is also forwarded to each microservice.
Step 7
A Lambda function is invoked to access the private service hosted in a VPC through AWS PrivateLink. PrivateLink provides private connectivity between the credit service VPC and Lambda on the private AWS network. All services are secured in a way that only the main AWS AppSync API is granted access.
Step 8
The credit service is hosted in AWS Fargate containers in a private VPC, and payment information is persisted in Amazon DynamoDB.
Step 9
Customer requests are evaluated against their pseudo credit rating using a service from the Amazon SageMaker elastic inference and provided a near real-time decision.
Well-Architected Pillars
The architecture diagram above is an example of a Solution created with Well-Architected best practices in mind. To be fully Well-Architected, you should follow as many Well-Architected best practices as possible.
Operational Excellence
Security
For secure authentication and authorization, IAM roles are in place for all service interactions in the environments. All data is fully encrypted in transit and in storage. AWS services provide HTTPS endpoints using TLS for communication, thus providing encryption in transit when communicating with the AWS APIs.
Reliability
To support a highly available network, all components scale automatically; account limits should be clearly defined for the supported product range. To adapt to changes in demand, the solution is modular and can scale with user adoption.
Performance Efficiency
Serverless architectures help to provision the exact resources that the workload needs. Strategies for storage lifecycle management and ensuring auto capacity scaling are used for ingestion and read and write access patterns. To meet the requirements of scaling, traffic and data access patterns, serverless architectures help to provision the exact resources that the workload needs.
Cost Optimization
This Guidance is designed to be fully optimized for cost, using only resources where necessary and only accessing data using the services appropriate for the business need.
Sustainability
By using managed services and dynamic scaling, you minimize the environmental impact of the backend services. This Guidance uses technologies that support data and access storage patterns that need to be monitored. Doing so ensures that assets, such as data, are stored in the optimum solution based on the read and write access patterns, with close attention to scaling of compute resources closely aligned to the demand.
Related content
How AWS is supporting Buy Now Pay Later (BNPL)
This post demonstrates how financial services institutions can leverage AWS to create BNPL journeys.