Architecture Design – Woodwork Manufacturing Company Migrating to AWS

1. Introduction / Business Objective

This architecture is for a major Woodwork Manufacturing organization in Cameroon with an annual turnover of over 150 million Xaf and 40 staff. Her current challenges include but are not limited to: long production times due to lack of proper monitoring and planning systems, delays to generate invoices as they are all done manually, loss of customer and production data, misuse of production materials, and lack of a centralized database.

Thus, the proposal to migrate core operations to AWS Cloud to improve efficiency, automate order and production tracking, secure customer data, digitize invoicing, and implement real-time customer updates not leaving out proper data management and insights.

2. Key Requirements

  1. Predict material needs and costing as well as job duration from past data
  2. Auto-generate invoices/confirmation of orders
  3. Notify production supervisor when a new job is confirmed and prior to expected delivery date
  4. Digitally store signed invoices (auto archive)
  5. Alert customers on order status via SMS
  6. Customer data/order management
  7. Generate business insights from data for decision making
organization solution

3. Key Services

  1. API Gateway:

    An API Gateway acts as a single-entry point for all client requests, routing them to the appropriate backend services. This is the entry point for external interactions. A Sales Agent initiates an action, placing an order, via an API request.

  2. Lambda Functions:

    AWS Lambda is a serverless, event-driven compute service that lets you run code without provisioning or managing servers. Various Lambda functions handle all event-driven compute needs, from initial request processing and real-time DynamoDB stream events, to orchestrating complex invoice generation workflows within Step Functions, and managing confirmations and notifications.

  3. DynamoDB:

    Fully managed NoSQL database offering single-digit millisecond latency. Processed orders and customer data are stored here as the operational database.

  4. DynamoDB Streams:

    Time-ordered sequence of item-level modifications in a DynamoDB table, capturing data changes in near real-time. Captures changes (new orders, updates) and triggers a Lambda function to perform specific tasks.

  5. S3 Buckets:

    Object storage for:

  6. Amazon EventBridge:

    Serverless event bus that routes events to targets based on rules. Used to send events to SNS two weeks before expected delivery dates.

  7. Amazon SNS:

    Pub/sub messaging service for sending SMS/email messages:

  8. AWS Step Functions:

    Visually orchestrates distributed applications and serverless workflows. Manages multi-step processes such as:

  9. S3 Glacier:

    Cold storage used for:

  10. Amazon SageMaker (Training):

    Uses prepared data from S3 Glacier to train ML models for job duration, cost prediction, and material needs.

  11. SageMaker Endpoint Model:

    Real-time deployment of the trained ML model. Invoked by Lambda to generate predictions used during invoicing.

  12. AWS Glue & Glue Data Catalog:

    Serverless ETL for:

  13. AWS Athena:

    Serverless SQL query engine for analyzing S3 data. Utilizes schemas from the Glue Catalog for business reporting.

  14. AWS QuickSight:

    Business Intelligence service used to create interactive dashboards and visual reports sourced from Athena queries.

  15. Amazon GuardDuty:

    Threat detection for monitoring suspicious activity in the account and across services.

  16. AWS WAF:

    Web Application Firewall to protect API Gateway from common attacks such as SQL injection and XSS.

4. Service Workflow

  1. A Sales Agent Places an Order

    The agent logs in securely (via VPN) and submits a new or updated job request. The request is sent through API Gateway, where AWS WAF filters incoming traffic for threats.

  2. The Order Enters the Processing System

    AWS Step Functions receives and inspects the data. It determines whether the order is new or an update and transforms the data into a processable format.

  3. Order Is Stored

    The system writes the job information to DynamoDB. If it’s a modification, the item is updated; if new, it’s inserted. DynamoDB Streams notify other services of the change.

  4. Smart Prediction Starts (ML Integration)

    A Lambda triggers Amazon SageMaker to run a real-time prediction from its deployed model. Outputs include job duration, material quantity needed, and minimum pricing with margin (e.g., 35% profit).

  5. Invoice Is Automatically Generated

    Step Functions coordinate a series of Lambda functions that use the prediction result to:

  6. Invoice Is Stored in the Cloud

    The finished invoice is stored in an S3 bucket. Lifecycle rules archive it to S3 Glacier after a retention period. The linked DynamoDB item is updated to reflect storage status.

  7. Everyone Is Notified

    SNS sends messages to the production supervisor and the customer. EventBridge schedules future alerts based on expected delivery timelines.

  8. Data Is Analyzed and Shared

  9. Security and Monitoring

5. Positioning with the Well-Architected Framework

SN Pillar Components Benefits in Our Architecture
1 Operational Excellence
  • Automated operations
  • Observability
  • Step Functions and Lambda automate various workflows (e.g., invoicing), reducing manual effort.
  • CloudWatch is integrated to monitor Lambda, DynamoDB Streams, and API Gateway for performance and logs.
2 Security
  • Least privilege access
  • Data protection
  • Network security
  • Secure communication
  • IAM roles are assigned to staff and AWS services with fine-grained permissions (e.g., DynamoDB, Glue Crawler).
  • S3 uses SSE-KMS encryption; TLS enforced for API Gateway; HTTPS enabled for S3; data is encrypted in transit.
  • API Gateway is protected by WAF to filter threats and secure incoming traffic.
  • VPC Endpoints ensure data flow between services like S3 and DynamoDB stays within the AWS network.
3 Reliability
  • Fault isolation
  • Automated recovery
  • Scalable service design
  • Backup and versioning
  • Lambda functions are stateless and event-driven, isolating faults and reducing impact.
  • Step Functions provide automatic retries and fallback mechanisms on failure.
  • Services like API Gateway, S3, DynamoDB, and SNS are serverless and automatically scaled.
  • DynamoDB has Point-in-Time Recovery (PITR) and on-demand backups. S3 uses versioning and lifecycle policies to archive data to Glacier.
4 Cost Optimization
  • Pay-as-you-go
  • Data lifecycle management
  • Right-sized resources
  • Monitoring costs
  • Lambda, Step Functions, and S3 are consumption-based services that reduce cost during low usage.
  • S3 lifecycle policies move stale data (e.g., invoices) to Glacier to reduce storage cost.
  • SageMaker endpoint is invoked on-demand, avoiding idle compute time.
  • CloudWatch monitors request volumes, latency, and errors; AWS Budgets provides cost visibility.
5 Performance Efficiency
  • Serverless compute
  • Data processing optimization
  • Asynchronous messaging
  • Lambda and Step Functions scale automatically to meet real-time demand.
  • Glue and Athena transform and query large datasets without preloading into databases.
  • SNS, DynamoDB Streams, and EventBridge decouple systems to handle spikes and reduce latency.
6 Sustainability
  • Efficient compute usage
  • Optimized storage
  • Lambda only runs when needed, reducing idle resource waste.
  • S3 uses lifecycle policies to shift cold data to Glacier for greener, cost-saving storage.

Walson Baiye Mboe / AWS Certified Solutions Architect – Associate / AWS Certified Cloud Practitioner / Student Optima IT (2025)