1 Build data products for relevant domains (like flight, passenger, Airline initiatives to build operations data stores and related services often don’t adapt to change due to rigid schemas, long implementation times, siloed operations and analytics systems, and on-premises scaling limitations to add new domains. Use this data platform architecture to relieve or replace your on-premises data platform load, increasing development agility and cost savings; it uses the Implementing Travel & Hospitality Data Mesh reference architecture as its foundation for a domain-owned design approach, open data standards, purpose-built databases, and extensible serverless architecture. AWS Cloud events Amazon Simple Storage Service (Amazon S3) 3 AWS Step Functions customer event processing Amazon API Gateway API clients • mobile customer service apps • customer service dashboard events . real-time staging • flight schedule • PNRs • tickets • ancillaries • loyalty membership Amazon Kinesis 1 systems of record • flight ops • Passenger service system (PSS) • Departure control system (DCS) • loyalty • baggage sortation • booking engine • surveys system external data • demographic data • address data data events data sets batch ingestion AWS Transfer Family AWS Data Exchange Amazon AppFlow • baggage sortation batch staging • demographic data • address data • loyalty transactions • shopping, regrets, denials, abandoned • surveys • flight schedule • PNRs • tickets parquet • ancillaries • loyalty membership • loyalty transactions • baggage sortation • surveys • market segment revenue • loyalty life time value reportable data sets • loyalty revenue index • loyalty customer service index • shopping data insights Reviewed for technical accuracy August 23, 2022 © 2022, Amazon Web Services, Inc. or its affiliates. All rights reserved. AWS Lambda stage data transform Amazon EMR AWS Glue AWS Glue Data Catalog transformations AWS Fargate 2 Amazon DynamoDB customer 360 data data stream enhance Amazon Redshift customer data marts 5 Amazon SageMaker customer analysis reportable data sets Amazon Athena AWS Reference Architecture • customer website AWS Step Functions microservices • mobile customer apps baggage reconciliation AWS Step Functions complex event processing 4 World Tracer Net Tracer published customer events customer event subscribers Amazon QuickSight customer analytics dashboard users and loyalty), separating storage from compute. 2 3 4 5 In the operational data store, use managed services and purposebuilt databases with microservice and event-driven patterns. This enables you to deprecate expensive on-premises infrastructure (like operational databases, service-oriented architecture (SOA) infrastructure, and message-oriented middleware), replacing it with architecture components that allow you to scale based on client adoption. Use open standards to build a data lake using the same data as in the operational data platform. Use a read-pattern schema to make the raw data and curated data readily available for all user roles. For well-known query patterns, build standard enterprise-datawarehouse schemas and data marts in Amazon Redshift. For ad-hoc query requirements, publish the data catalog in AWS Glue and use Amazon Athena for querying the data lake directly. Extend the data warehouse based on your specific needs. Use Amazon SageMaker to provide standard artificial intelligence (AI) and machine learning (ML) models for
