Data flows into AWS through batch processing, 1 real-time data, Secure File Transfer Protocol Implementing Travel & Hospitality Data Mesh (SFTP) and IoT sensors. The travel and hospitality industries are facing new challenges when generating, accessing, and analysing data at scale. Use this new approach to build a data Data sources are managed by the Business platform that serves both operational and analytical needs by using domain owned design, maintained data properties, open data standards, purpose-built 2 databases, and extensible serverless architecture. This data delivery platform architecture helps relieve and eventually replace the on-premises data platform Domain. Producers use organization-level load, leading to cost savings and an agile environment. blueprints to provide core services such as security, governance, and open standards. Producers build the Operational Data Store using External AWS Cloud DynamoDB and Amazon DynamoDB 2 Data Producer Domain Accounts Central IT Platform 3 Consumer Domain Accounts Accelerator (DAX) to provide consistent single- 4 digit-millisecond latency performance. Producer 1 Account – Metadata Reservation Domain Consumer 1 Account- Batch 5 Marketing Domain Consumer Domain uses process data sets from 3 AWS Glue Amazon Amazon multiple producer domains based on business 1 Data Catalog Neptune DynamoDB needs. Consumers build a data mart using Events Data Amazon Amazon Amazon Ingestion Crawler Amazon Redshift and analytics using Amazon EMR Athena (Data mart) Athena and Governance QuickSight. Integration

Amazon S3 Amazon Metadata is managed via multiple services. AWS AWS Amazon AWS Cloud Amazon S3 Amazon S3 Derived Data Sets QuickSight 4 CloudTrail CloudWatch Formation Glue is used for data cataloguing. Data lineage is Raw Data Process Data AWS Transfer Raw Data (Visualization) stored in . Data contracts are System of Family Batch Real Time Sets Records AWS Security & Access stored in DynamoDB. Management Services Consumer accounts 2 – Revenue Management Domains Consumers use online, text-based search and Amazon AWS Step AWS Amazon 5 EventBridge filtering of data sets recorded in the central data Batch Functions Lambda DynamoDB (ODS) AWS AWS catalog to collect data. They can search for AWS AWS Consumer accounts 3..m – Secrets Single IAM KMS relevant data by name, contents, sensitivity, or Manager Sign-On Other Domains such as Sales, Finance, Customer, Airport, other any other custom labels you have defined. Amazon Kinesis Producer 2 Account - Events and Operation Loyalty Domain AWS Lake Formation provides centralized Amazon Amazon Amazon AWS Firewall 7 GuardDuty Inspector Macie Manager Account(s) - 6 management of security, governance and auditing policies including fine-grain permissions. It also AWS Data Producer 3 Account - Exchange enables automatic schema discovery and Engineering Domain Discovery Amazon Amazon conversion to the required format. IOT Sensor SageMaker Athena Data Batch AWS IoT Core Producer accounts – Other Amazon Crawler AWS Glue ML (Machine Learning) is performed using Lake Domains such as Digital, Athena DataBrew 7 Formation. Use Amazon SageMaker to provide Revenue, and HR Amazon S3 AI Services standard AI/ML models for customer AWS Lake Formation 6 Predictive Models AWS IoT Greengrass Events segmentations and lifetime value. Other AI services such as Amazon Personalize can be utilized to get actionable insight.

Reviewed for technical accuracy August 3, 2021 AWS Reference Architecture © 2021, , Inc. or its affiliates. All rights reserved. Data flows into AWS through batch processing, Implementing Travel & Hospitality Data Mesh 1 real-time data, SSH File Transfer Protocol (SFTP) and IoT sensors. A view of all of the services in the data mesh and how they connect and work together on AWS. Provide staging for ingesting all batch and real- 2 time data using cost effective storage classes in External Data AWS Cloud 3 4 Amazon Simple Storage Service (Amazon S3).

An AWS Glue crawler creates the table metadata AWS Glue AWS Glue AWS Glue Amazon EMR 6 3 Crawler Data Catalog in the data catalog. The AWS Glue data catalog AWS Transfer Amazon Redshift is an index of the location, schema, and runtime Datamart metrics of your data. Use the information in the 1 Amazon QuickSight Reporting and Analytics data catalog to create and monitor your ETL Analytics AWS Storage Dashboard (extract, transform, and load) jobs. Gateway Amazon S3 Amazon S3 Amazon S3 Amazon S3 Amazon S3 Raw Data Raw Data Processed Derived Data Predictive Amazon Athena Batch Staging Real-time Data Sets Sets Models SageMaker Studio IDE Use open standards to build the data lake using AWS Data 4 the same data as the operational data platform. Exchange Use a read pattern schema to make the raw data … … … … … Amazon SageMaker Amazon SageMaker Amazon SageMaker Studio Notebooks Ground Truth Build and curated data readily available seamlessly for all user roles using a workflow such as Apache 2 … … … Mobile Customer Airflow. System of … … Service Apps Records Amazon SageMaker Amazon SageMaker Amazon SageMaker Model Monitor Deploy and Manage Train and Tune Use purpose-built databases such as Amazon … … … … … 5 DynamoDB and serverless architecture to deliver 7 Customer Service Amazon SageMaker Dashboard microservices and events to the operational data Prediction Models Debugger store. … … … … …

5 Customer Website Build all reportable data sets in Amazon S3 and Events 6 leverage Amazon Redshift and Amazon Athena Lambda function Amazon API Gateway 1 Amazon DynamoDB for analytics. Optionally, for heavily used Amazon Kinesis Amazon Kinesis Operational Data Store Data Streams Data Firehose analytics, build data marts in Amazon Redshift. AWS Step Functions AWS Step Functions Mobile Customer For ad-hoc requirements, use Athena to analyse Customer Data Processing Complex Event Processing Apps data in the data lake with standard SQL.

IOT Sensor DynamoDB Amazon Kinesis Published Events Lambda function Lambda function Use Amazon SageMaker to provide standard Data Lambda function Lambda function Streams Data Streams Events Data Events 7 AI/ML models for customer segmentations and 1 AWS Security and Access Management Services Governance lifetime value. Other AI services such as Amazon Subscribers Personalize can be utilized to get actionable 8 AWS IoT AWS IoT insight. AWS Greengrass Core AWS AWS AWS Secrets AWS Single Amazon Amazon Amazon AWS Firewall AWS Amazon IAM KMS Manager Sign-On GuardDuty Inspector Macie Manager CloudTrail CloudWatch CloudFormation As a best practice, use a multi-account strategy 8 to provide resources and security isolation for workloads. This helps with categorization and Reviewed for technical accuracy August 3, 2021 AWS Reference Architecture reducing blast radius. © 2021, Amazon Web Services, Inc. or its affiliates. All rights reserved.