Serverless Architecture for Global Applications.pdf
免费下载
© 2020, Amazon Web Services, Inc. or its affiliates. All rights reserved.
AWS Reference Architecture
Sync
Write
Query
CDC Event Stream
Amazon Kinesis
Cross-Region
Replication
S3 CRR
S3 Access Points
Cross-Region
Replication
Global Database
RDS Proxy
Cross-Region
Replication
Global Tables
DAX
Propagate changes across Regions with
active-active data replication.
8
Store for data access patterns: e.g.,
key-value store for Single View of
Entities, relational for Reporting, and
data lake for Analytics and AI/ML.
5
Process transactional logic via serverless
functions and run long running tasks
with containers. Use the Saga pattern to
orchestrate distributed transactions with
eventual consistency. Add intelligence
via Machine Learning and Augmented AI.
4
Process asynchronous commands by
adding them to the event bus. Source
additional events from external systems.
3
Front each Region with an entry API
implementing the CQRS pattern.
For query requests, read from the data
layer. For synchronous write commands,
invoke the logic layer.
2
Route traffic from edge locations based
on request path, allowing a gradual
migration of single-CNAME legacy API
operations. Then route requests to the
Region with the least latency for the
requester.
1
Serverless Architecture for Global Applications
Improve customer experience on your global services by deploying into multiple AWS Regions and applying event-driven
architectural patterns to reduce latency and increase performance, such as event sourcing, saga orchestration, and CQRS.
Region
REST/SOAP
Amazon
API Gateway
AWS Cloud
Single View
Amazon
DynamoDB
Report DB
Amazon
Aurora
Data Lake
Amazon S3
Lake Formation
Region
REST/SOAP
Amazon
API Gateway
Single View
Amazon
DynamoDB
Report DB
Amazon
Aurora
Data Lake
Amazon S3
Lake Formation
Serverless
Microservices
AWS Lambda
Long Running
Containers
AWS Fargate
Data Layer
Logic Layer
Ingestion
Amazon S3
Async
Write
Latency
Routing
Amazon
Route 53
Query
CQRS
Write
Path-based
Routing
Amazon
CloudFront
Systems of
Record
File Systems
AWS Storage
Gateway
Users
Database Changes
File System Changes
1
2
3
5
8
9
For heavy reading scenarios on other
Regions, send write requests to primary
Regions. For fast global writes, replicate
the event and command logic on all
Regions.
9
DB CDC Events
AWS DMS
Saga Pattern
AWS Step
Functions
Event Bus
Amazon EventBridge
Machine
Learning
Amazon A2I
Application Events
Systems of Record
Logic Events
Data Events
Write
4
Raise data events after changes to the
canonical data model on the data layer,
reducing redundance in the logic layer.
7
Raise logic events from the logic layer to
run event-driven workflow steps,
following the Transformation pattern.
6
6
7
of 1
免费下载
【版权声明】本文为墨天轮用户原创内容，转载时必须标注文档的来源（墨天轮），文档链接，文档作者等基本信息，否则作者和墨天轮有权追究责任。如果您发现墨天轮中有涉嫌抄袭或者侵权的内容，欢迎发送邮件至：contact@modb.pro进行举报，并提供相关证据，一经查实，墨天轮将立刻删除相关内容。
下载排行榜
文档被以下合辑收录
评论