2022-09-22.jcon.integration-and-conversation-patterns.foundations.pdf

© 2022, Amazon Web Services, Inc. or its affiliates. All
rights reserved. © 2022, Amazon Web Services, Inc. or its affiliates. All rights reserved. Integration and conversation patterns - Foundations Dirk Fröhner (he/him) Principal Solutions Architect Amazon Web Services J C O N O N L I N E , 2 0 2 2 - 0 9 - 2 2

rights reserved. Motivation Patterns Use cases Resources Agenda

rights reserved. If your application is cloud-native, large-scale, or distributed, and doesn’t include a messaging component, that’s probably a bug. Tim Bray General-purpose internet-software geek

rights reserved. Potential drawbacks of synchronous systems Synchronous systems are inherently tightly coupled Problems in downstream systems can have immediate impact on upstream callers Retries from upstream callers can easily fan out and amplify problems Photo: Dirk Fröhner

rights reserved. Potential drawbacks of synchronous systems Synchronous systems are inherently tightly coupled Problems in downstream systems can have immediate impact on upstream callers Retries from upstream callers can easily fan out and amplify problems Photo: Dirk Fröhner . . . and some things simply take too much time to wait, or are asynchronous by nature

rights reserved. © 2022, Amazon Web Services, Inc. or its affiliates. All rights reserved. Patterns Gregor Hohpe and Bobby Woolf: Enterprise Integration Patterns Designing, Building, and Deploying Messaging Solutions

rights reserved. Message exchange One-way

rights reserved. Message exchange One-way Sender Message channel Receiver A A No response expected Command-, document-, event-message Message intent

rights reserved. Message exchange One-way Asynchronous request-response Sender Message channel Receiver A A No response expected Command-, document-, event-message Message intent

rights reserved. Message exchange One-way Asynchronous request-response Message channel Requester Message channel Responder Sender Message channel Receiver A A A A B B Response expected Return address Correlation ID No response expected Command-, document-, event-message Message intent

rights reserved. Message exchange One-way Asynchronous request-response Message channel Requester Message channel Responder Sender Message channel Receiver Integration pattern A A A A B B No response expected Command-, document-, event-message Message intent Response expected Return address Correlation ID

rights reserved. Message exchange One-way Asynchronous request-response Message channel Requester Message channel Responder Sender Message channel Receiver Conversation pattern A A A A B B Integration pattern No response expected Command-, document-, event-message Message intent Response expected Return address Correlation ID

rights reserved. Message channels Point-to-point (queue)

rights reserved. Message channels Point-to-point (queue) Sender(s) Queue Receivers B A C B A C Each message consumed by one receiver Competing consumers Buffering load balancer

rights reserved. Message channels Point-to-point (queue) Publish-subscribe / fan-out (topic) Sender(s) Queue Receivers B A C B A C Each message consumed by one receiver Competing consumers Buffering load balancer

rights reserved. Message channels Point-to-point (queue) Publish-subscribe / fan-out (topic) Publisher(s) Topic Subscribers Sender(s) Queue Receivers B A C B A C B A C B A C B A C B A C Each message consumed by one receiver Competing consumers Each message consumed by each subscriber How to scale consumers? Conceptually no buffering Buffering load balancer

rights reserved. Message channels Publish-subscribe / fan-out (message bus) Prosumer Prosumer Message bus Scope: components & applications Prosumers over producers vs consumers Additional functionality, but also constraints, e.g.: content-based router, schema validation, payload format constraint, adapters for 3P software Prosumer Prosumer Prosumer

rights reserved. Message channels Point-to-point (queue) Publish-subscribe / fan-out (topic) Publisher(s) Subscribers Sender(s) Receivers B A C B A C B A C B A C B A C B A C Amazon Simple Queue Service (Amazon SQS) Cloud-native & serverless Amazon Simple Notification Service (Amazon SNS) Cloud-native & serverless AWS services implementing various message channel patterns

rights reserved. Message channels Point-to-point (queue) Publish-subscribe / fan-out (topic) Publisher(s) Subscribers Sender(s) Receivers B A C B A C B A C B A C B A C B A C Amazon MQ – managed RabbitMQ or Apache ActiveMQ For applications constrained to industry- standard protocols like JMS or AMQP Amazon MQ – managed RabbitMQ or Apache ActiveMQ For applications constrained to industry- standard protocols like JMS or AMQP AWS services implementing various message channel patterns

rights reserved. Message channels Publish-subscribe / fan-out (topic) Publisher(s) Subscribers B A C B A C B A C B A C AWS IoT Core – Message Broker MQTT, MQTT over WebSocket, and HTTPS (publish only) for apps in the IoT context AWS services implementing various message channel patterns

rights reserved. Message channels Publish-subscribe / fan-out (message bus) Prosumer Prosumer Message bus Amazon EventBridge Cloud-native & serverless For building event-driven applications at scale across AWS, existing systems, or SaaS apps Prosumer Prosumer Prosumer AWS services implementing various message channel patterns Amazon EventBridge

rights reserved. Message channels What about . . .

rights reserved. Message channels What about . . . Streams?

rights reserved. Excursus: Message vs. stream processing Message processing Stream processing A stream is yet another fan-out pattern, but:

rights reserved. Excursus: Message vs. stream processing Message processing Stream processing Each individual message is a unit of work Messages deleted after consumption Photo: Christian Müller A stream is yet another fan-out pattern, but:

rights reserved. Excursus: Message vs. stream processing Message processing Stream processing Each individual message is a unit of work Messages deleted after consumption Batches of messages are units of work Messages retained after consumption (until expiration) Photo: Christian Müller Photo: Christian Müller A stream is yet another fan-out pattern, but:

rights reserved. Excursus: Message vs. stream processing Message processing Stream processing Each individual message is a unit of work Messages deleted after consumption Batches of messages are units of work Messages retained after consumption (until expiration) Photo: Christian Müller Photo: Christian Müller A stream is yet another fan-out pattern, but: Operational overhead

rights reserved. Excursus: Message vs. stream processing AWS services implementing the stream pattern

rights reserved. Excursus: Message vs. stream processing AWS services implementing the stream pattern Amazon Kinesis Data Streams Amazon Kinesis Data Firehose Amazon Kinesis Data Analytics Amazon Kinesis

rights reserved. Excursus: Message vs. stream processing AWS services implementing the stream pattern Amazon Kinesis Data Streams Amazon Kinesis Data Firehose Amazon Kinesis Data Analytics Amazon Kinesis Amazon Managed Streaming for Apache Kafka

rights reserved. Message channels Composite pattern: Topic-queue-chaining

rights reserved. Message channels Durable subscriber pattern: How to avoid missing messages while not listening? Composite pattern: Topic-queue-chaining

rights reserved. Message channels Producer Topic B A C Durable subscriber pattern: How to avoid missing messages while not listening? Composite pattern: Topic-queue-chaining

rights reserved. Message channels Producer Topic B A C Queue Queue A C B A C B Consumers B A C B A C Durable subscriber pattern: How to avoid missing messages while not listening? Composite pattern: Topic-queue-chaining

rights reserved. Message channels Consumers B A C Producer Topic B A C Queue Queue Queue A C B A C B A C B Consumers B A C B A C Durable subscriber pattern: How to avoid missing messages while not listening? Concurrent consumer pattern: How to scale consumers of publish-subscribe channels? Composite pattern: Topic-queue-chaining

rights reserved. Message channels Consumers B A C Producer Topic B A C Queue Queue Queue A C B A C B A C B Consumers B A C B A C Durable subscriber pattern: How to avoid missing messages while not listening? Concurrent consumer pattern: How to scale consumers of publish-subscribe channels? Best of both worlds: fan-out & consumer scale-out & buffering load balancing all at the same time Composite pattern: Topic-queue-chaining

rights reserved. Message channels Dead-letter queue (DLQ)

rights reserved. Message channels Dead-letter queue (DLQ) Producer Queue A C B

rights reserved. Message channels Dead-letter queue (DLQ) Producer Queue A C B Consumers B A C

rights reserved. Message channels Dead-letter queue (DLQ) Producer Queue A C B Consumers B A C Processing message C : Exception! C

rights reserved. Message channels Dead-letter queue (DLQ) Producer Queue A C B Consumers B A C Processing message C : Exception! Repeatedly! C

rights reserved. Message channels Dead-letter queue (DLQ) Producer Queue A C B Consumers B A C Processing message C : Exception! Repeatedly! C Transient failure mitigation Poison pill handling

rights reserved. Message channels Dead-letter queue (DLQ) DLQ Producer Queue A C B Consumers B A C C Transient failure mitigation Poison pill handling

rights reserved. Message channels Dead-letter queue (DLQ) Application ops DLQ C Producer Queue A C B Consumers B A C C Transient failure mitigation Poison pill handling No need for open-heart surgery – investigate without the hassle of production

rights reserved. Message channels Visibility timeout, inflight messages, and message acknowledgement Case 1: Fast enough consumer

rights reserved. Message channels Visibility timeout, inflight messages, and message acknowledgement Time

rights reserved. Message channels Visibility timeout, inflight messages, and message acknowledgement A Time snd_msg P1 A

rights reserved. Message channels Visibility timeout, inflight messages, and message acknowledgement A Time snd_msg P1 A Message status: Stored

rights reserved. Message channels Visibility timeout, inflight messages, and message acknowledgement A rcv_msg A A Time snd_msg P1 A C1 Message status: Stored

rights reserved. Message channels Visibility timeout, inflight messages, and message acknowledgement A A rcv_msg A A Time snd_msg P1 A C1 Visibility timeout Message status: Stored Message status: In flight

rights reserved. Message channels Visibility timeout, inflight messages, and message acknowledgement A A A rcv_msg A rcv_msg A Time snd_msg P1 C2 A C1 Visibility timeout Message status: Stored Message status: In flight

rights reserved. Message channels Visibility timeout, inflight messages, and message acknowledgement A A A rcv_msg A rcv_msg A Time snd_msg P1 C1 C2 A A A C1 ack_msg Visibility timeout Message status: Stored Message status: In flight

rights reserved. Message channels Visibility timeout, inflight messages, and message acknowledgement A A A rcv_msg A rcv_msg A Time snd_msg P1 C1 C2 A A A C1 ack_msg Visibility timeout Message status: Stored Message status: In flight Message status: Deleted

rights reserved. Message channels Visibility timeout, inflight messages, and message acknowledgement Case 2: Not fast enough consumer

rights reserved. Message channels Visibility timeout, inflight messages, and message acknowledgement Time

rights reserved. Message channels Visibility timeout, inflight messages, and message acknowledgement A Time snd_msg P1 A

rights reserved. Message channels Visibility timeout, inflight messages, and message acknowledgement A Time snd_msg P1 A Message status: Stored

rights reserved. Message channels Visibility timeout, inflight messages, and message acknowledgement A rcv_msg A A Time snd_msg P1 A C1 Message status: Stored

rights reserved. Message channels Visibility timeout, inflight messages, and message acknowledgement A A rcv_msg A A Time snd_msg P1 A C1 Visibility timeout Message status: Stored Message status: In flight

rights reserved. Message channels Visibility timeout, inflight messages, and message acknowledgement A A A rcv_msg A rcv_msg A Time snd_msg P1 C2 A C1 Visibility timeout Message status: Stored Message status: In flight

rights reserved. Message channels Visibility timeout, inflight messages, and message acknowledgement A A A rcv_msg A rcv_msg A A snd_msg P1 C1 C2 A Visibility timeout Message status: Stored Message status: In flight Time

rights reserved. Message channels Visibility timeout, inflight messages, and message acknowledgement A A A rcv_msg A rcv_msg A A snd_msg P1 C1 C2 A A Visibility timeout Message status: Stored Message status: In flight Message status: Stored Time

rights reserved. Message channels Visibility timeout, inflight messages, and message acknowledgement A A A A rcv_msg A rcv_msg rcv_msg A A A snd_msg P1 C1 C2 C2 A A Visibility timeout Message status: Stored Message status: In flight Message status: Stored Time

rights reserved. Message channels Visibility timeout, inflight messages, and message acknowledgement A A A A rcv_msg A rcv_msg rcv_msg A A A A Time snd_msg P1 C1 C2 C2 A Message status: Stored Message status: In flight Message status: Stored A Message status: In flight Visibility Visibility timeout

rights reserved. Message channels FIFO queues

rights reserved. Message channels FIFO queues Strict order for message delivery

rights reserved. Message channels FIFO queues Producer FIFO queue C B D Consumer Strict order for message delivery A C B D A Order!

rights reserved. Message channels FIFO queues Producer FIFO queue C B D Consumer Strict order for message delivery All magic comes with a price, deary A C B D A Order!

rights reserved. Message channels FIFO queues What happens with multiple consumers?

rights reserved. Message channels FIFO queues What happens with multiple consumers? Concurrent message consumption!

rights reserved. t m Message channels FIFO queues Producer FIFO queue C B D Consumers What happens with multiple consumers? Concurrent message consumption! A Protocol

rights reserved. t m Message channels FIFO queues Producer FIFO queue C B D Consumers What happens with multiple consumers? Concurrent message consumption! A Protocol A

rights reserved. t m Message channels FIFO queues Producer FIFO queue C B D Consumers What happens with multiple consumers? Concurrent message consumption! A Protocol A B A

rights reserved. t m Message channels FIFO queues Producer FIFO queue C B D Consumers What happens with multiple consumers? Concurrent message consumption! A Protocol A B C B A

rights reserved. t m Message channels FIFO queues Producer FIFO queue C B D Consumers What happens with multiple consumers? Concurrent message consumption! A Protocol A B C B A C

rights reserved. t m 0 B Message channels FIFO queues Producer FIFO queue C B D Consumers What happens with multiple consumers? Concurrent message consumption! A Protocol A B C B A C

rights reserved. t m 0 B Message channels FIFO queues Producer FIFO queue C B D Consumers What happens with multiple consumers? Concurrent message consumption! A Protocol A B C B D A C

rights reserved. t m 0 B Message channels FIFO queues Producer FIFO queue C B D Consumers What happens with multiple consumers? Concurrent message consumption! A Protocol A B C B D A D C

rights reserved. t m 0 B 1 A Message channels FIFO queues Producer FIFO queue C B D Consumers What happens with multiple consumers? Concurrent message consumption! A Protocol A B C B D A D C

rights reserved. t m 0 B 1 A 2 D Message channels FIFO queues Producer FIFO queue C B D Consumers What happens with multiple consumers? Concurrent message consumption! A Protocol A B C B D A D C

rights reserved. t m 0 B 1 A 2 D 3 C Message channels FIFO queues Producer FIFO queue C B D Consumers What happens with multiple consumers? Concurrent message consumption! A Protocol A B C B D A D C

rights reserved. t m 0 B 1 A 2 D 3 C Message channels FIFO queues Producer FIFO queue C B D Consumers What happens with multiple consumers? Concurrent message consumption! Message processing time can vary and lead to order violation in consumer processes A Protocol A B C B D A D C

rights reserved. Message channels Message groups

rights reserved. Message channels Message groups Messages grouped by discriminator attribute

rights reserved. Message channels Message groups Messages grouped by discriminator attribute Example: Amazon SQS message group IDs

rights reserved. t m Message channels Message groups Producer FIFO queue C B D Consumers A Protocol Messages grouped by discriminator attribute Example: Amazon SQS message group IDs

rights reserved. t m Message channels Message groups Producer FIFO queue C B D Consumers A Protocol A Messages grouped by discriminator attribute Example: Amazon SQS message group IDs

rights reserved. t m Message channels Message groups Producer FIFO queue C B D Consumers A Protocol A A Messages grouped by discriminator attribute Example: Amazon SQS message group IDs

rights reserved. t m Message channels Message groups Producer FIFO queue C B D Consumers A Protocol A B A Messages grouped by discriminator attribute Example: Amazon SQS message group IDs

rights reserved. t m Message channels Message groups Producer FIFO queue C B D Consumers A Protocol A B B A Messages grouped by discriminator attribute Example: Amazon SQS message group IDs

rights reserved. t m 0 B Message channels Message groups Producer FIFO queue C B D Consumers A Protocol A B B A Messages grouped by discriminator attribute Example: Amazon SQS message group IDs

rights reserved. t m 0 B Message channels Message groups Producer FIFO queue C B D Consumers A Protocol A B B D A Messages grouped by discriminator attribute Example: Amazon SQS message group IDs Not C!

rights reserved. t m 0 B Message channels Message groups Producer FIFO queue C B D Consumers A Protocol A B B D A D Messages grouped by discriminator attribute Example: Amazon SQS message group IDs

rights reserved. t m 0 B 1 A Message channels Message groups Producer FIFO queue C B D Consumers A Protocol A B B D A D Messages grouped by discriminator attribute Example: Amazon SQS message group IDs

rights reserved. t m 0 B 1 A Message channels Message groups Producer FIFO queue C B D Consumers A Protocol A B B D A D Messages grouped by discriminator attribute Example: Amazon SQS message group IDs C

rights reserved. t m 0 B 1 A Message channels Message groups Producer FIFO queue C B D Consumers A Protocol A B B D A D Messages grouped by discriminator attribute Example: Amazon SQS message group IDs C C

rights reserved. t m 0 B 1 A Message channels Message groups Producer FIFO queue C B D Consumers A Protocol A B B A D C Messages grouped by discriminator attribute Example: Amazon SQS message group IDs D C

rights reserved. t m 0 B 1 A 2 D Message channels Message groups Producer FIFO queue C B D Consumers A Protocol A B B A D C Messages grouped by discriminator attribute Example: Amazon SQS message group IDs D C

rights reserved. t m 0 B 1 A 2 D 3 C Message channels Message groups Producer FIFO queue C B D Consumers A Protocol A B B A D C Messages grouped by discriminator attribute Example: Amazon SQS message group IDs D C

rights reserved. t m 0 B 1 A 2 D 3 C Message channels Message groups Producer FIFO queue C B D Consumers A Protocol A B B A D C Messages grouped by discriminator attribute Example: Amazon SQS message group IDs Order retained within each message group Consumer scale-out on # of message group IDs D C

rights reserved. t m 0 B 1 A 2 D 3 C Message channels Message groups Producer FIFO queue C B D Consumers A Protocol A B B A D C Messages grouped by discriminator attribute Example: Amazon SQS message group IDs Order retained within each message group Consumer scale-out on # of message group IDs D C Local vs. global message order?

rights reserved. t m 0 B 1 A 2 D 3 C Message channels Message groups Producer FIFO queue C B D Consumers A Protocol A B B A D C Messages grouped by discriminator attribute Example: Amazon SQS message group IDs Order retained within each message group Consumer scale-out on # of message group IDs D C What if a global order is needed?

rights reserved. t m 0 B 1 A 2 D 3 C Message channels Message groups Producer FIFO queue C B D Consumers A Protocol A B B A D C Messages grouped by discriminator attribute Example: Amazon SQS message group IDs Order retained within each message group Consumer scale-out on # of message group IDs D C What if a global order is needed? One message group for all messages, with respective implications

rights reserved. Message channels FIFO topics

rights reserved. Message channels FIFO topics Producer FIFO topic C B D Consumers A C B D A Order! C B D A Order! C B D A Order!

rights reserved. Message channels FIFO topics Producer FIFO topic C B D Consumers No concurrent consumers A C B D A Order! C B D A Order! C B D A Order!

rights reserved. Message channels FIFO topics Producer FIFO topic C B D Consumers No concurrent consumers What happens for topic-queue-chaining? A C B D A Order! C B D A Order! C B D A Order!

rights reserved. Message channels FIFO topics Producer FIFO topic Consumers No concurrent consumers What happens for topic-queue-chaining? Order! Order! Order! C B D A C B D A C B D A C B D A SNS FIFO topics also use message groups (MG)

rights reserved. Message channels FIFO topics Producer FIFO topic No concurrent consumers What happens for topic-queue-chaining? Order! Order! Order! C B D A C B D A C B D A C B D A MG IDs to be forwarded into queue messages SNS FIFO topics also use message groups (MG) FIFO Queue FIFO Queue FIFO Queue Order! Order! Order! C B D A C B D A C B D A Consumers

rights reserved. Message routing

rights reserved. Message routing Message filter

rights reserved. Message routing Message filter Receive only relevant subset of messages

rights reserved. Message routing Message filter Receive only relevant subset of messages Publisher(s) Topic Subscribers C A D B C B D A C B D A

rights reserved. Message routing Message filter Receive only relevant subset of messages Publisher(s) Topic Subscribers C A D B C B D A C B D A color == red color == green

rights reserved. Message routing Message filter Receive only relevant subset of messages Controlled by subscriber Publisher remains completely unaware Publisher(s) Topic Subscribers C A D B C B D A C B D A color == red color == green

rights reserved. Message routing Message filter Recipient list Receive only relevant subset of messages Controlled by subscriber Send only relevant subset of messages to each subscriber Publisher remains completely unaware Publisher(s) Topic Subscribers C A D B C B D A C B D A color == red color == green

rights reserved. Message routing Message filter Recipient list Publisher(s) Subscribers Receive only relevant subset of messages Controlled by subscriber Send only relevant subset of messages to each subscriber Publisher remains completely unaware Publisher(s) Topic Subscribers C A D B C B D A C B D A color == red color == green C B D A C A D B C B D A color == red color == green Recipient list component

rights reserved. Message routing Message filter Recipient list Publisher(s) Subscribers Receive only relevant subset of messages Controlled by subscriber Send only relevant subset of messages to each subscriber Controlled directly by publisher or a separate component Publisher remains completely unaware Publisher(s) Topic Subscribers C A D B C B D A C B D A color == red color == green C B D A C A D B C B D A color == red color == green Recipient list component

rights reserved. Message routing Scatter-gather

rights reserved. How to distribute a request to relevant/interested parties and capture their individual responses? Message routing Scatter-gather

rights reserved. How to distribute a request to relevant/interested parties and capture their individual responses? Message routing For election or parallel processing scenarios, i.e., search for best response or accumulate responses Scatter-gather

rights reserved. How to distribute a request to relevant/interested parties and capture their individual responses? Message routing Scatter-gather Requester For election or parallel processing scenarios, i.e., search for best response or accumulate responses

rights reserved. How to distribute a request to relevant/interested parties and capture their individual responses? Message routing Scatter-gather Requester Topic A For election or parallel processing scenarios, i.e., search for best response or accumulate responses

rights reserved. How to distribute a request to relevant/interested parties and capture their individual responses? Message routing Scatter-gather Requester Topic A A A A For election or parallel processing scenarios, i.e., search for best response or accumulate responses Responders

rights reserved. How to distribute a request to relevant/interested parties and capture their individual responses? Message routing Scatter-gather Requester Topic A A A A B1 For election or parallel processing scenarios, i.e., search for best response or accumulate responses Responders B3 B2

rights reserved. How to distribute a request to relevant/interested parties and capture their individual responses? Message routing Scatter-gather Requester Requester Topic A Queue A A A B1 For election or parallel processing scenarios, i.e., search for best response or accumulate responses Responders B2 B3

rights reserved. How to distribute a request to relevant/interested parties and capture their individual responses? Message routing Scatter-gather Requester Requester Topic A A A A B1 For election or parallel processing scenarios, i.e., search for best response or accumulate responses Responders B3 B1 B2 B3 B2 Queue

rights reserved. How to distribute a request to relevant/interested parties and capture their individual responses? Message routing Scatter-gather Requester Requester Topic A A A A B1 For election or parallel processing scenarios, i.e., search for best response or accumulate responses Responders Processor Aggregator B3 B1 B2 B3 B2 Queue

rights reserved. Context: Wild Rydes, Inc.

rights reserved. Submit ride completion U S E C A S E Wild Rydes unicorn

rights reserved. Submit ride completion U S E C A S E POST /... HTTP/1.1 ... { "customer": ..., "from-to" : ..., "distance": ..., "duration": ..., "fare" : ... } Wild Rydes unicorn

rights reserved. Submit ride completion U S E C A S E Ride management service POST /... HTTP/1.1 ... { "customer": ..., "from-to" : ..., "distance": ..., "duration": ..., "fare" : ... } Wild Rydes unicorn

rights reserved. Submit ride completion U S E C A S E Ride management service POST /... HTTP/1.1 ... { "customer": ..., "from-to" : ..., "distance": ..., "duration": ..., "fare" : ... } Managed API Gateway Wild Rydes unicorn

rights reserved. Submit ride completion U S E C A S E Ride management service POST /... HTTP/1.1 ... { "customer": ..., "from-to" : ..., "distance": ..., "duration": ..., "fare" : ... } Managed API Gateway Completed rides processor Wild Rydes unicorn

rights reserved. Submit ride completion U S E C A S E Ride management service POST /... HTTP/1.1 ... { "customer": ..., "from-to" : ..., "distance": ..., "duration": ..., "fare" : ... } Completed rides store Managed API Gateway Completed rides processor Wild Rydes unicorn

rights reserved. Submit ride completion U S E C A S E 201 Created ... { < new resource representation > } Ride management service POST /... HTTP/1.1 ... { "customer": ..., "from-to" : ..., "distance": ..., "duration": ..., "fare" : ... } Completed rides store Managed API Gateway Completed rides processor Wild Rydes unicorn

rights reserved. Submit ride completion U S E C A S E Ride management service Customer notification service 201 Created ... { < new resource representation > } POST /... HTTP/1.1 ... { "customer": ..., "from-to" : ..., "distance": ..., "duration": ..., "fare" : ... } Completed rides store Managed API Gateway Completed rides processor Wild Rydes unicorn

rights reserved. Submit ride completion U S E C A S E Ride management service Customer notification service Customer accounting service 201 Created ... { < new resource representation > } POST /... HTTP/1.1 ... { "customer": ..., "from-to" : ..., "distance": ..., "duration": ..., "fare" : ... } Completed rides store Managed API Gateway Completed rides processor Wild Rydes unicorn

rights reserved. Submit ride completion U S E C A S E Ride management service Customer notification service Customer accounting service Customer loyalty service 201 Created ... { < new resource representation > } POST /... HTTP/1.1 ... { "customer": ..., "from-to" : ..., "distance": ..., "duration": ..., "fare" : ... } Completed rides store Managed API Gateway Completed rides processor Wild Rydes unicorn

rights reserved. Submit ride completion U S E C A S E Ride management service Customer notification service Customer accounting service Customer loyalty service Data lake ingestion service 201 Created ... { < new resource representation > } POST /... HTTP/1.1 ... { "customer": ..., "from-to" : ..., "distance": ..., "duration": ..., "fare" : ... } Completed rides store Managed API Gateway Completed rides processor Wild Rydes unicorn

rights reserved. Submit ride completion U S E C A S E Ride management service Customer notification service Customer accounting service Customer loyalty service Data lake ingestion service Extraordinary rides service 201 Created ... { < new resource representation > } POST /... HTTP/1.1 ... { "customer": ..., "from-to" : ..., "distance": ..., "duration": ..., "fare" : ... } Completed rides store Managed API Gateway Completed rides processor Wild Rydes unicorn

rights reserved. Submit ride completion U S E C A S E Ride management service Customer notification service Customer accounting service Customer loyalty service Data lake Ingestion service Extraordinary rides service Interested in rides with: fare >= x distance >= y 201 Created ... { < new resource representation > } POST /... HTTP/1.1 ... { "customer": ..., "from-to" : ..., "distance": ..., "duration": ..., "fare" : ... } Completed rides store Managed API Gateway Completed rides processor Wild Rydes unicorn

rights reserved. Submit ride completion U S E C A S E Ride management service Customer notification service Managed API Gateway Customer accounting service Customer loyalty service Data lake ingestion service Extraordinary rides service Interested in rides with: fare >= x distance >= y 201 Created ... { < new resource representation > } POST /... HTTP/1.1 ... { "customer": ..., "from-to" : ..., "distance": ..., "duration": ..., "fare" : ... } Completed rides store Completed rides processor Integration via database? Wild Rydes unicorn

rights reserved. Submit ride completion U S E C A S E Ride management service Customer notification service Managed API Gateway Customer accounting service Customer loyalty service Data lake ingestion service Extraordinary rides service Interested in rides with: fare >= x distance >= y Wild Rydes unicorn app Photo: Dirk Fröhner 201 Created ... { < new resource representation > } POST /... HTTP/1.1 ... { "customer": ..., "from-to" : ..., "distance": ..., "duration": ..., "fare" : ... } Completed rides store Completed rides processor Integration via database? Oh my! Photo: Dirk Fröhner

rights reserved. Submit ride completion U S E C A S E Ride management service Customer notification service Managed API Gateway Customer accounting service Customer loyalty service Data lake ingestion service Extraordinary rides service Interested in rides with: fare >= x distance >= y 201 Created ... { < new resource representation > } POST /... HTTP/1.1 ... { "customer": ..., "from-to" : ..., "distance": ..., "duration": ..., "fare" : ... } Completed rides store Completed rides processor Integration via REST APIs? Wild Rydes unicorn

rights reserved. Submit ride completion U S E C A S E Ride management service Customer notification service Managed API Gateway Customer accounting service Customer loyalty service Data lake ingestion service Extraordinary rides service Interested in rides with: fare >= x distance >= y 201 Created ... { < new resource representation > } POST /... HTTP/1.1 ... { "customer": ..., "from-to" : ..., "distance": ..., "duration": ..., "fare" : ... } Completed rides store Completed rides processor Integration via REST APIs? Absolutely, but . . . Wild Rydes unicorn

rights reserved. Submit ride completion U S E C A S E Recipient-list Ride management service Customer notification service Customer accounting service Customer loyalty service Data lake ingestion service Extraordinary rides service https://... https://... https://... https://... https://... https://... 201 Created ... { < new resource representation > } POST /... HTTP/1.1 ... { "customer": ..., "from-to" : ..., "distance": ..., "duration": ..., "fare" : ... } Completed rides store Managed API Gateway Completed rides processor Wild Rydes unicorn

rights reserved. Submit ride completion U S E C A S E Recipient-list service Ride management service Customer notification service Customer accounting service Customer loyalty service Data lake ingestion service Extraordinary rides service https://... https://... https://... https://... https://... https://... 201 Created ... { < new resource representation > } POST /... HTTP/1.1 ... { "customer": ..., "from-to" : ..., "distance": ..., "duration": ..., "fare" : ... } Completed rides store Request distribution service Managed API Gateway Completed rides processor Wild Rydes unicorn

rights reserved. Submit ride completion U S E C A S E Self-managed filtering Customer notification service Customer accounting service Customer loyalty service Data lake ingestion service Extraordinary rides service https://... https://... https://... https://... https://... Interested in rides with: fare >= x distance >= y https://... 201 Created ... { < new resource representation > } POST /... HTTP/1.1 ... { "customer": ..., "from-to" : ..., "distance": ..., "duration": ..., "fare" : ... } Completed rides store Request distribution service Managed API Gateway Completed rides processor Wild Rydes unicorn Ride management service

rights reserved. Submit ride completion U S E C A S E Self-managed filtering Ride management service Customer notification service Request distribution service Managed API Gateway Customer accounting service Customer loyalty service Data lake ingestion service Extraordinary rides service https://... https://... https://... https://... https://... Interested in rides with: fare >= x distance >= y https://... 201 Created ... { < new resource representation > } POST /... HTTP/1.1 ... { "customer": ..., "from-to" : ..., "distance": ..., "duration": ..., "fare" : ... } Completed rides store Completed rides processor Integration via messaging? Absolutely! Wild Rydes unicorn

rights reserved. Submit ride completion U S E C A S E Publish-subscribe (topic) Ride management service Customer notification service Customer accounting service Customer loyalty service Data lake ingestion service Extraordinary rides service 201 Created ... { < new resource representation > } POST /... HTTP/1.1 ... { "customer": ..., "from-to" : ..., "distance": ..., "duration": ..., "fare" : ... } Completed rides store Ride completion topic Managed API Gateway Completed rides processor Wild Rydes unicorn

rights reserved. Submit ride completion U S E C A S E Message filter Ride management service Customer notification service Customer accounting service Customer loyalty service Data lake ingestion service Extraordinary rides service SNS message filter: fare >= x distance >= y 201 Created ... { < new resource representation > } POST /... HTTP/1.1 ... { "customer": ..., "from-to" : ..., "distance": ..., "duration": ..., "fare" : ... } Completed rides store Ride completion topic Managed API Gateway Completed rides processor Wild Rydes unicorn

rights reserved. Submit ride completion U S E C A S E Topic-queue-chaining Ride management service Customer notification service Customer accounting service Customer loyalty service Data lake ingestion service Extraordinary rides service SNS message filter: fare >= x distance >= y 201 Created ... { < new resource representation > } POST /... HTTP/1.1 ... { "customer": ..., "from-to" : ..., "distance": ..., "duration": ..., "fare" : ... } Completed rides store Ride completion topic Managed API Gateway Completed rides processor Wild Rydes unicorn

rights reserved. Prebooking campaigns U S E C A S E All at once Massive campaign with CTA: Book this week!

rights reserved. Prebooking campaigns U S E C A S E All at once Wild Rydes customer

rights reserved. Prebooking campaigns U S E C A S E All at once POST /... HTTP/1.1 ... { "customer": ..., "from" : ..., "to" : ..., "when" : ... } Wild Rydes customer

rights reserved. Ride booking service Prebooking campaigns U S E C A S E All at once AWS Cloud POST /... HTTP/1.1 ... { "customer": ..., "from" : ..., "to" : ..., "when" : ... } Wild Rydes customer

rights reserved. Ride booking service Prebooking campaigns U S E C A S E All at once AWS Cloud POST /... HTTP/1.1 ... { "customer": ..., "from" : ..., "to" : ..., "when" : ... } Managed API Gateway Wild Rydes customer

rights reserved. Ride booking service Prebooking campaigns U S E C A S E All at once AWS Cloud POST /... HTTP/1.1 ... { "customer": ..., "from" : ..., "to" : ..., "when" : ... } Managed API Gateway Prebooking processing resource Wild Rydes customer

rights reserved. Ride booking service Prebooking campaigns U S E C A S E All at once AWS Cloud POST /... HTTP/1.1 ... { "customer": ..., "from" : ..., "to" : ..., "when" : ... } Managed API Gateway Prebooking processing resource Prebooking data store Wild Rydes customer

rights reserved. Ride booking service Prebooking campaigns U S E C A S E All at once AWS Cloud 201 Created ... { < prebooking representation > } POST /... HTTP/1.1 ... { "customer": ..., "from" : ..., "to" : ..., "when" : ... } Managed API Gateway Prebooking processing resource Prebooking data store Wild Rydes customer

rights reserved. Ride booking service Prebooking campaigns U S E C A S E All at once AWS Cloud 201 Created ... { < prebooking representation > } POST /... HTTP/1.1 ... { "customer": ..., "from" : ..., "to" : ..., "when" : ... } Managed API Gateway Prebooking Processing Resource Massive campaign with CTA: Book today! Prebooking Data Store Wild Rydes customer

rights reserved. Prebooking campaigns U S E C A S E Internal decoupling Wild Rydes customer

rights reserved. Prebooking campaigns U S E C A S E Internal decoupling POST /... HTTP/1.1 ... { "customer": ..., "from" : ..., "to" : ..., "when" : ... } Wild Rydes customer

rights reserved. Ride booking service Prebooking campaigns U S E C A S E Internal decoupling AWS Cloud POST /... HTTP/1.1 ... { "customer": ..., "from" : ..., "to" : ..., "when" : ... } Wild Rydes customer

rights reserved. Ride booking service Prebooking campaigns U S E C A S E Internal decoupling AWS Cloud POST /... HTTP/1.1 ... { "customer": ..., "from" : ..., "to" : ..., "when" : ... } Managed API Gateway Wild Rydes customer

rights reserved. Ride booking service Prebooking campaigns U S E C A S E Internal decoupling AWS Cloud Prebooking preprocessing resource POST /... HTTP/1.1 ... { "customer": ..., "from" : ..., "to" : ..., "when" : ... } Managed API Gateway Wild Rydes customer

rights reserved. Ride booking service Prebooking campaigns U S E C A S E Internal decoupling AWS Cloud Prebooking forwarding topic Prebooking preprocessing resource POST /... HTTP/1.1 ... { "customer": ..., "from" : ..., "to" : ..., "when" : ... } Managed API Gateway Wild Rydes customer

rights reserved. Ride booking service Prebooking campaigns U S E C A S E Internal decoupling AWS Cloud Prebooking forwarding topic Prebooking preprocessing resource POST /... HTTP/1.1 ... { "customer": ..., "from" : ..., "to" : ..., "when" : ... } Managed API Gateway 202 Accepted ... { < task & status representation > } Wild Rydes customer

rights reserved. Ride booking service Prebooking campaigns U S E C A S E Internal decoupling AWS Cloud 202 Accepted ... { < task & status representation > } Prebooking processing queue Prebooking processing resource Prebooking forwarding topic Prebooking preprocessing resource POST /... HTTP/1.1 ... { "customer": ..., "from" : ..., "to" : ..., "when" : ... } Managed API Gateway Ride booking service Wild Rydes customer

rights reserved. Ride booking service Prebooking campaigns U S E C A S E Internal decoupling AWS Cloud 202 Accepted ... { < task & status representation > } Data lake ingestion service Data lake ingestion processing Data lake ingestion queue Prebooking processing queue Prebooking processing resource Prebooking forwarding topic Prebooking preprocessing resource POST /... HTTP/1.1 ... { "customer": ..., "from" : ..., "to" : ..., "when" : ... } Managed API Gateway Ride booking service Wild Rydes customer

rights reserved. Ride booking service Prebooking campaigns U S E C A S E Internal decoupling AWS Cloud 202 Accepted ... { < task & status representation > } Data lake ingestion service Data lake ingestion processing Data lake ingestion queue Prebooking processing queue Prebooking processing resource Prebooking forwarding topic Prebooking preprocessing resource POST /... HTTP/1.1 ... { "customer": ..., "from" : ..., "to" : ..., "when" : ... } Managed API Gateway Ride booking service Wild Rydes customer No space left for the database icon

rights reserved. Ride booking service Prebooking campaigns U S E C A S E Internal decoupling AWS Cloud 202 Accepted ... { < task & status representation > } Data lake ingestion service Data lake ingestion processing Data lake ingestion queue Prebooking processing queue Prebooking processing resource Prebooking forwarding topic Prebooking preprocessing resource POST /... HTTP/1.1 ... { "customer": ..., "from" : ..., "to" : ..., "when" : ... } Managed API Gateway Ride booking service Wild Rydes customer

rights reserved. Ride booking service Prebooking campaigns U S E C A S E Internal decoupling AWS Cloud 202 Accepted ... { < task & status representation > } Data lake ingestion service Data lake ingestion processing Data lake ingestion queue Prebooking processing queue Prebooking processing resource Prebooking forwarding topic Prebooking preprocessing resource POST /... HTTP/1.1 ... { "customer": ..., "from" : ..., "to" : ..., "when" : ... } Managed API Gateway Ride booking service Wild Rydes customer Massive campaign with CTA: Book now!

rights reserved. Ride booking service Prebooking campaigns U S E C A S E Even more internal decoupling AWS Cloud 200 OK ... <SendMessageResponse> ... </SendMessageResponse> POST /... HTTP/1.1 ... Action=SendMessage &MessageBody=... Prebooking forwarding topic Prebooking preprocessing resource Prebooking submission queue Data lake ingestion service Ride booking service Data lake ingestion processing Data lake ingestion queue Prebooking processing queue Prebooking processing resource Wild Rydes customer

rights reserved. Ride booking service Prebooking campaigns U S E C A S E Even more internal decoupling AWS Cloud 202 Accepted ... { < task & status representation > } Prebooking forwarding topic Prebooking preprocessing resource Prebooking submission queue Managed API Gateway Data lake ingestion service Ride booking service Data lake ingestion processing Data lake ingestion queue Prebooking processing queue Prebooking processing resource POST /... HTTP/1.1 ... { "customer": ..., "from" : ..., "to" : ..., "when" : ... } Wild Rydes customer

rights reserved. Ride booking service Prebooking campaigns U S E C A S E Even more internal decoupling AWS Cloud 202 Accepted ... { < task & status representation > } Prebooking forwarding topic Prebooking preprocessing resource Prebooking submission queue Managed API Gateway Ride booking service Prebooking processing queue Prebooking processing resource POST /... HTTP/1.1 ... { "customer": ..., "from" : ..., "to" : ..., "when" : ... } Data lake ingestion service Data lake ingestion processing Data lake ingestion queue Wild Rydes customer

rights reserved. Prebooking campaigns U S E C A S E Stream for datalake ingestion AWS Cloud 202 Accepted ... { < task & status representation > } Modern data platform Data lake raw data tier Data lake ingestion stream Prebooking forwarding topic Prebooking preprocessing resource Prebooking submission queue Managed API Gateway Ride booking service Ride booking service Prebooking processing queue Prebooking processing resource POST /... HTTP/1.1 ... { "customer": ..., "from" : ..., "to" : ..., "when" : ... } Wild Rydes customer

rights reserved. How to distribute a request to relevant/interested parties and capture their individual responses? Message routing Scatter-gather Requester Requester Topic A A A A B1 For election or parallel processing scenarios, i.e., search for best response or accumulate responses Responders Processor Aggregator B3 B1 B2 B3 B2 Queue Remember the scatter-gather pattern?

rights reserved. How to distribute a request to relevant/interested parties and capture their individual responses? Message routing Scatter-gather Requester Requester Topic A A A A B1 For election or parallel processing scenarios, i.e., search for best response or accumulate responses Responders Processor Aggregator B3 B1 B2 B3 B2 Queue

rights reserved. Instant ride RFQ U S E C A S E Wild Rydes customer Scatter-gather

rights reserved. Instant ride RFQ U S E C A S E Scatter-gather POST /<submit> HTTP/1.1 ... { "customer": ..., "from" : ..., "to" : ... } Wild Rydes customer

rights reserved. Instant ride RFQ U S E C A S E Ride booking service Scatter-gather POST /<submit> HTTP/1.1 ... { "customer": ..., "from" : ..., "to" : ... } Wild Rydes customer

rights reserved. Instant ride RFQ U S E C A S E Ride booking service API Gateway Scatter-gather POST /<submit> HTTP/1.1 ... { "customer": ..., "from" : ..., "to" : ... } Wild Rydes customer

rights reserved. Instant ride RFQ U S E C A S E Ride booking service RFQ intake processor Scatter-gather POST /<submit> HTTP/1.1 ... { "customer": ..., "from" : ..., "to" : ... } Wild Rydes customer API Gateway

rights reserved. Instant ride RFQ U S E C A S E Ride booking service RFQ intake processor RFQ store Scatter-gather POST /<submit> HTTP/1.1 ... { "customer": ..., "from" : ..., "to" : ... } Wild Rydes customer API Gateway

rights reserved. Instant ride RFQ U S E C A S E Ride booking service RFQ share topic RFQ intake processor RFQ store Scatter-gather POST /<submit> HTTP/1.1 ... { "customer": ..., "from" : ..., "to" : ... } Wild Rydes customer API Gateway

rights reserved. Instant ride RFQ U S E C A S E Ride booking service RFQ share topic RFQ intake processor RFQ store Scatter-gather POST /<submit> HTTP/1.1 ... { "customer": ..., "from" : ..., "to" : ... } HTTP/1.1 202 Accepted ... { "links": { <link-to-refresh> }, "status": "running", "eta" : ... } Wild Rydes customer API Gateway

rights reserved. Instant ride RFQ U S E C A S E Ride booking service RFQ share topic RFQ intake processor Unicorn management service RFQ store Scatter-gather POST /<submit> HTTP/1.1 ... { "customer": ..., "from" : ..., "to" : ... } HTTP/1.1 202 Accepted ... { "links": { <link-to-refresh> }, "status": "running", "eta" : ... } Wild Rydes customer API Gateway

rights reserved. Instant ride RFQ U S E C A S E Ride booking service RFQ share topic RFQ Intake Processor Unicorn management service RFQ store . . . Scatter-gather POST /<submit> HTTP/1.1 ... { "customer": ..., "from" : ..., "to" : ... } HTTP/1.1 202 Accepted ... { "links": { <link-to-refresh> }, "status": "running", "eta" : ... } UMR UMR UMR UMR: Unicorn management resource Wild Rydes customer API Gateway

rights reserved. Instant ride RFQ U S E C A S E Ride booking service RFQ share topic RFQ intake processor Unicorn management service RFQ store . . . Scatter-gather POST /<submit> HTTP/1.1 ... { "customer": ..., "from" : ..., "to" : ... } HTTP/1.1 202 Accepted ... { "links": { <link-to-refresh> }, "status": "running", "eta" : ... } UMR UMR UMR UMR: Unicorn management resource Wild Rydes customer API Gateway

rights reserved. Instant ride RFQ U S E C A S E Ride booking service RFQ share topic RFQ intake processor Unicorn management service RFQ store Unicorn n Unicorn 2 Unicorn 1 . . . Scatter-gather POST /<submit> HTTP/1.1 ... { "customer": ..., "from" : ..., "to" : ... } HTTP/1.1 202 Accepted ... { "links": { <link-to-refresh> }, "status": "running", "eta" : ... } UMR UMR UMR UMR: Unicorn management resource Wild Rydes customer API Gateway

rights reserved. Instant ride RFQ U S E C A S E Ride booking service RFQ share topic RFQ intake processor Unicorn management service RFQ store RFQ response queue Unicorn n Unicorn 2 Unicorn 1 . . . Scatter-gather POST /<submit> HTTP/1.1 ... { "customer": ..., "from" : ..., "to" : ... } HTTP/1.1 202 Accepted ... { "links": { <link-to-refresh> }, "status": "running", "eta" : ... } UMR UMR UMR UMR: Unicorn management resource Wild Rydes customer API Gateway

rights reserved. Instant ride RFQ U S E C A S E Ride booking service RFQ share topic RFQ intake processor Unicorn management service RFQ store RFQ response queue Response aggregator Unicorn n Unicorn 2 Unicorn 1 . . . Scatter-gather POST /<submit> HTTP/1.1 ... { "customer": ..., "from" : ..., "to" : ... } HTTP/1.1 202 Accepted ... { "links": { <link-to-refresh> }, "status": "running", "eta" : ... } UMR UMR UMR UMR: Unicorn management resource Wild Rydes customer API Gateway

rights reserved. Instant ride RFQ U S E C A S E Ride booking service RFQ share topic RFQ intake processor Unicorn management service RFQ store RFQ response queue Response aggregator Unicorn n Unicorn 2 Unicorn 1 . . . Scatter-gather POST /<submit> HTTP/1.1 ... { "customer": ..., "from" : ..., "to" : ... } HTTP/1.1 202 Accepted ... { "links": { <link-to-refresh> }, "status": "running", "eta" : ... } UMR UMR UMR Wild Rydes customer API Gateway What can the client do?

rights reserved. Instant ride RFQ U S E C A S E Ride booking service RFQ share topic RFQ intake processor Unicorn management service RFQ store RFQ response queue Response aggregator Unicorn n Unicorn 2 Unicorn 1 . . . Scatter-gather POST /<submit> HTTP/1.1 ... { "customer": ..., "from" : ..., "to" : ... } HTTP/1.1 202 Accepted ... { "links": { <link-to-refresh> }, "status": "running", "eta" : ... } UMR UMR UMR Wild Rydes customer API Gateway What can the client do? Retrieve a status update

rights reserved. Instant ride RFQ U S E C A S E Ride booking service RFQ share topic RFQ intake processor Unicorn management service RFQ store RFQ response queue Response aggregator Unicorn n Unicorn 2 Unicorn 1 . . . Scatter-gather POST /<submit> HTTP/1.1 ... { "customer": ..., "from" : ..., "to" : ... } HTTP/1.1 202 Accepted ... { "links": { <link-to-refresh> }, "status": "running", "eta" : ... } UMR UMR UMR Wild Rydes customer API Gateway

rights reserved. Instant ride RFQ U S E C A S E Ride booking service RFQ share topic RFQ Intake processor Unicorn management service RFQ store RFQ response queue Response aggregator Unicorn n Unicorn 2 Unicorn 1 . . . Scatter-gather GET /<refresh> HTTP/1.1 ... UMR UMR UMR Wild Rydes customer API Gateway

rights reserved. Instant ride RFQ U S E C A S E Ride booking service RFQ share topic RFQ intake processor Unicorn management service RFQ store RFQ response queue Response aggregator Unicorn n Unicorn 2 Unicorn 1 . . . Scatter-gather GET /<refresh> HTTP/1.1 ... UMR UMR UMR Wild Rydes customer API Gateway

rights reserved. Instant ride RFQ U S E C A S E Ride booking service RFQ share topic RFQ intake processor Unicorn management service RFQ store RFQ result processor RFQ response queue Response aggregator Unicorn n Unicorn 2 Unicorn 1 . . . Scatter-gather GET /<refresh> HTTP/1.1 ... UMR UMR UMR Wild Rydes customer API Gateway

rights reserved. Instant ride RFQ U S E C A S E Ride booking service RFQ share topic RFQ intake processor Unicorn management service RFQ store RFQ result processor RFQ response queue Response aggregator Unicorn n Unicorn 2 Unicorn 1 . . . Scatter-gather GET /<refresh> HTTP/1.1 ... HTTP/1.1 200 OK ... { "links": { <link-to-refresh> }, "status": "running", "eta" : ... } UMR UMR UMR Wild Rydes customer API Gateway

rights reserved. Instant ride RFQ U S E C A S E Ride booking service RFQ share topic RFQ intake processor Unicorn management service RFQ store RFQ result processor RFQ response queue Response aggregator Unicorn n Unicorn 2 Unicorn 1 . . . Scatter-gather GET /<refresh> HTTP/1.1 ... HTTP/1.1 200 OK ... { "links": { <link-to-refresh> }, "status": "running", "eta" : ... } UMR UMR UMR Wild Rydes customer API Gateway What can the client do? Retrieve another status update

rights reserved. Instant ride RFQ U S E C A S E Ride booking service RFQ share topic RFQ intake processor Unicorn management service RFQ store RFQ response queue Response aggregator Unicorn n Unicorn 2 Unicorn 1 . . . Scatter-gather GET /<refresh> HTTP/1.1 ... RFQ result processor UMR UMR UMR Wild Rydes customer API Gateway

rights reserved. Instant ride RFQ U S E C A S E Ride booking service RFQ share topic RFQ intake processor Unicorn management service RFQ store RFQ result processor RFQ response queue Response aggregator Unicorn n Unicorn 2 Unicorn 1 . . . Scatter-gather GET /<refresh> HTTP/1.1 ... UMR UMR UMR Wild Rydes customer API Gateway

rights reserved. Instant ride RFQ U S E C A S E Ride booking service RFQ share topic RFQ intake processor Unicorn management service RFQ store RFQ result processor RFQ response queue Response aggregator Unicorn n Unicorn 2 Unicorn 1 . . . Scatter-gather GET /<refresh> HTTP/1.1 ... HTTP/1.1 200 OK ... { "links": { <link-to-result> }, "status": "done" } UMR UMR UMR Wild Rydes customer API Gateway

rights reserved. Instant ride RFQ U S E C A S E Ride booking service RFQ share topic RFQ intake processor Unicorn management service RFQ store RFQ result processor RFQ response queue Response aggregator Unicorn n Unicorn 2 Unicorn 1 . . . Scatter-gather GET /<refresh> HTTP/1.1 ... HTTP/1.1 200 OK ... { "links": { <link-to-result> }, "status": "done" } UMR UMR UMR Wild Rydes customer API Gateway What can the client do? Retrieve RFQ result

rights reserved. Instant ride RFQ U S E C A S E Ride booking service RFQ share topic RFQ intake processor Unicorn management service RFQ store RFQ response queue Response aggregator Unicorn n Unicorn 2 Unicorn 1 . . . Scatter-gather GET /<result> HTTP/1.1 ... RFQ result processor UMR UMR UMR Wild Rydes customer API Gateway

rights reserved. Instant ride RFQ U S E C A S E Ride booking service RFQ share topic RFQ intake processor Unicorn management service RFQ store RFQ result processor RFQ response queue Response aggregator Unicorn n Unicorn 2 Unicorn 1 . . . Scatter-gather GET /<result> HTTP/1.1 ... UMR UMR UMR Wild Rydes customer API Gateway

rights reserved. Instant ride RFQ U S E C A S E Ride booking service RFQ share topic RFQ intake processor Unicorn management service RFQ store RFQ result processor RFQ response queue Response aggregator Unicorn n Unicorn 2 Unicorn 1 . . . Scatter-gather GET /<result> HTTP/1.1 ... HTTP/1.1 200 OK ... { "links": { <links> }, "ride-data": ..., "quotes" : ... } UMR UMR UMR Wild Rydes customer API Gateway

rights reserved. Resources and call to action AWS blog series Read again about the use cases and patterns from this talk (and more)

rights reserved. Resources and call to action AWS blog series Read again about the use cases and patterns from this talk (and more) AWS hands-on workshop Gain hands-on experience with building the use cases from this talk (and more)

rights reserved. Resources and call to action AWS blog series Read again about the use cases and patterns from this talk (and more) AWS hands-on workshop Gain hands-on experience with building the use cases from this talk (and more) Reach out to your friendly AWS solutions architect

rights reserved. Resources and call to action AWS blog series Read again about the use cases and patterns from this talk (and more) AWS hands-on workshop Gain hands-on experience with building the use cases from this talk (and more) Reach out to your friendly AWS solutions architect and keep in mind

rights reserved. Resources and call to action AWS blog series Read again about the use cases and patterns from this talk (and more) AWS hands-on workshop Gain hands-on experience with building the use cases from this talk (and more) Reach out to your friendly AWS solutions architect and keep in mind Loose coupling is always better than lousy coupling

2022-09-22.jcon.integration-and-conversation-patterns.foundations.pdf

2022-09-22.jcon.integration-and-conversation-patterns.foundations.pdf

More Decks by Dirk Fröhner

Featured

Transcript