Summary of Distributed Sagas: A Protocol for Coordinating Microservices - Caitie McCaffrey - JOTB17

00:00:00 - 00:40:00

In the video, Caitie McCaffrey discusses the distributed saga protocol for coordinating microservices. The protocol is designed to be easier to implement and use, and has a number of trade-offs that depends on the specific architecture and implementation.

• 00:00:00 The speaker discusses the benefits of microservice architectures, and how distributed sagas can help coordinate these services. She also discusses how ferrel concurrency control can be implemented within microservices to ensure that application invariants are upheld.
• 00:05:00 Distributed sagas are a protocol for coordinating microservices, which can be difficult to build and maintain. Spanner is a globally distributed database that Google claims to have achieved five nines of availability.
• 00:10:00 Distributed sagas are a protocol for coordinating microservices, which can help to improve availability and latency.
• 00:15:00 Distributed sagas are a protocol for coordinating microservices. They allow requests to be queued and executed in a serial or parallel fashion, and can be used to ensure that requests are completed successfully.
• 00:20:00 The video discusses a protocol for coordinating microservices, dubbed the "distributed saga." The protocol allows for requests to be executed in a distributed manner without the need for atomicity or isolation. The protocol is very durable, persist across failures, and requires only a single process to execute requests.
• 00:25:00 In this video, Caitie McCaffrey explains how a distributed sagas protocol can be used to coordinate microservices. A saga is a collection of messages between microservices that describes a unit of work. Each node in the graph that is referenced by a parent microservice is complete, which means the microservice can start executing the requested operation. Before any requests can be processed, a start hotel message must be logged to the saga log. This has to happen before any other requests can be processed. Once the start hotel message has been logged, the car note can be processed. Once the car note is completed, the flight node can be worked on. Once the flight node is completed, the book car request message can be processed. Once the book car request message is processed, the flight node and the car service can be marked as completed. The flight log can then be updated accordingly. In the event that a request fails, a rollback can be initiated. The rollback will start from the bottom of the graph and work its way to the top. Once the rollback has completed, the system will be in a state where all requests have successfully executed. Finally, the cancellation of the flight reservation can be logged to the
• 00:30:00 Distributed sagas are a protocol for coordinating Microservices. They help maintain state between Microservices and can be used to compose services together.
• 00:35:00 The presenter discusses a protocol for coordinating microservices, which is based on the concept of distributed sagas. The protocol is designed to be easier to implement and use, and has a number of trade-offs that depends on the specific architecture and implementation.
• 00:40:00 Caitie McCaffrey provides a protocol for coordinating microservices, which involves storing state in a local cache and using two-phase commit to ensure that the state is always consistent. If a service fails, the rest of the services can still operate normally by using hot failover.