TalksAWS re:Invent 2025 - [NEW LAUNCH] Deep Dive on AWS Lambda durable functions (CNS380)
AWS re:Invent 2025 - [NEW LAUNCH] Deep Dive on AWS Lambda durable functions (CNS380)
Summary of AWS re:Invent 2025 - Deep Dive on AWS Lambda Durable Functions (CNS380)
Introduction
Presenters: Eric Johnson (Principal Developer Advocate) and Michael (Product Manager)
This session focused on the newly announced AWS Lambda Durable Functions, a significant evolution in serverless computing.
The Serverless Journey
The presenters discussed the evolution from monolithic architectures to microservices and the challenges faced, such as complexity and coordination issues.
AWS introduced serverless computing with AWS Lambda, providing stateless, short-lived compute without infrastructure management.
To address the need for orchestration, AWS introduced AWS Step Functions in 2016 and Amazon EventBridge in 2019 for event-driven architectures.
However, developers still faced challenges with application logic orchestration, leading to the development of AWS Lambda Durable Functions.
AWS Lambda Durable Functions
Durable Functions are regular AWS Lambda functions with added "superpowers" for reliable, long-running, and stateful execution.
Key features of Durable Functions:
Checkpointing: Allows functions to checkpoint progress, enabling recovery from failures and suspension of execution.
Replay: Replays the function's event handler from the top, skipping over completed checkpoints to resume execution.
Durable Execution SDK: Provides higher-level abstractions like "steps" and "weights" to simplify the use of checkpointing and replay.
Durable Functions support various runtimes, including Node.js, Python, and OCI (Open Container Initiative) containers.
They integrate with existing AWS Lambda event sources, synchronous and asynchronous invocations, and can invoke other Lambda functions (durable or non-durable).
Technical Deep Dive
Checkpoint and Replay: Durable Functions use a "durable execution" model, where the function's progress is checkpointed, and upon failure or suspension, the function is replayed from the last checkpoint.
Weights and Suspension: Durable Functions can use "weights" to suspend execution and wait for external events, callbacks, or conditions before resuming.
Compensation and Undo: Durable Functions make it easier to implement compensation logic (undo operations) using the step-based execution model.
Development and Testing
Durable Functions integrate with AWS SAM (Serverless Application Model) and other infrastructure-as-code tools like AWS CloudFormation and AWS CDK.
The SDK provides a testing framework for both local and cloud-based integration testing, leveraging mocking and assertion capabilities.
Pricing and Comparison to AWS Step Functions
Durable Functions pricing is based on the number of operations (checkpoints and weights), data storage, and existing Lambda compute charges.
The presenters discussed the differences between Durable Functions and AWS Step Functions, highlighting that Durable Functions are more focused on application logic orchestration, while Step Functions are better suited for workflow orchestration across AWS services.
Conclusion and Feedback
The presenters encouraged the audience to try out Durable Functions, provide feedback, and reach out through AWS forums, support channels, and the open-source SDKs.
They emphasized the importance of building applications like "monoliths" (single-screen applications) and deploying them as microservices, taking advantage of the reliability and flexibility offered by Durable Functions.
These cookies are used to collect information about how you interact with this website and allow us to remember you. We use this information to improve and customize your browsing experience, as well as for analytics.
If you decline, your information won’t be tracked when you visit this website. A single cookie will be used in your browser to remember your preference.