We know that watching all the re:Invent session videos can be a daunting task, but we don't want you to miss out on the gold that is often found in them! In this blog, you can find quick summaries of all the 300 and 400 level sessions, grouped by track. Enjoy!
OPN305 The pragmatic serverless Python developer
The presentation at AWS re:Invent 2023, led by Heitor Lessa, Chief Architect for Powertools for AWS, and Ran Isenberg, a principal software architect at CyberArk, focused on pragmatic approaches for serverless Python developers. They discussed the importance of project structure, clear responsibilities for Lambda handlers, and the separation of business domain and integration code. Emphasizing the significance of testing, they explored unit, integration, and end-to-end tests, advocating for a development environment that stays within the IDE and uses real cloud resources. They demonstrated testing techniques for both synchronous and asynchronous processes, stressing the value of understanding the most common code paths and resource usage.
In terms of production readiness, the session highlighted AWS Powertools, a toolset designed to simplify serverless applications with features like data validation, API routing, and open API integration. They discussed tools like Tuna, PySpy, and PyInstrument for analyzing code start-up time and resource usage, allowing developers to make informed decisions about potential optimizations. The speakers also introduced PyBenchmark, a tool to automate benchmarking processes.
As a practical takeaway, Lessa and Isenberg provided a GitHub repository with documented code and decisions made during their project. They also offered a template for initializing projects with best practices already integrated. The session concluded with an invitation to continue discussions at the AWS booth and a reminder to provide feedback through the session survey.
AWS re:Invent 2023 - The pragmatic serverless Python developer (OPN305)
API302 Building next-generation applications with event-driven architecture
In the AWS re:Invent 2023 session on "Building next-generation applications with event-driven architecture," Eric Johnson, a principal developer advocate for serverless at AWS, delves into the fundamentals of event-driven architecture (EDA) and its practical implementation. He begins by sharing his extensive experience in development and architecture, emphasizing his passion for infrastructure as code and automation. Johnson humorously introduces his session rules, including a unique approach to counting and quotation gestures. He then highlights the importance of understanding coupling in system integration, explaining how tightly coupled systems can lead to dependencies and issues in scalability and maintenance.
Johnson discusses various integration patterns, including synchronous request-response models and asynchronous point-to-point queues, using Amazon SQS as an example. He explains the benefits and drawbacks of these patterns, emphasizing the importance of decoupling and efficient message processing. Johnson also introduces Amazon EventBridge, a service that facilitates event-driven architecture by allowing different AWS services and custom applications to communicate through events. He explains EventBridge's functionality, demonstrating how it can simplify application architecture by handling event routing through rules, thereby reducing the need for complex routing logic in the application code.
The session concludes with a focus on idempotency in distributed systems, which ensures that operations have the same effect, regardless of how many times they are performed. Johnson shares a personal anecdote to illustrate the issues caused by duplicates in systems and emphasizes the significance of idempotency in avoiding such problems. He explains various strategies to achieve idempotency, such as using tokens and careful naming conventions to avoid duplication. Johnson also highlights AWS services like SQS FIFO queues and Step Functions, which inherently support idempotent operations. The talk ends with resources for further learning and encouragement to provide feedback on the session.
AWS re:Invent 2023 - Building next-generation applications with event-driven architecture (API302)
API303 Navigating the journey to serverless event-driven architecture
In this session at AWS re:Invent 2023, Dave Boyne, a Developer Advocate at AWS, delved into the journey towards adopting serverless event-driven architecture. He emphasized the importance of understanding behavior before implementation, highlighting tools like event storming for visualizing system behavior and identifying events, boundaries, and common language. Boyne discussed various architectural and messaging patterns, including point-to-point communication, publish-subscribe, and event streaming, and how services like AWS Lambda, EventBridge, and DynamoDB streams facilitate event-driven architecture.
Boyne also covered operational and maintenance aspects of event-driven architecture. He stressed the significance of establishing standards for events, ensuring discoverability, and maintaining effective communication within teams. Tools like EventBridge schema discovery and ACNKP for documenting event-driven architectures were mentioned. He underscored the concept of "event first thinking," treating events as first-class citizens, and the need for continuous evaluation and adaptation of event schemers.
The session concluded with Boyne emphasizing the evolution and complexity inherent in event-driven architectures. He suggested embracing the ubiquitous language across domains, understanding the coupling between components, and utilizing AWS services effectively for managing events. He also highlighted the importance of resources like the free service badge for serverless learning and Eda visuals for a visual learning approach to event-driven solutions. Boyne encouraged attendees to continue exploring and innovating within the realm of event-driven architecture.
AWS re:Invent 2023 - Navigating the journey to serverless event-driven architecture (API303)
API309 Advanced integration patterns & tradeoffs for loosely coupled systems
The AWS re:Invent 2023 session, led by Dirk, a principal solutions architect at AWS, and Gregor, from the serverless engineering team, delved into advanced integration patterns and the trade-offs involved in building loosely coupled systems. They highlighted how integration patterns, like synchronous messaging, continue to be relevant today. They used real-world analogies, like the operation of a coffee shop, to explain queuing and two-phase commits. The discussion also included the evolution from monolithic architectures to more flexible, service-oriented architectures, emphasizing the importance of integration for effective communication between different services, whether they are closely or loosely coupled.
The session further explored the complexities of managing control flow and flow control in distributed systems. They explained the significance of queues in achieving loose coupling by allowing independent control flows and managing the flow of messages. The speakers used practical examples, such as the integration of Amazon EventBridge and FIFO (First-In-First-Out) queues, to demonstrate how different AWS services handle control flow and flow control, including considerations for message ordering, delivery semantics, and error handling. They emphasized the importance of understanding the underlying control flow and its impact on system behavior, particularly in scenarios involving retries and error handling.
Lastly, Dirk and Gregor discussed the importance of cloud automation in implementing these integration patterns. They argued that automation in the cloud is more than just a convenience; it's essential for defining the architecture of modern applications. They proposed a new way of thinking about cloud automation, not just as infrastructure provisioning but as a means to define application architecture, including aspects like flow control, error handling, and retries. The session concluded with a call to embrace failure as an inevitable aspect of distributed systems, advocating for the use of patterns like dead-letter queues and idempotent consumer implementations to handle failures gracefully.
AWS re:Invent 2023 - Advanced integration patterns & trade-offs for loosely coupled systems (API309)
API310 Scale interactive data analysis with Step Functions Distributed Map
This presentation at AWS re:Invent 2023 focused on scaling interactive data analysis using AWS Step Functions Distributed Map. Adam Wagner from AWS and Roberto Iturralde from Vertex Pharmaceuticals shared insights on the application and benefits of this technology. The talk began with an introduction to the challenges of handling large-scale data processing, particularly in the context of processing half a million invoice files. Adam outlined the limitations of traditional approaches and highlighted the advantages of distributed processing, including speed, scalability, and fault tolerance.
Roberto Iturralde shared a case study from Vertex Pharmaceuticals, where they used AWS Step Functions Distributed Map for efficient analysis of thousands of images generated during drug discovery experiments. The previous system, based on a static server fleet, faced challenges with scalability, cost, and maintenance. The new serverless solution significantly improved scalability and reduced costs, achieving 11 times faster processing at 90% lower costs. This system not only enhanced efficiency but also offered better visibility into operations and error handling.
The session concluded with Adam highlighting the versatility of Step Functions Distributed Map across various industries, including life sciences, financial modeling, and unstructured file processing. He emphasized the ease of use, rapid startup times, and cost-effectiveness of the tool. Adam also recommended resources for further learning and encouraged attendees to explore AWS Lambda Powertools for additional functionalities in Lambda functions. The session ended with a Q&A session and a reminder to complete the session survey.
AWS re:Invent 2023 - Scale interactive data analysis with Step Functions Distributed Map (API310)
API401 Advanced serverless workflow patterns and best practices
The AWS re:Invent 2023 presentation on "Advanced Serverless Workflow Patterns and Best Practices (API401)" by Ben Smith, a principal developer advocate at AWS, focused on serverless architecture, specifically AWS Step Functions. Smith emphasized the utility and efficiency of AWS Step Functions in building serverless applications. He discussed various patterns and practices, such as using Step Functions for orchestrating AWS services, handling asynchronous tasks, and managing failures. Smith also highlighted the cost-effectiveness of Step Functions, explaining how to optimize workflows for cost and performance.
The presentation also delved into practical aspects of using AWS Step Functions, demonstrating patterns like the Saga and Circuit Breaker for managing complex workflows and handling failures. Smith introduced new features like the HTTP state, enabling direct integration with external APIs and the ability to re-run failed workflow states. He shared examples from real-world AWS applications, such as Serverless Video and Serverless Land, to illustrate how these patterns and practices are applied in production environments.
Lastly, Smith updated the audience on the latest AWS Step Functions features, including the Distributed Map state for enhanced parallel processing and integration with Bedrock for invoking models directly from Step Functions. He also announced the availability of the Step Functions Workflow Studio in IDEs through App Composer, facilitating more accessible and integrated workflow design. These updates underscore AWS's commitment to evolving serverless architecture, making it more robust and user-friendly for developers.
AWS re:Invent 2023 - Advanced serverless workflow patterns and best practices (API401)
SVS301 Building APIs: Choosing the best API solution and strategy for workloads
The presentation at AWS re:Invent 2023 focused on building APIs and choosing the best API solution and strategy for workloads. The speakers, Josh Khan and AY Jonathan, discussed the shift towards modern, modular architectures that offer improved scalability, performance, and innovation. They emphasized the importance of using APIs to integrate modular systems or to provide data to applications like mobile or web apps. The talk aimed to guide through the design decisions in building real-life systems, highlighting the immediate applicability of the concepts discussed.
The session delved into various aspects of API design, considering both technical and non-technical factors. The speakers introduced different API architectural styles like REST, GraphQL, gRPC, and AYAP, each suited for specific scenarios and needs. The choice of an API style depended on several factors, including access patterns, API visibility, SLA requirements, API topology, and team skills. Josh and AY demonstrated these principles through a practical example – building an extension or plugin for a larger application, involving a leaderboard service with multiple APIs.
Towards the end, they presented a comprehensive architecture for the leaderboard plugin, utilizing different API styles and AWS services. This included using Amazon EventBridge for asynchronous metrics ingestion, Amazon VPC Lattice for an internal BPC-based metrics tracking service, and AWS AppSync for a public-facing GraphQL API. They concluded by emphasizing a product mindset for API design and encouraged the use of managed services and infrastructure as code for operational simplicity and standardization of best practices. They also provided resources for further learning and implementation examples on GitHub.
AWS re:Invent 2023 - Building APIs: Choosing the best API solution & strategy for workloads (SVS301)
SVS305 Refactoring to Serverless
The AWS re:Invent 2023 session on "Refactoring to Serverless" focused on how serverless technology can not only relieve concerns about infrastructure but also reduce the need for application code. The speakers, Sindhu and Gregor, emphasized that serverless refactoring involves shifting from writing and maintaining large amounts of application code to leveraging automation code. This approach not only streamlines development processes but also enhances the robustness and scalability of applications. They introduced the concept of using AWS services like Lambda Destinations and EventBridge Pipes to automate processes that were traditionally handled by application code, thereby enabling a more efficient and maintainable architecture.
During the session, various examples of serverless refactoring were presented to illustrate how certain tasks and processes within applications can be optimized. For instance, replacing custom code with Lambda Destinations allows the automated routing of function results to other AWS services, reducing code complexity and maintenance requirements. Another example involved using AWS Step Functions for orchestrating workflows, which replaces the need for complex choreography between Lambda functions. These examples highlighted the shift from a focus on infrastructure to a focus on application and data flow, showcasing the benefits of serverless in simplifying application architecture.
The presentation also covered practical aspects of implementing serverless refactoring within organizations. The speakers discussed team structures and tooling, suggesting that continuous refactoring should be an integral part of software delivery, similar to continuous integration and deployment. They emphasized the importance of functional test coverage and automation in supporting frequent and painless refactoring. The session concluded with a call to action for attendees to actively engage in serverless refactoring, contributing to the community-driven catalog of refactoring patterns and benefiting from the streamlined, efficient, and scalable nature of serverless architectures.
AWS re:Invent 2023 - Refactoring to serverless (SVS305)
SVS307 Scaling serverless data processing with Amazon Kinesis and Apache Kafka
The presentation at AWS re:Invent 2023 focused on scaling serverless data processing using Amazon Kinesis and Apache Kafka. The speaker, Julian Wood, discussed various aspects of integrating Lambda with these streaming data services. He highlighted the different types of audience members, including those familiar with either Kinesis or Kafka, and those interested in understanding how Lambda could enhance their data streaming processes. The session aimed to bridge knowledge gaps between different user groups and provide insights into effective data processing.
Julian emphasized the importance of understanding streaming data concepts like brokers, partitions, and records. He delved into the specifics of Kafka and Kinesis, explaining their architecture, including clusters, brokers, topics, and partitions in Kafka, and streams and shards in Kinesis. The talk also covered the Lambda event source mapping, a critical component for processing streaming data. This mapping allows for the automatic scaling of data processing, managing consumer offsets, and ensuring efficient handling of streaming data.
The presentation also addressed practical strategies for managing performance and scaling in data streaming scenarios. Julian provided insights into optimizing Lambda functions for better performance, including effective partition key strategies, error handling, and monitoring using CloudWatch metrics. He explained the parallelization factor in Kinesis and the use of enhanced fan-out for improved throughput. Julian concluded by highlighting resources for continued learning in serverless architectures and encouraged attendees to explore further at the re:Invent Expo and ServerlessLand.com.
AWS re:Invent 2023 - Scaling serverless data processing with Amazon Kinesis and Apache Kafka(SVS307)
SVS309 Improve productivity by shifting more responsibility to developers
This transcript from AWS re:Invent 2023 details a presentation by Heeki Park, a specialist SA at AWS, and Sam Dengler, a senior distinguished engineer at Capital One. They discuss the importance of governance in application deployment, particularly focusing on AWS Lambda. Their talk highlights the necessity of balancing rapid development with compliance and safety, emphasizing the need to manage serverless applications effectively.
The presentation covers various tools and techniques to enhance governance and productivity. They introduce AWS services like AWS Config, CloudFormation Guard, and Amazon Inspector to ensure compliance and safety in serverless application deployment. The speakers illustrate the use of AWS Config for proactive and detective controls, CloudFormation Guard for template validation, and Amazon Inspector for runtime scanning of code for vulnerabilities. They also discuss the importance of code signing to ensure the integrity and origin of code, as well as strategies like centralized artifact repositories for managing dependencies securely.
Finally, the talk addresses the ongoing nature of governance, suggesting practices for addressing compliance findings, owner outreach, and documentation. It emphasizes incorporating governance activities into standard development workflows, such as agile remediation and using issue management systems. The presentation concludes with a reminder of the continuous evolution of governance controls and the necessity of adapting to changing requirements and standards in serverless computing.
AWS re:Invent 2023 - Improve productivity by shifting more responsibility to developers (SVS309)
SVS323 I didn’t know Amazon API Gateway did that
The presentation at AWS re:Invent 2023 by Eric Johnson, a Principal Developer Advocate for Serverless at AWS, focused on the lesser-known capabilities of Amazon API Gateway. Johnson began by sharing his initial challenges with API Gateway and how he overcame them by deeply learning and teaching about it. He emphasized his passion for serverless technologies and his extensive experience in software development and architecture.
Johnson delved into various features and functionalities of API Gateway, highlighting its versatility and power. He covered different aspects such as authorization methods (IAM permissions, Lambda authorizers), caching capabilities, logging and monitoring with CloudWatch, and the use of stages for managing different API versions. Johnson also discussed the use of canary releases for API updates, resource policies for security, and the integration of client-side SSL certificates for backend verification.
The presentation also covered the benefits of using OpenAPI for API development in API Gateway, the significance of identity-based policies, and the utility of mutual TLS for enhanced security. Johnson stressed the importance of proper API management and monitoring, using throttling and logging to maintain API health and efficiency. He concluded by pointing attendees to resources for further learning and encouraged them to explore Serverless Land for patterns and templates related to serverless architectures.
AWS re:Invent 2023 - I didn’t know Amazon API Gateway did that (SVS323)
SVS401 Best practices for serverless developers
At AWS re:Invent 2023, Julian Wood and Chris Muns presented key practices for serverless developers, emphasizing the strategic mindset and approach needed for building applications in the cloud. They focused on leveraging AWS's expertise in delivering well-architected outcomes like security, scalability, performance, and availability. This enables developers to concentrate on creating value for customers without the complexities of managing infrastructure. They discussed how serverless has evolved beyond just running code to encompass broader application constructs like databases, functions, queues, and workflows. They highlighted the importance of speed and quality in development, facilitated by serverless capabilities, which help in quick prototyping and rapid feedback loops.
The session also covered practical aspects like the use of frameworks for building serverless applications, testing strategies, and operational best practices. Tools like AWS SAM (Serverless Application Model), CDK (Cloud Development Kit), and Application Composer were recommended for efficient application development and deployment. The speakers emphasized the importance of testing throughout the development lifecycle, including local testing, integration testing in the cloud, and production testing. They also discussed operational practices in a serverless environment, focusing on setting up guardrails and employing platform engineering to manage and govern applications effectively.
Towards the end, Wood and Muns delved into AWS Lambda, discussing performance optimization, concurrency models, and cost-effective strategies. They stressed the importance of understanding the Lambda execution environment lifecycle, including handling cold starts and leveraging provisioned concurrency for better performance. By understanding these aspects, developers can efficiently scale their serverless applications and maintain high performance. The session concluded with resources for further learning and an invitation to continue exploring AWS services for serverless development.
AWS re:Invent 2023 - Best practices for serverless developers (SVS401)
Conclusion
These are summaries of all the 300 and 400 level Serverless sessions. We hope you found these helpful in both getting an overview of the new serverless content as well as deciding which sessions to go watch.
Brian Tarbox
