September 2022 Summaries

Kenny Chamberlin's blog post discusses the challenges and proposed solutions for handling exceptions in programming, particularly focusing on languages like Kotlin, Rust, and C++. He reflects on his experiences with exception handling in languages such as C# and Java, noting the difficulties in ensuring bugs are addressed, particularly in Java's approach to exceptions. Chamberlin advocates for a design philosophy where errors are treated as return values rather than exceptions, promoting compiler-enforced correctness and reducing the need for developers to handle exceptions manually. This approach, which he refers to as "Exceptions Are Bugs," emphasizes using the type system to model possible outcomes, encouraging developers to write robust and error-resistant code. The post highlights how various languages have incorporated features to support this paradigm, such as Rust's Result<> and Option<> and Kotlin's sealed types. Chamberlin suggests this method results in clearer, safer, and more cooperative development, despite potentially requiring more code and adaptation in languages that do not yet fully support pattern matching.

The blog post discusses the tuning and optimization of the Momento JavaScript cache client, focusing on enhancing performance, resilience, and debuggability to provide an excellent user experience. Emphasizing simplicity, the article details how the client libraries handle configuration challenges such as CPU utilization and network I/O, particularly in a Node.js environment that limits execution to a single CPU core. It explores the impact of gRPC settings, such as the grpc.use_local_subchannel_pool, on client performance, demonstrating significant improvements in latency and throughput by adjusting the number of concurrent requests and gRPC channels. The post highlights that optimal configurations can vary between development and production environments, and announces upcoming pre-built configurations in the Momento JavaScript client to cater to different scenarios. Concluding with a look ahead, it teases future insights into Python client tuning, reinforcing Momento's commitment to handling complex tuning for users.

The blog post discusses enhancing an AWS-based API service by implementing new endpoints for user avatar images and securing them with custom authentication logic using AWS API Gateway custom authorizers. By introducing a new /profile-pic and /cached-profile-pic endpoint, the service fetches and returns the base64 encoded user images, allowing for improved security and performance. The custom authorizer, a Lambda function, checks if the requester's user ID is a follower of the profile picture's owner, allowing or denying access accordingly. The post highlights the benefits of using a centralized cache, demonstrating that integrating caching with serverless architectures like AWS Lambda and API Gateway can significantly reduce latency and enhance performance. The author also evaluates performance through testing and demonstrates that enabling a remote cache for the custom authorizer and service results in lower response times, underscoring the importance of caching in distributed, function-based applications.

Momento, a company utilizing Twitter's open-source caching framework Pelikan, is discussed on the Open Source Startup Podcast by co-founder Daniela and Twitter Principal Engineer Yao Yue. The conversation highlights Momento's approach to providing a streamlined caching service, akin to an airline offering a complete travel experience, in contrast to just offering an engine like Memcached or Redis. This metaphor illustrates Momento's focus on delivering an end-to-end experience that eliminates the complexities of managing caching services. The podcast also delves into the challenges of messaging and positioning for technical companies and emphasizes the significance of fostering an active developer community.

Momento offers a straightforward and transparent pricing model for its cache service, aiming to simplify the often complex and unpredictable costs associated with cloud services. The pricing is based on a single dimension of $0.50 per gigabyte for both inbound and outbound data, with no additional storage charges, and a free tier allows developers to use the first 5GB each month without cost. This approach is designed to eliminate hidden fees and make it easy for customers to project costs and scale their applications without encountering unexpected expenses. By focusing on simplicity and elasticity, Momento seeks to optimize the total cost of ownership for cloud databases and enable faster development cycles for its users.