InfoQ Homepage Optimization Content on InfoQ
-
Inside Uber’s Query Architecture: Simplifying Layers and Improving Observability
Uber rebuilt its Apache Pinot query architecture, replacing the Presto-based Neutrino system with a lightweight proxy called Cellar and Pinot’s Multi-Stage Engine Lite Mode. The redesign simplifies SQL execution, improves resource management, and ensures predictable performance for large-scale analytics workloads.
-
Meta Open Sources OpenZL: a Universal Compression Framework for Structured Data
Meta’s OpenZL changes the way data is compressed by maximizing efficiency for structured datasets, outperforming traditional methods like Zstandard. With a universal decompressor and custom compression plans, it simplifies operational deployment while achieving superior compression ratios and speeds, making it an essential tool for modern data infrastructures.
-
Cloudflare Achieves 99.99% Warm Start Rate for Workers with 'Shard and Conquer' Consistent Hashing
Cloudflare's innovative "Shard and Conquer" technique revolutionizes its serverless platform by slashing cold start rates by 90%. Utilizing a consistent hash ring, it routes traffic efficiently, keeping Workers warm and minimizing latency. Enhanced for larger applications, this approach ensures optimal performance while accommodating user demands for richer functionalities.
-
Agoda Leverages ChatGPT in the CI/CD Process for SQL Stored Procedure Optimization
Agoda started utilizing ChatGPT to optimize SQL stored procedures (SP) as part of their CI/CD process. After introducing the automated LLM-assisted step, the company observed shortened stored procedure optimization times, which lightened the load on DB developers. Agora works on making ChatGPT more accessible for SP optimization outside of the CI/CD pipeline.
-
Challenges of Creating iOS App Extensions at Lyft
In a recent article, Lyft engineers Artur Stepaniuk and Max Husar described how Lyft handles the complexity of creating an app extension for their iOS app without breaking the tight RAM and binary size constraints set by Apple nor impair user experience.
-
Netflix Rolls Out Service-Level Prioritized Load Shedding to Improve Resiliency
Netflix extended its prioritized load-shedding implementation to the individual service level to further improve system resilience. The approach uses cloud capacity more efficiently by shedding low-priority requests only when necessary instead of maintaining separate clusters for failure isolation.
-
Compiler Explorer Provides Insights into Low-Level Android App Optimization
Android engineers at Google added support for the Java and Kotlin programming languages to Compiler Explorer, an open source tool aimed at exploring how compilers work by compiling code in real-time. Using Compiler Explorer, Android engineers can optimize the performance of their apps by observing how the compiler works under the hood instead of using a set of pre-defined best practices.
-
Azure Advisor Well-Architected Assessment in Public Preview to Optimize Cloud Infrastructure
Microsoft Azure recently announced the public preview of the Advisor Well-Architected assessment. This self-guided questionnaire aims to provide tailored, actionable recommendations to optimize Azure resources while aligning with the Azure Well-Architected Framework (WAF) principles.
-
How Amazon Aurora Serverless Manages Resources and Scaling for Fleets of 10K+ Instances
AWS engineers published a paper describing the evolution and latest design of resource management and scaling for the Amazon Aurora Serverless platform. Aurora Serverless uses a combination of components at different levels to create a holistic approach for dynamically scaling and adjusting resources to satisfy the needs of customer workloads.
-
Project Leyden Announces Early Access Build: 2-3x Start-up Improvements for Java Applications
The OpenJDK has reached a milestone by announcing the Early Access (EA) build for Project Leyden. This build represents over a year of development efforts to enhance Java application performance, particularly focusing on start-up times. The preliminary testing has shown impressive results, with popular application frameworks experiencing a 2-3x improvement in start-up times.
-
InfoQ Dev Summit Boston: Optimizing Java Applications on Kubernetes - Beyond the Basics
At the InfoQ Dev Summit in Boston, Bruno Borges, who has been principal PM manager at Microsoft for over six years, shared insights on optimizing Java applications on Kubernetes. His session focused primarily on leveraging JVM ergonomics, understanding the impact of CPU throttling, and effectively managing garbage collection processes.
-
Zendesk Moves from DynamoDB to MySQL and S3 to Save over 80% in Costs
Zendesk reduced its data storage costs by over 80% by migrating from DynamoDB to a tiered storage solution using MySQL and S3. The company considered different storage technologies and decided to combine the relational database and the object store to strike a balance between querybility and scalability while keeping the costs down.
-
LinkedIn Migrates Espresso to HTTP2 and Reduces Connections by 88% and Latency by 75%
LinkedIn was able to dramatically improve the scalability and performance of its Espresso database by migrating it from HTTP1.1 to HTTP2, resulting in a reduction in the number of connections, latency, and garbage collection times. To achieve these gains, the team had to optimize the Netty’s default HTTP2 stack to make it fit their needs.
-
How Spotify Fights Global Warming by Taking Their Mobile App Size under Control
Adding new features has often a hidden impact on app size, with several undesired effects, including user disengagement and unnecessary carbon emissions. To keep app size under control, Spotify engineers established pre-PR and post-PR processes to help prevent an estimated 10MB app size growth over six months.
-
Hugging Face's Guide to Optimizing LLMs in Production
When it comes to deploying Large Language Models (LLMs) in production, the two major challenges originate from the huge amount of parameters they require and the necessity of handling very long input sequences to represent contextual information. Hugging Face has documented a list of techniques to tackle those hurdles based on their experience serving such models.