InfoQ Homepage Performance & Scalability Content on InfoQ
-
Producing the World's Cheapest Tokens: A How-to Guide
Meryem Arik explains how to cut AI inference costs by up to 90% by optimizing batch sizes, hardware selection, and request scheduling for high-volume, non-real-time workloads.
/presentations/ai-token-price/en/mediumimage/meryem-arik-medium-1785844873071.jpg)
-
Keeping ChatGPT Fast as AI Development Accelerates
Martin Spier shares how agentic coding accelerates release velocity at OpenAI and explains how always-on AI agents redefine performance engineering to keep ChatGPT fast at massive scale.
/presentations/openai-performance-engineering-agentic-coding/en/mediumimage/MartinSpier-medium-1785314358072.jpeg)
-
From ms to µs: OSS Valkey Architecture Patterns for Modern AI
Dumanshu Goyal explains how moving from proxied architectures to direct access in Redis/Valkey slashes tail latency to microseconds, lowers costs, and eliminates single points of failure.
/presentations/valkey-architecture-patterns/en/mediumimage/DumanshuGoyal-medium-1786082640468.jpg)
-
Million PDFs: Building a Modern Document Infrastructure with Rust and Typst
Erik Steiger explains how to replace heavy, outdated PDF engines like Crystal Reports and Puppeteer with a high-performance, serverless Rust and Typst pipeline, dropping render times below 2ms.
/presentations/document-infrastructure-rust-typst/en/mediumimage/ErikSteiger-medium-1782220478687.jpg)
-
Architecting a Centralized Platform for Data Deletion at Netflix
Netflix Engineers Vidhya Arvind and Shawn Liu discuss the pillars of safe, large-scale data deletion. They explain strategies to eliminate data ghosts and manage tombstone resource contention.
/presentations/architecting-deletion-system/en/mediumimage/medium-1779869686290.jpg)
-
How to Build an Exchange: Sub Millisecond Response Times and 24/7 Uptimes in the Cloud
Frank Yu explains how Coinbase builds ultra-high-performance exchanges. He discusses using single-threaded, deterministic core logic and Raft consensus to achieve sub-millisecond P99s in the cloud.
/presentations/exchange-systems-cloud/en/mediumimage/frank-yu-medium-1776173818222.jpeg)
-
Scaling Cloud and Distributed Applications: Lessons and Strategies from chase.com, #1 Banking Portal in the US
Durai Arasan shares how Chase.com achieved a 71% latency reduction. He explains strategies for efficient scaling, multi-region resilience, and automated "repaving" to secure large-scale systems.
/presentations/chase-scale-cloud-distributed/en/mediumimage/durai-arasan-medium-1765444482482.jpg)
-
Architecting Planet Scale, Modern Apps in the Cloud
George Mao discusses the five stages of maturity for building a planet-scale, global, and highly available architecture using fully serverless, cloud-native services (GCP/AWS).
/presentations/maturity-architecture-cloud/en/mediumimage/george-mao-medium-1764845691312.jpeg)
-
Scaling an Embedded Database for the Cloud – Challenges and Trade-Offs
Stephanie Wang discusses the tradeoffs and lessons learned from building a cloud-native data warehouse by scaling an embedded database, going from an in-process system to one with cloud capabilities.
/presentations/scale-embedded-database/en/mediumimage/stephanie-wang-medium-1757423720784.jpg)
-
GenAI at Scale: What it Enables, What it Costs, and How to Reduce the Pain
Mark Kurtz discusses scaling GenAI and optimizing LLM deployments. He shares how to overcome technical and financial challenges using open-source tools like vLLM, LLM Compressor, and InstructLab.
/presentations/genai-scale/en/mediumimage/mark-kurtz-medium-1756207670091.jpeg)
-
Architecture in the Lead: Scaling Today, Shaping Tomorrow
Ian Arundale and Matthew Clark share how the BBC's architecture delivers for huge live events, discussing the importance of elasticity, resilience, and security, and the crucial human skills.
/presentations/architecture-scale-change/en/mediumimage/matthew-clark-ian-arundale-medium-1756282334948.jpg)
-
Beyond Durability: Database Resilience and Entropy Reduction with Write-Ahead Logging at Netflix
Prudhviraj Karumanchi and Vidhya Arvind share how Netflix built a Write-Ahead Log to guarantee data durability and reliability, tackling issues like data loss, corruption, and replication at scale.
/presentations/netflix-write-ahead-logging/en/mediumimage/medium-2-1750404618074.jpg)