The Hitchhiker’s guide to how stuff works

Inspired by Douglas Adams and grounded in 50 years of computing history, this podcast explores HPE NonStop (formerly Tandem) - the system built on the assumption that everything will break. From process pairs and fail-fast architecture to real-time card swipes powering global payment switches, discover how a "hive" of modest servers quietly ran the world's money for half a century.

Traditional network perimeters are dead. Welcome to the era of Zero Trust Network Architecture (ZTNA), where every device, connection, and packet is continuously scrutinized .

We explore how modern software-defined networks decouple the control plane from the data plane, leveraging service meshes like Envoy for mutual TLS (mTLS) while routing traffic securely across microservices and global Content Delivery Networks.

This comprehensive overview explores the evolution of artificial intelligence, tracing its journey from early neural networks to massively scaled Large Language Models and reasoning engines. It examines the foundational impact of the Transformer architecture and self-attention mechanisms, which unlocked massive parallelization and advanced capabilities. Furthermore, the text details the critical role of specialized hardware like NVIDIA GPUs, infrastructure challenges in production such as the KV Cache, and the importance of data curation and Retrieval-Augmented Generation. Finally, it highlights the ongoing race toward Artificial General Intelligence, open-source breakthroughs, and the significant energy demands required to power modern AI datacenters.

Netflix manages its massive global streaming platform by dividing its infrastructure into two main systems: a control plane hosted on cloud servers that handles user browsing and account management, and a custom data plane that efficiently delivers video content locally through internet service providers. To prevent widespread outages, the company transitioned from a fragile monolithic application to thousands of autonomous microservices, utilizing dynamic service discovery and automated circuit breakers to isolate failures. Furthermore, the platform employs chaos engineering by intentionally breaking its own components during business hours to ensure system resilience, while massive data pipelines and machine learning models continually analyze user activity to personalize recommendations and optimize streaming quality.

In this episode, we tackle the massive, ever-expanding universe of Amazon Web Services (AWS). Think of AWS as an infinite, pay-by-the-hour megacity where you can rent anything from a single storage locker (S3) to an entire skyscraper of supercomputers (EC2). We’ll break down how the cloud actually works, why companies ditched their own server rooms, and revisit the infamous days when the 'us-east-1' region took a nap and brought half the internet down with it. Grab your towel and learn how to navigate the cloud without accidentally racking up a $10,000 billing surprise overnight.