Designing And Building Datacenters - Tapa blanda

Sinopsis

Designing and Building Datacenters offers a complete, end-to-end view of the engineering that powers the modern internet and AI workloads. Readers will walk through every critical decision-from choosing a site with the right power, fiber, and water resources to laying the concrete foundation, installing racks, and configuring the complex power and cooling chains that keep a facility running 24/7. By the end of the book you will understand how to balance redundancy, efficiency, and cost while meeting the stringent reliability demands of today's digital services. The guide dives deep into the core infrastructure that makes a datacenter function. You'll learn how utility feeds are transformed into clean, uninterruptible power through UPS systems, generators, and dual-path distribution, and how heat is removed using everything from traditional air containment and hot/cold aisle strategies to cutting-edge direct-to-chip and immersion liquid cooling. Chapters on building architecture, whitespace and gray space planning, rack selection, and structured cabling show you how to create a physical layout that optimizes airflow, power delivery, and maintainability while supporting future growth. Networking receives equal attention, with clear explanations of leaf-spine fabrics, ECMP routing, BGP in the data center, and the optical transport technologies-DWDM, coherent transceivers, and CPO-that move terabits of data across continents. You'll also explore compute platforms ranging from general-purpose x86 servers to AI-focused GPU machines, storage systems built on erasure coding and software-defined storage, and the software layers of virtualization, containers, and Kubernetes that turn hardware into a flexible, self-healing resource pool. Automation, provisioning, and Infrastructure as Code are detailed so you can treat the entire facility as programmable infrastructure. Reliability, observability, and security are treated as ongoing disciplines rather than afterthoughts. You'll discover how to measure availability with MTBF/MTTR, apply the Uptime Institute tiering model, and implement predictive maintenance and change-management processes that prevent human-induced outages. Monitoring, telemetry, DCIM, and observability tools are explained to give you real-time insight into power, cooling, and network health, while zero-trust architecture, microsegmentation, and encryption strategies protect data and applications from ever-evolving cyber threats. Sustainability metrics like PUE, WUE, and CUE are covered alongside renewable energy procurement, waste-heat reuse, and circular-economy hardware practices. Finally, the book looks ahead to the forces reshaping datacenters: edge and modular deployments that bring compute closer to users, AI-ready facilities designed for extreme power densities, widespread liquid cooling, and the push toward 800GbE and co-packaged optics. You'll gain insight into disaggregated infrastructure enabled by CXL, the integration of datacenters into district heating systems, and the evolving role of the facility as a hub for both centralized cloud and distributed edge workloads. Armed with this knowledge, you'll be equipped to design, build, operate, and future-proof datacenters that meet the performance, efficiency, and resilience demands of the next generation of digital services.

"Sinopsis" puede pertenecer a otra edición de este libro.