Libri GmbH Architecture réseau résiliente : Conception redondance, basculement, récupération et connectivité tolérante aux pannes

Two routers. Two circuits. Two cloud regions. And still-one outage.Redundancy looks reassuring on an architecture diagram. Production failure is different. A backup circuit may share the same conduit as the primary. Two firewalls may receive the same faulty configuration. Multiple cloud regions may depend on one global identity or policy service. A failover path may work perfectly-until it receives more traffic than it can carry.Resilient Network Architecture shows network engineers, architects, cloud professionals, infrastructure teams, SREs, security practitioners, and technical leaders how to design connectivity that continues delivering useful service when real-world assumptions break.Rather than treating resilience as a shopping list of redundant hardware, Rion Frost builds an end-to-end framework around failure domains, shared fate, blast radius, critical flows, graceful degradation, surviving capacity, recovery dependencies, and testable evidence.Learn how to evaluate active-active and active-passive architectures, N+1 capacity, path diversity, first-hop redundancy, multi-chassis designs, routed access, BFD, OSPF, IS-IS, BGP, RPKI, ECMP, fast reroute, anycast, SD-WAN, DNS traffic management, leaf-spine fabrics, VXLAN/EVPN, and multi-region cloud connectivity.The book goes beyond Layer 3. You will see how load balancers, health checks, timeouts, retries, circuit breakers, queues, session state, identity systems, PKI, security policy, observability, automation, and application behavior can determine whether a network failure remains minor-or becomes a cascading incident.Dedicated chapters examine hybrid and multi-cloud design, security without availability fragility, network observability, disaster readiness, RTO and RPO, infrastructure as code, chaos engineering, game days, digital twins, intent-based networking, and guarded automation.Real incident analysis and practical scenarios show why apparently distributed systems can still contain hidden global dependencies-and why recovery capability must be proven rather than assumed.An extensive Resilience Architecture Review Toolkit helps you map critical flows, identify shared failure domains, define degraded-service contracts, calculate capacity after failure, control change blast radius, demand recovery evidence, design game days, and build measurable improvement cycles.Stop designing networks that merely have backups. Build systems that remain understandable, survivable, and recoverable when the diagram is no longer green.