Cloud Architecture

Your landing zone set up the org, OUs, and baseline SCPs — then most teams stall, duplicating resources per account and wiring brittle cross-account role chains. Since re:Invent 2024 the toolkit changed: RCPs bound what can be done TO a resource (even by external principals), declarative policies enforce EC2/VPC/EBS config state that survives new APIs, and one Route 53 Profile can push DNS to up to 5,000 VPCs. Here is the mechanism-by-job decision matrix and a rollout order that avoids lockouts.

A composite mid-market manufacturer exited a colo of 150 workloads and ~400 TB in 21 months across 8 waves, with 47 minutes of unplanned cutover downtime (industry median per Gartner: ~4 hours per major cutover). The single-wave alternative bid came in at ~$4.8M; the wave-based program ran ~$2.1M. Here is how the wave plan, dependency cutoff, and rollback layer fit together against a hard landlord deadline.

July 2026: EU Data Act switching charges hit zero by 12 Jan 2027; AWS free DTO-on-exit still requires Support approval and typically a 90-day move window. Exit cost is re-platform time (~$113k in our model), not the ~$4,400 egress line.

A stack of ALB + EC2 + RDS Multi-AZ + S3 composes to ~99.83% availability—so promising customers 99.9% is a check you cannot cash. This guide does the composition math, converts it to an error budget (99.9% = 43.2 min/month), and shows why AWS service credits never fund your SLA penalties.

July 2026: Barclays ~86% plan some repatriation; IDC ~8–9% full exit — both true. Per-workload placement beats multi-cloud-by-default. Scored matrix + daily tax of “just in case” portability.

A CCoE that only publishes standards decks fails within two quarters. This 2026 operating model ties platform RFCs, delegated-admin guardrails, Well-Architected reviews, and FinOps chargeback—benchmarked on a 14-account estate that cut deploy exceptions from 23/month to 6 in 90 days.

July 2026: MAP Mobilize should not start with more than five failed Platform/Security controls. 47-point readiness checklist (People, Platform, Security, FinOps) for Assess gates — CAF 3.0 + Migration Lens, post–Migration Hub.

July 2026: opinionated 6-phase AWS learning path — 3-month full-time benchmark, skip Cloud Practitioner, ship projects before SAA, with phase exit criteria you can print.

CAF 3.0 organizes six perspectives and 47 capabilities—up from 31 in CAF 2.0—plus four phases (Envision, Align, Launch, Scale). Here is how to connect those workshops to Control Tower, MAP, and Well-Architected without treating the framework as a slide deck.

Global Accelerator charges about $0.025 per provisioned accelerator per hour—even while disabled—and adds Data Transfer-Premium on top of normal data transfer. Two static Anycast IPv4 addresses (or four addresses in dual-stack: two IPv4 and two IPv6) front ALBs, NLBs, EC2, or EIPs across Regions; that pricing model changes whether you beat CloudFront or Route 53 latency records alone.

May 2026 guidance with a reproducible scaffold—9 artefact files, 0 Terraform-managed creates at plan baseline, anchored to official OpenTofu, HashiCorp, and Ansible project documentation.

July 2026 field note: Aurora failover windows still force app timeouts above tens of seconds; DynamoDB hot partitions and short transactions remain the scaling levers — with a partition-math worksheet you can reproduce.