The path· 6
- 1Intermediate 7 minNew York suspends new data centers above 50 MWGovernor Kathy Hochul signed an order pausing new 50 MW+ data centers for up to a year while the state runs an environmental study. It targets AI's pressure on the grid, water use and cost-shifting to the public — and signals a new regulatory risk for anyone designing cloud and AI workloads.
- 2Advanced 9 minECS Auto Scaling: High-Resolution Metrics vs. Traditional Scaling ApproachesThe launch of high-resolution (20-second) metrics for Amazon ECS reshapes the auto scaling calculus for latency-sensitive financial workloads. This article compares the four primary scaling approaches — target tracking with high resolution, step scaling, scheduled scaling, and predictive — with concrete trade-offs, real numbers, and a decision matrix for production environments.
- 3Advanced 10 minAWS Lambda MicroVMs: technical review of a new serverless primitiveAWS Lambda MicroVMs fills a genuine gap between ephemeral functions and heavy VMs, delivering hypervisor-level isolation with near-instant resume latency and state preserved for up to 8 hours. As an architect operating financial-grade multi-tenant environments, I see genuine potential here — and equally real pitfalls that must be addressed before any production adoption.
- 4Expert 9 minOne Org or Many? The Postmortem Nobody Wants to WriteThe choice between a single AWS Organization and multiple organizations looks strategic on paper but becomes operational at the worst possible moment — during an incident. This article reconstructs the most common failure pattern I observe in financial-grade environments and proposes concrete remediations grounded in the AWS Well-Architected pillars.
- 5Expert 10 minOracle HA on AWS: FSx for ONTAP as a Lever for Gradual ModernizationThe most relevant signal in this cycle for architects still carrying critical Oracle workloads is not to abandon them — it is to orchestrate them with cloud-native primitives that deliver real HA without forcing a rewrite. FSx for NetApp ONTAP changes the storage equation for Oracle on AWS in concrete ways. This briefing analyzes what changes in practice, the real trade-offs, and how to position this layer within a gradual modernization strategy.
- 6Expert 9 minEC2 G7 & NVIDIA Blackwell: GPU Inference Architecture for ProductionEC2 G7 instances, accelerated by NVIDIA RTX PRO 4500 Blackwell Server Edition GPUs, represent a generational leap that goes well beyond benchmark numbers. In this analysis, I examine the real GPU-to-GPU communication mechanisms, failure patterns in multi-node clusters, and the architectural decisions that separate a functional inference deployment from a financial-grade fault-tolerant system.
Deep-dive studies
guideLambda vs ECS: the architect's compute decision guide on AWSChoosing between Lambda and ECS is not about preference — it is about matching the unit of scale to the load pattern. This guide covers every Lambda type (including MicroVMs and Managed Instances) and ECS type (Fargate, EC2, Managed Instances), the decision framework I use in practice, and the real impact on engineering, business, and customer experience.playbookPlaybook: 5 AI Architecture Shifts — and what to do about each oneWhat was advanced in 2025 became default in 2026: the bottleneck moved past the prompt, agents went to managed production, MCP standardized tooling, security became layered, and token FinOps entered the roadmap. This playbook maps each shift to concrete actions — for people building systems, not demos.guideInside AI Agents (3/3): Agents in Production on AWS with Bedrock AgentCoreThe third and final part of the series descends the elevator to the technical floor: how to run an AI agent in production on AWS using Amazon Bedrock AgentCore. We cover the full component map (Runtime, Gateway, Memory, Identity, Observability), model selection by cost/latency/reasoning, security with guardrails and session isolation, and FinOps to keep real costs under a few dollars per month.adrADR: Running AI-Generated Code — Lambda MicroVMs vs. Isolated ContainersArchitecture decision on how to safely execute AI-agent-generated or user-submitted code in a multi-tenant environment, comparing Lambda MicroVMs (AWS launch June 2026), ephemeral containers with gVisor/Firecracker on ECS/EKS, and standard Lambda. The recommended decision is per-session Lambda MicroVMs orchestrated by Bedrock AgentCore, providing VM-level isolation without the operational overhead of managing your own virtualization stack.adrADR: Nextflow Profiles on HealthOmics — Separating Config from LogicAWS HealthOmics now supports Nextflow profiles, enabling predefined execution settings to be activated at runtime without modifying workflow source code. This analysis examines the architectural decision behind that separation of concerns, its real trade-offs, and the operational consequences for teams running bioinformatics pipelines at scale.postmortemAWS us-east-1 (2025): Datacenter Thermal Event Takes Down EC2 and EBSA cooling system failure in an us-east-1 datacenter caused servers to overheat, triggering thermal protection shutdowns that degraded EC2 and EBS across the region. The incident reignites critical debates about workload concentration in us-east-1, the role of physical AZ isolation, and the need for graceful degradation in production architectures. Open source to explore
queue-advisor-pricing-appCompare SQS, Kinesis, EventBridge and MSK costs before committing to an AWS queue.aws-event-driven-finops-platformEvent-driven AWS banking reference architecture with FinOps, security, and a live frontend.aws-agentic-ai-reference-architectureAWS reference architecture for production agentic AI — security, observability, and DevSecOps.DeskbuddyESP32 touchscreen smart desk dashboard — firmware, web UI, and browser installer in one repo.solution-architecture-mcp-toolkitBilingual MCP toolkit for ADRs, threat modeling, and governed Well-Architected reviews.mcp-aws-solution-architectMCP server that turns any AI assistant into a copilot for AWS Solution Architects.