The three cloud service models. IaaS gives you raw infrastructure (VMs, networks). PaaS gives you a managed platform (you bring code). SaaS gives you finished software (you just use it).

A Region is a geographic location; an Availability Zone (AZ) is an isolated datacenter within it. Spread workloads across AZs for fault tolerance and across regions for disaster recovery.

Recovery Time Objective = how fast you must recover after an outage. Recovery Point Objective = how much data (in time) you can afford to lose. Together they drive your DR strategy.

Defining infrastructure in version-controlled files instead of clicking in a console. Tools: CloudFormation, Terraform, Bicep, CDK, Pulumi. Enables reviews, repeatability, and audit.

The cloud provider secures the cloud (hardware, hypervisor); you secure what you put in the cloud (data, access, configuration). Misconfiguration — not provider breaches — causes most incidents.

Vertical = bigger machine (more CPU/RAM). Horizontal = more machines behind a load balancer. Cloud-native designs prefer horizontal scaling of stateless services.

Elastic Compute Cloud — resizable virtual machines. Choose instance families (t/m/c/r), AMIs, and pay per second. The foundation of IaaS on AWS.

Serverless functions. Event-driven, scales to zero, billed per invocation + duration. Max 15-minute runtime; mind cold starts.

Simple Storage Service — object storage with 11 nines of durability. Storage classes (Standard, IA, Glacier, Intelligent-Tiering) trade cost for retrieval speed.

Azure object storage within a Storage Account. Access tiers: Hot, Cool, Archive. Redundancy: LRS, ZRS, GRS, GZRS.

Virtual Private Cloud — your isolated network. Contains subnets (public/private), route tables, Internet Gateway, and NAT Gateway.

Identity and Access Management — defines who (users/roles) can do what (policies) to which resources. Follow least privilege; never use the root account day-to-day.

Azure's cloud identity provider (formerly Azure AD). Combined with RBAC and Conditional Access for authentication and authorization.

Managed relational databases (MySQL, Postgres, etc.). Multi-AZ for HA (synchronous standby), read replicas for read scaling (asynchronous).

Managed NoSQL key-value/document DB with single-digit-millisecond latency. Design around access patterns; choose partition keys to avoid hot partitions.

Globally distributed, multi-model NoSQL database with five tunable consistency levels and single-digit-ms latency.

AWS-native IaC using YAML/JSON templates organized into stacks. Supports parameters, outputs, and drift detection.

A clean domain-specific language that compiles to ARM templates. Easier to read and write than raw ARM JSON.

A logical container for related Azure resources sharing a lifecycle. Sits under a Subscription, which sits under a Management Group.

The fundamental unit of isolation in GCP — groups resources, IAM, billing and quotas. Organized under Folders and an Organization.

Serverless, petabyte-scale data warehouse. Pay for storage + bytes scanned. Use partitioning and clustering to control cost; train models with BigQuery ML.

Google Kubernetes Engine. Standard mode (you manage nodes) or Autopilot (Google manages nodes; you just deploy workloads).

Run any stateless container that listens on a port. Fully managed, scales to zero, request-based billing. Built on Knative.

Global, scalable messaging for event-driven systems and streaming ingestion. Pairs with Dataflow for processing.

A tradable certificate representing the right to emit one tonne of CO2-equivalent. Verified by registries like Verra and Gold Standard.

GHG Protocol emission categories. Scope 1 = direct (owned sources). Scope 2 = purchased energy. Scope 3 = value chain (usually the largest and hardest to measure).

Carbon Neutral typically offsets emissions. Net-Zero means deeply cutting emissions (often via SBTi-validated targets) and only offsetting the unavoidable residual.

Normalized Difference Vegetation Index — a satellite-derived measure of vegetation health. Used to estimate biomass and carbon sequestration in farming projects.

The 6th pillar of the AWS Well-Architected Framework: minimize the environmental impact of cloud workloads through region choice, right-sizing, and efficient design.

GCP service that reports gross carbon emissions per project, which can be exported to BigQuery and visualized for decarbonization tracking.