Commitment Discount
A discount is not the same as a saving. Check what usage qualifies, which payments remain due, and how changes affect coverage and utilization.
-
-
Serverless Containers
Cloud Run, ECS on Fargate, and EKS on Fargate expose different operating models. Check workload fit and billing rather than assuming idle is free.
-
Configuration Drift
Compare declared, built, and running state. Correct the responsible change path while distinguishing permitted variation from unobserved properties.
-
Golden Image
A tested machine image supplies a shared starting artifact. Replacement still needs state handling, launch controls, and service checks.
-
Support Lifecycle
Check the fixes and assistance your exact release and components receive, then plan the next transition before that coverage ends.
-
Cloud Bursting
Bursting adds capacity for demand peaks. It may still depend on the primary environment, so it does not by itself demonstrate disaster recovery.
-
Data Gravity
Data placement affects transfer and processing costs. Compare local execution, replicas, and remote reads using measured volume, lag, and workload needs.
-
Multicloud
Multiple clouds add operating responsibilities. Assess their purpose, identity lifecycle, observability, and the cost of moving significant workloads.
-
Hybrid Cloud
A connected datacenter and cloud do not by themselves establish portability. Specify the deployment model and test the capabilities required between them.
-
Shared VPC
Shared VPC separates network and workload administration. A host can share multiple VPCs, while each network retains its own scope and limits.
-
Resource Hierarchy
Resource position affects inherited controls, but billing, quota, and isolation scopes differ. Review each relationship before reorganizing the hierarchy.
-
Cost Allocation Tag
Tags, account ownership, and usage records answer different allocation questions. Measure missing coverage and reconcile the result against billing.
-
Service Control Policy
SCPs limit covered AWS principals without granting access. Resource-policy grants and principal identity determine which restrictions apply.
-
Landing Zone
A landing zone is a deployed, governed cloud environment for workloads. Supplying new ones repeatably takes a request and handover process; without it, waiting can push teams toward accounts of their own.
-
Data Transfer Cost
Map the actual flows and price each path. Assess replication, batching, and placement against correctness, latency, recovery, and availability requirements.
-
Private Endpoint
A working private path does not prove public access is closed. Test the application’s actual destination and the service’s access policies.
-
Virtual Network
Plan address capacity and non-overlap. Adding a supported range differs from changing a subnet or renumbering deployed resources.
-
Cold Start
Measure initialization code, runtime startup, and dependency delays against the workload deadline before choosing a startup mitigation.
-
Shared Responsibility Model
Managed services still leave customer duties for data, settings, identities, and access decisions. Identify the service-specific duties and who operates each control.
-
Dynamic Resource Allocation
Returning executors can reduce contention. Billing savings depend on node scaling and pricing; shuffle files and cached blocks affect when executors can leave.
-
Synapse SQL Pools
A dedicated SQL pool reserves warehouse compute you pay for while it runs; a serverless pool charges for the data each query processes. They fit different jobs.
-
OneLake
Every Fabric tenant has exactly one OneLake. Workloads store and read Delta or Iceberg tables in it, so one copy of data can serve Spark, SQL, and Power BI.
-
OneLake Shortcut
A shortcut makes an S3 bucket or another workspace’s table look like a local folder in Fabric. No separate copy is maintained, though optional caching can store files locally, and deletes and access can reach the source.
-
Capacity Unit (Microsoft Fabric)
An F64 capacity provides 64 capacity units, shared by every workspace assigned to it. The number sets compute, licensing rules, and who slows down whom.
-
Capacity Bursting and Smoothing (Microsoft Fabric)
A spike can run faster than the capacity you bought, because Fabric spreads the extra usage over the next minutes or hours. The borrowed compute still counts.
-
Cloud Infrastructure
Cloud infrastructure is the hardware and software that make computing self-service, pooled, elastic, and metered — and it sets what you manage versus rent.
-
Archive Rehydration
Plan how offline archived objects become readable, track progress and temporary copies, and validate content and consumer access before declaring recovery.
-
Minimum Storage Duration
Separate minimum billable residence from retention requirements, use the correct tier-entry clock, and include early-deletion charges in a comparison.
-
Storage Lifecycle Policy
Define storage transitions and expiration with explicit filters and clocks, protect referenced versions, and verify the objects a policy actually affects.
-
SaaS (Software as a Service)
SaaS provides access to a provider’s cloud applications. Understand the settings consumers control and their remaining integration, data, and access responsibilities.
-
Compute
Learn how compute executes data and model workloads, and why more processors do not fix every delay.
-
Storage
Understand storage as a platform responsibility, including persistence, versions, access, and recovery.