Hybrid cloud done smart, not "all-in-cloud" with bills that explode
Not everything that runs on-prem today makes sense in the cloud. Nor the other way around. I help you decide rationally which workloads are worth migrating to Google Cloud or Azure, what stays on-prem for cost/latency/compliance, and how to tie them together into a coherent architecture. Certified across 6 Google Cloud courses.
01 When cloud makes sense
Cloud isn't a destination, it's a tool. Here's when it's the right choice — and when it isn't:
It makes sense to migrate to cloud when:
- Bursty workloads (campaigns, end-of-month reports, batch jobs) — you pay only for what you use
- Secondary Disaster Recovery site — built cheaply with pull-on-demand replication
- Managed services that reduce TCO: Cloud SQL, BigQuery, Pub/Sub, managed auth
- New greenfield applications — cloud-native from the start, not migrated later
- Geographically distributed teams — global access, CDN, load balancing
It does NOT make sense to migrate (or stay all-cloud) when:
- Steady, predictable, 24/7 workloads — hardware amortized over 5 years beats cloud on cost
- Data processing with large constant volumes (egress bills destroy the economics)
- Strict compliance with RO/EU data residency without suitable regions
- Ops team without cloud skills and no training budget
- Legacy applications that can't be refactored cost-effectively
The truth rarely told
Many "all-in-cloud" projects ended up costing 2-3x more than the equivalent on-prem version, especially after years 2-3 when cloud egress bills and managed services become visible. A properly designed hybrid keeps the best of both worlds.
Many "all-in-cloud" projects ended up costing 2-3x more than the equivalent on-prem version, especially after years 2-3 when cloud egress bills and managed services become visible. A properly designed hybrid keeps the best of both worlds.
02 What I deliver
Cloud readiness assessment
Workload inventory, cloud-fit scoring per application (6R: Rehost/Replatform/Refactor/Retire/Retain/Repurchase), 3-year TCO projection.
Hybrid architecture
End-to-end design: VPN/Interconnect connectivity, federated IAM, unified monitoring, cross-environment backup, identity & access governance.
Lift-and-shift / replatform migration
Controlled execution: VM migration (Migrate for Compute Engine, Azure Migrate), replatforming on managed services when cost decreases.
FinOps & cost optimization
Actual billing audit, waste identification (idle resources, over-provisioning), committed use discounts, reserved instances, continuous rightsizing.
Docker / GKE containerization
Packaging applications in Docker, deployment on managed Kubernetes (GKE, AKS), CI/CD with Cloud Build or GitHub Actions.
Governance & guardrails
Organization policies, standard tagging/labeling, budget alerts, least-privilege IAM, audit logging, landing zone template.
03 Types of hybrid architectures I implement
On-prem VMware/Proxmox
+
Google Cloud
Hybrid DR + burst
Production stays on-prem (stable cost, zero latency), DR site in Google Cloud (async replication, on-demand activation), plus capacity for bursty workloads (reports, analytics, batch).
On-prem Active Directory
+
Azure AD / Entra ID
Hybrid identity
AD Connect for synchronization, SSO to SaaS applications (Microsoft 365, Google Workspace, cloud ERP), centralized MFA, conditional access policies.
On-prem databases
→
Cloud SQL / BigQuery
Data platform
I keep OLTP close to the application (latency), but extract analytics to BigQuery / Snowflake where scale is cheap. ETL pipeline with Pub/Sub + Dataflow or equivalent.
04 The cloud migration process
PHASE 01
Readiness Assessment
Inventory, 6R scoring per workload, TCO projection, quick wins identification.
PHASE 02
Landing Zone
Setup organization, projects, VPC, IAM, logging, monitoring — the foundation we build on.
PHASE 03
Pilot Migration
2-3 non-critical applications migrated, lessons learned, process adjustments.
PHASE 04
Migration in waves
Planned waves, rollbacks prepared, stakeholder communication, intense monitoring per wave.
PHASE 05
Optimize
FinOps review, rightsizing, committed use discounts, replatforming to managed services where costs drop.
PHASE 06
Operate
Runbooks, alerting, dashboards, ops team training, optional retainer for BAU.
05 Use cases
Manufacturer with seasonal peak
24/7 on-prem production, burst capacity in GCP for end-of-quarter reports. Cost 40% lower than permanent on-prem expansion.
Scaling SaaS provider
Progressive migration from individual VPS to GKE, full CI/CD, cost-per-customer reduced by 35%.
Company with small-budget DR
Primary site on-prem, DR with Veeam Cloud Connect to GCP. RPO 4h, RTO 2h, monthly cost €400 vs. €4,000 for a secondary datacenter.
Company fully migrated to cloud, wants hybrid
AWS cost reached $25k/month. FinOps analysis, reshift predictable workloads back on-prem, 55% reduction without service loss.
06 Packages
Package 01
Cloud Readiness Assessment
from €3,000 / project
- Workload inventory
- 6R scoring per application
- 3-year TCO projection
- Phased recommendations
- Duration: 3-4 weeks
Package 02 · Recommended
Hybrid Architecture Design
from €7,500 / project
- Everything in Package 01
- Target hybrid architecture
- Landing zone template
- Terraform IaC for foundations
- Pilot of 2-3 migrated applications
- Duration: 6-10 weeks
Package 03
FinOps Review
from €2,000 / project
- Full cloud billing audit
- Waste + quick wins identification
- Rightsizing recommendations
- Committed use / reserved strategy
- Target 15-30% bill reduction
- Duration: 2 weeks
07 FAQ
Why focus on Google Cloud instead of AWS?
I have direct certifications and experience on GCP (Core Infrastructure, Essential, Elastic, Compute Engine). On Azure I work in enterprise projects. On AWS my experience is limited to specific workloads. For an AWS-heavy client, I partner with an AWS-certified colleague.
Can you help with a full Kubernetes migration?
For workloads that make sense on k8s (microservices, stateless applications, modern CI/CD) — yes, on GKE or AKS. For legacy monoliths, it usually doesn't make economic sense for SMBs — I recommend traditional VMs or simple containers.
How do we decide whether to go all-cloud or hybrid?
The Readiness Assessment gives you the answer based on your concrete workloads. As a rule, for companies with 50+ VMs and predictable workloads, hybrid beats all-cloud on TCO in 85% of cases.
What if I already have poorly documented cloud deployments?
First, an audit: what exists, what it costs, what it does. Then we decide whether to consolidate, migrate back, or refactor. Often, 20% of workloads generate 80% of the bill.
Can I contract just FinOps without migration?
Yes. Package 03 is exactly that — cost audit & optimization for your existing deployments, without major restructuring.
Let's start with a free cost check
Give me read-access on billing (AWS/GCP/Azure) for 2 months or CSV exports. In a week I'll send you 3 concrete optimization recommendations with estimated impact. Zero obligations.