Practice
Cloud & DevOps · Application Modernization
Datacenter Exit · Microsoft Azure
Exit the datacenter.
Compound the savings.
A century-old global manufacturer was spending $2.35M annually on a datacenter that had become a liability — not an asset. Kansoft migrated 80+ enterprise workloads to Microsoft Azure. Zero business disruption. Savings that compound from year one.
~$2.35M → $1.77M annual hosting 80+ enterprise servers migrated Zero business disruption
Context
A hundred years of precision engineering. A datacenter that had stopped keeping up.
The client manufactures premium components for global markets — a business built on operational precision, with manufacturing, finance, identity, EDI, and analytics workloads running around the clock across multiple continents.
Their on-premise datacenter had become a $2.35M annual obligation — and the trajectory was upward. Aging hardware, growing support costs, and an infrastructure model that couldn't flex with the business. Leadership made the call to exit.
The question wasn't whether to move to cloud. It was how to execute a full datacenter exit without disrupting a business that ships, invoices, and reports every single day.
The Constraint
This wasn't a migration. It was a global enterprise removing the floor from under itself while the business kept running.
Three constraints pulled in different directions. The engineering plan had to hold all three.
01
Cost trajectory was unsustainable
Annual hosting, support, and infrastructure spend had crossed $2.35M, with the curve pointed upward. Continuing on-prem was not a steady-state — it was an accelerating obligation.
02
The estate was both old and entangled
80+ enterprise servers, hybrid networking, ADFS, EDI, PLM, analytics gateways — each with dependencies on the others, and none documented in a way that survived a clean cut-over.
03
The business does not pause for an exit
Manufacturing operations across multiple regions. EDI traffic with global partners. Reporting cycles that don't accept a maintenance window. Migration had to happen inside the running business.
Approach
A datacenter exit is a sequencing problem, not a lift-and-shift problem.
We structured the engagement in five phases — each gated, each measurable, each designed to make the next one safer rather than larger. Cloud rationalization came before migration. Hybrid networking came before workload moves. Cost modeling ran throughout.
PHASE 01
Cloud Rationalization
Inventoried every workload. Classified each for migration, re-platform, or decommission. The cheapest workload to run on Azure is the one you no longer need.
PHASE 02
Azure Target Architecture
Designed VM and PaaS landing zones with cost-aware SKU choices, private endpoints, and load balancing — not lift-and-shift defaults that quietly inflate bills.
PHASE 03
Hybrid Networking
MPLS connectivity and hybrid cloud routing established before workload migration began, so on-prem and Azure could co-exist during transition rather than be stitched together under pressure.
PHASE 04
Phased Migration
Workloads moved in dependency-aware waves — ADFS and identity first, application tiers next, analytics gateways last. Every wave had a tested rollback path.
PHASE 05
Datacenter Exit & FinOps
Decommissioned the legacy estate. Locked in optimized Azure spend with reserved capacity, right-sized SKUs, and a hosting model the finance team could forecast.
What We Built
62 VMs. 15 PaaS resources. One hybrid architecture that holds together when the datacenter is no longer there.
The deliverable was not the migration. The deliverable was a new operating model for enterprise infrastructure — one where compute scales with the business, networking spans cloud and remaining on-prem assets without fragility, and the next modernization initiative starts from a cloud-native foundation instead of fighting a legacy estate.
Enterprise Azure footprint
62 Azure Virtual Machines and 15 PaaS resources spanning identity, application, integration, file storage, and analytics tiers — provisioned with cost-aware SKU selection and right-sizing from day one.
Hybrid cloud architecture
MPLS-backed connectivity, private endpoints, and load balancing tied Azure workloads to the remaining datacenter footprint cleanly. No "we'll fix the networking later" technical debt.
Modern cloud operating model
Power BI Gateway, Integration Runtime, Azure File Share, ADFS, SMTP relay, Web EDI, EDM Library, and PLM QA — each repositioned on Azure with the operating model the business will use for the next decade.
Datacenter exit, fully executed
Not a phased reduction. A complete exit from the legacy estate, with infrastructure rationalization carried through end-to-end so the cost baseline actually resets rather than partially recedes.
Migration Footprint
Workloads Migrated
The work, named.
Every workload below crossed from the legacy datacenter into Azure with dependencies preserved, identity intact, and the operational runbook updated rather than rewritten under fire.
Power BI Gateway
Analytics & reporting bridge
Azure File Share
Enterprise file services
Integration Runtime
Data orchestration
ADFS Infrastructure
Federated identity
SMTP Relay
Mail transport
Web EDI
Global partner exchange
EDM Library Apps
Engineering document mgmt
PLM QA Workloads
Product lifecycle quality
What Changed
The infrastructure stopped depreciating. The savings started compounding.
These are the operating numbers the client now plans against. The gap between the old run-rate and the new is the kind of figure CFOs bring to the board.
~$1.26M
Recurring annual savings
FY27 onwards, against a $2.35M legacy baseline
~25%
Hosting cost reduction
$2.35M → $1.77M in the first full year
80+
Workloads migrated
62 Azure VMs, 15 PaaS, plus enterprise apps
0
Business disruption events
Manufacturing and reporting cycles ran through migration
| Measure | Before | After |
|---|---|---|
| Infrastructure hosting model | Legacy on-prem datacenter | Microsoft Azure + optimized hybrid |
| Annual hosting cost | ~$2.35M | ~$1.77M |
| FY26 cost reduction | Baseline | ~$618K saved |
| FY27 onwards (recurring) | Baseline | ~$1.26M annual savings |
| Infrastructure scalability | Limited, hardware-bound | Cloud-elastic |
| Datacenter dependency | High | Significantly reduced |
| Operational flexibility | Low | Improved across workloads |
| Architecture posture | Legacy estate | Cloud-native foundation |
What It Means for the Business
Beyond the cost line: an infrastructure posture the next decade of strategy can build on.
Cost savings were the trigger. What the business gained was optionality — for modernization, for analytics, for AI-readiness — starting from a cloud-native foundation rather than fighting a legacy estate.
Cost optimization at scale
The hosting baseline reset rather than partially declined. ~$618K removed in year one. ~$1.26M removed every year that follows. Predictable, modeled, and visible to finance.
Infrastructure modernisation
Workloads now sit on Azure-native services with elasticity, observability, and a deployment posture that supports — rather than slows — the next wave of modernization initiatives.
Reduced operational overhead
Hosting, patching, and support operations consolidated. Engineering hours that were going into keeping the lights on are now available for work that compounds.
Strategic datacenter exit, complete
Not a deferred decommission. Not a partial reduction. A full exit from the legacy estate, executed without disturbing the business that depends on it.
Cloud-native starting point for what comes next
Every modernization initiative that follows — analytics, AI, ERP upgrades, security hardening — now starts from Azure, not from a legacy estate. That's a fundamentally different strategic position.
Hybrid done properly
Cloud and remaining on-prem assets connect through MPLS-backed hybrid networking — not held together by improvisation. The architecture survives the next decade, not just the migration window.
Engagement
How the work was shaped.
Cloud Platform
Microsoft Azure (VMs, PaaS, File Share, Load Balancers, Private Endpoints)
Data & Integration
Power BI Gateway, Integration Runtime
Networking
MPLS connectivity, hybrid cloud architecture
Why Kansoft
Datacenter exits fail in predictable ways. This one didn't.
Under-scoped dependencies. Networking stitched together under pressure. Cost models that look right until the first Azure bill arrives. The engagement was structured specifically to prevent each of those failure modes — not respond to them.
Azure migration executed, not learned Multi-region, multi-workload migrations across regulated and mission-critical environments. The patterns from prior engagements prevented problems this client never saw.
Full datacenter exit, start to finish Not a partial decommission. Not a phased reduction that leaves half the legacy cost intact. A structured exit from a live enterprise estate — delivered without disrupting operations.
FinOps built into the architecture Cloud spend optimization was designed in from day one — right-sized SKUs, reserved capacity, cost-aware landing zones. The CFO-visible savings came from decisions made at the architecture stage, not after the bill arrived.
Hybrid networking that holds MPLS-backed connectivity and private endpoints established before migration — not improvised during it. The architecture connects Azure and remaining on-prem assets cleanly, and was designed to outlast the migration window by years.
Planning a datacenter exit?
Your legacy run-rate is a strategic decision. Treat it like one.
If you're carrying infrastructure that costs more every year than it enables, the calculus has already shifted. The question now is how to execute an exit without making the next 18 months harder than the last five years. Our Cloud & DevOps practice has done this before — we can show you what a structured Azure migration looks like for your specific estate.
Speak with our Cloud & DevOps teamConfidential · No procurement obligation · Response within one business day