On-Premises VMware to Kubernetes Migration
Confidential Enterprise (Financial Services)
Context
A major financial services organisation running workloads on traditional VMware infrastructure needed to modernise to Kubernetes whilst maintaining strict compliance and security requirements. The existing estate supported hundreds of solution teams across multiple business units.
Challenge
Design and deliver a fully automated Kubernetes platform on VMware (in-house data centre) supporting Kubeadm-driven clusters, including SDN and network architecture, automated provisioning, SecOps compliance, and onboarding of over 100 solution teams.
Approach
Architected Kubernetes clusters on VMware with comprehensive automation using Terraform, Go, YAML, and scripting pipelines via Thoughtworks GO CD. Implemented OpenEBS-based storage automation, cluster monitoring with NewRelic and Splunk, SecOps controls for compliance, and fully automated Helm-based deployment pipelines for both cluster addons and solution workloads. Extended to AWS and Azure public cloud with KOPS-driven clusters and CircleCI automation.
Delivery
Multi-phase engagement: VMware cluster architecture and automation (8 weeks), Helm-based CD pipeline development (4 weeks), SecOps implementation (4 weeks), solution team onboarding programme with technical authoring and workshops (ongoing). AWS and Azure cloud clusters delivered in parallel workstream.
Outcomes
100+ teams onboarded
Solution teams onboarded with comprehensive technical workshops and documentation for SREs and developers
Fully automated CD
Helm-based deployment pipeline for cluster addons and solution workloads across on-prem and cloud
SecOps compliance
Security and compliance requirements met across all clusters with automated policy enforcement
Legacy & Sustainability
Full-stack developer platform with API-based cluster orchestration, automated onboarding, and CRD/Operator-style backing services.
Stack
Timeline
7+ years (continuous engagement)
What's Next
Progressive migration to cloud-native tooling patterns. GitOps-based deployment models replacing legacy pipelines.
Client identity is confidential. Detailed references and outcomes available under NDA.
Request References