Skip to content
BankingCloudManaged services

Cutting cloud costs without cutting service.

Confidential European bank

3 Min

Two engineers with laptops in a corridor beside illuminated server racks

The case study in 30 seconds

Paying more for cloud, and getting less from it.

A European bank’s AWS footprint had become expensive, with many legacy instances in a hybrid cloud model. Support cost too much and delivered too little: SLAs were poor, response was slow and the backlog of issues kept growing. DevOps and cloud management were visibly delaying service delivery and development.

Arenema took on DevOps as a managed service. A FinOps review, better resource management and reservations cut cloud costs by 25%. Automation and self-service reduced tickets by 23%, and a shared resource model delivered headcount savings of 47%.

The bank also gained follow-the-sun support around the clock, without asking anyone to work nights.

The challenge

Rising costs and a growing backlog.

The bank came to Arenema with four problems:

  1. 01A costly EC2 footprint, with many legacy instances in a hybrid cloud model
  2. 02High labour costs to support AWS, with poor SLAs and slow response
  3. 03A large backlog of issues caused by poor support, weak planning and a high level of toil
  4. 04Clear evidence that DevOps and cloud management were delaying service delivery and development

Cloud cost was the symptom. How the platform was run was the cause.

Our approach

Run it as a service, and fix the economics.

Arenema took responsibility for DevOps as a managed service, working on four fronts: what the bank paid for cloud, how its platforms were built, how much work was automated, and how support was staffed.

  1. 01. FinOps

    Paying for what the bank actually uses

    Better resource management, reservations and a FinOps review reduced cloud costs by 25%.

  2. 02. Architecture

    Certified expertise, scalable platforms

    Certified subject-matter experts and architects helped the bank make better product decisions. Kubernetes services, a service mesh and Red Hat OpenShift improved scalability, and a journey to cloud using VMware migration and VMware Cloud on AWS cleared on-premise data-centre space and improved disaster recovery compliance.

  3. 03. Testing and automation

    Safer releases, less toil

    The Istio service mesh enabled canary and blue-green testing. Automation reduced manual toil and sped up the resolution of support issues, and self-service scripts cut ticket numbers by 23%.

  4. 04. Managed service

    Follow-the-sun, not night shifts

    A shared resource model delivered headcount savings of 47%. Follow-the-sun coverage through the managed service replaced the need for teams to work night shifts.

Lower costs, fewer tickets.

What changed once DevOps ran as a managed service.

  • −25%

    cloud costs

  • −23%

    support tickets

  • 47%

    headcount savings

The outcome

A leaner platform, and support that keeps up.

The bank runs a leaner AWS estate at lower cost, with fewer tickets, faster resolution of support issues and coverage around the clock.

Moving workloads out of its on-premise data centre also improved its disaster recovery compliance.

Share

Categories:Banking · Cloud · Managed services

Next projects

View all

Every organisation faces different technology challenges. Let’s explore how we can help you move forward.