Capacity on demand and opex: growing core banking without betting your capital

By the Redsis Infrastructure team · October 2026 · 6 min read

The dilemma: grow now, without knowing for how long

The bank had budgeted its on-demand capacity to be fully consumed by August 2020. But from November 2019, business needs accelerated consumption, and at the same time changes to its regional consolidation strategy meant the platform had to stay in service through January 2022.

Then there was storage. The IBM DS8884 enterprise system behind the core was close to 90% full, and writes per second were already above the recommended guideline. This wasn't a future problem; it was a present operational risk.

The underlying question was as much financial as technical: how do you secure capacity for a horizon that might move again, without locking up capital in infrastructure whose future depends on strategic decisions still being made?

What IBM Power Capacity on Demand is

Enterprise-class IBM Power servers typically ship with more processors and memory installed than the client initially activates. Capacity on Demand (CoD) lets you switch those resources on when the business needs them, without replacing hardware or stopping operations. Broadly, IBM offers several models:

• Permanent activation of additional cores and memory, for structural growth.
• Temporary activation by the day or by usage, for business peaks, closes or campaigns.
• Shared schemes that let capacity move between servers wherever it's needed.

In this project, Redsis consolidated the bank's processor and memory resources into a single on-demand pool , from which capacity is allocated to each server as needed. That lets the bank expand quickly without giving up the flexibility, security and reliability its core applications demand.

Storage counts too

Adding compute without addressing storage would simply have moved the bottleneck. So the solution included expanding the IBM DS8884, part of IBM's enterprise storage family built for mission-critical workloads, with more capacity and more I/O operations to support growth over the whole period. Redsis handled implementation and migration of the storage behind the core, connected to Power servers running IBM i.

The financial model: an operating lease

The other half of the answer was financial. The entire solution, capacity, storage and services, was delivered under a 12-month operating lease with IBM Capital . And negotiating processor, memory and storage growth together gave the bank better commercial terms.

For an institution whose technology horizon is still taking shape, an opex model has clear advantages:

• It doesn't tie up capital in assets that may not be needed after consolidation.
• It aligns cost with usage and with the period when capacity is actually required.
• It simplifies the decision : one proposal that brings together technology, services and financing.

Five lessons for core banking capacity planning

1. Track actual consumption, not just the budget

The jump from 32 to 56 cores was visible because there was daily consumption data. Watching the trend, not just the contracted total, is what buys time to react before capacity runs out.

2. Treat compute and storage as one system

Disk occupancy and writes per second are as critical as cores. An upgrade that ignores either half only moves the problem.

3. Design for a horizon that can shift

Corporate strategies change dates. A solution that can be extended or adjusted, like on-demand capacity, protects the bank better than a purchase sized for a single date.

4. Get every stakeholder to the table early

The final proposal came out of several rounds between Redsis, IBM specialists and the bank's team. That joint work is what makes it possible to combine technology and financing in one answer.

5. Don't forget day-to-day operations

On-demand capacity has to be managed. Here, CoD administration was delivered through the Redsis Customer Service Center, so the bank could stay focused on its business.

The result: capacity when the business calls for it

The bank extended its compute and storage through January 2022 with no disruption to core banking, avoided the disk and write saturation risks, and gained the flexibility to handle business peaks without new capital investment. All on an opex model that kept pace with its regional consolidation roadmap.

Where to start

If your core is nearing its limits and your roadmap still has open questions, it pays to review actual consumption, storage health and your on-demand capacity and financing options before committing to a purchase. At Redsis we combine more than 25 years of infrastructure experience in Latin American banking with deep expertise in IBM Power, IBM i and IBM enterprise storage.