Datacenter migration
21 August 2025
At Global CIO, Ammar Shareef, CIO of KEENU, challenged the prevailing cloud-first narrative by sharing his team’s journey to a hybrid, on-premises (on-prem) — driven not by trend, but by regulation, performance, and cost realism.
KEENU adopted an OPEX hardware rental, stood up a hyper-converged private cloud, drilled Primary→DR→Primary failovers, and avoided hard-coded API endpoints.
The result: a hybrid architecture built not on trends but trade-offs — tailored to scale, compliance, and resilience. KEENU’s case shows why infrastructure strategy must start with context, not convention.
Speaker & Context
Ammar Shareef is the Chief Information Officer at KEENU, a regulated fintech company operating under an Electronic Money Institution (EMI) license granted by the State Bank of Pakistan. At KEENU, he led a complex data center migration project that began as a response to regulatory pressure and matured into a case study in operational adaptability and cost realism.
Migration Is Not One-Size-Fits-All
Migration has no universal template; strategy must be context-driven:
- Cost structure: CAPEX-heavy on-prem vs. flexible OPEX in the cloud
- Control and compliance: Local data residency mandates may rule out public cloud
- Latency and networking: High-performance or real-time workloads may demand physical proximity
At KEENU, all these variables were at play — and they shaped not only the destination but also the path of migration.
The KEENU Experience: Why We Moved
KEENU’s migration journey began with a regulatory trigger. In 2020, the team started exploring long-term infrastructure options. By 2024, they faced a clear decision point: the cloud provider they were using could not meet the Tier III data center certification newly required by the State Bank of Pakistan. The regulator did not prohibit cloud usage outright — but for latency-sensitive workloads like payment processing, compliance and performance both demanded local control.
In practice, recurring submarine-cable incidents periodically degraded cloud latency in Pakistan — unacceptable for payment transactions.
Rapid growth made scalability critical, yet a full on-prem shift faced CAPEX limits. Procuring enterprise-grade infrastructure in Pakistan involves not only supply chain delays but also dollar-denominated payments — a significant barrier in a tightly regulated financial sector.
To overcome this, Shareef and his team negotiated a rental-based OPEX model with hardware vendors — mimicking cloud OPEX while regaining control. Once hardware arrived, the team completed the full migration in approximately six months, balancing regulatory urgency with operational pragmatism.