What the work actually looks like.

Situation

A financial services business moving into a new multi-floor office, with a regulator-driven requirement that card-handling systems be demonstrably separated from the corporate network.

Constraint

The fit-out schedule was fixed by the lease, and segmentation had to be designed before any cable was pulled. Retrofitting separation into a flat network after occupancy is substantially more expensive.

Approach

• Survey and pathway design before the walls were closed
• Cat6A horizontal distribution with fibre backbone between floors
• Separate physical paths and clearly distinct labelling for regulated systems
• Segmented network design delivered as part of the same scope, not handed to a second vendor
• Certified per-port test results and as-built documentation at handover

Evidence produced

Full port-level test results, floor plans and cable schedules supplied as the audit evidence pack.

Situation

A production operation where editing throughput had become the bottleneck, with staff moving large media files across a network originally specified for office use.

Constraint

Production could not stop. Any change had to be staged around active editorial deadlines, with a rollback path at each step.

Approach

• Measured actual throughput and contention before proposing anything
• Fibre backbone and storage-network redesign sized against real file sizes, not nominal specification
• Phased switchover executed outside production hours
• Storage tiering so archive material stopped competing with active projects
• Documentation handed over so the in-house team could operate it

Evidence produced

Before-and-after throughput measurements on the same test transfer, recorded in the handover pack.

Situation

An international operator whose platform team was blocked behind infrastructure work, with deployments infrequent, manual and consequently high-risk.

Constraint

Traffic peaks around events meant capacity could not simply be reduced, and licensing obligations required a technical audit trail of changes.

Approach

• Kubernetes platform with autoscaling matched to event-driven peaks
• Infrastructure defined as code so environments became reproducible and reviewable
• CI/CD pipelines with an auditable record of every production change
• SAST and dependency scanning integrated at thresholds the team would keep
• Golden paths so product teams could deploy without infrastructure tickets

Evidence produced

Change records generated automatically by the pipeline, usable directly as licensing audit evidence.

Situation

A provider with a disaster recovery plan that had been written, filed, and never executed: a common and unmeasured risk.

Constraint

Clinical systems could not be taken offline for a test, so recovery had to be proven without disrupting care.

Approach

• Business impact analysis to set recovery objectives against clinical rather than commercial impact
• RTO and RPO agreed per system with clinical leadership
• Recovery exercises run against isolated restored copies rather than live systems
• Runbooks rewritten to be executable by someone who did not author them
• Findings from each exercise remediated and re-tested

Evidence produced

Dated recovery test records with measured restore times, retained as audit evidence.