Designed the data center facilities themselves, not only what ran inside them, including the disaster recovery site: floor plan and rack elevations, hot and cold aisle containment, dual UPS with battery plant and maintenance bypass, independent A and B electrical paths from separate panels and transformers, and overhead busway power distribution.
An architecture is only as available as the power and cooling underneath it, and the failure I have seen most often is a resilient logical design sitting on a single power path nobody drew.
Then instrumented what had been designed: data center infrastructure management monitoring the full power chain, UPS output across both feeds and generator power, underfloor water detection, humidity, and hot and cold aisle temperatures to confirm the containment was performing the way it was drawn. Designing the facility and then instrumenting it closes the loop — you find out whether the drawing was right.
Applied in
- Citywide Network Total Cost of Ownership StudyA weighted scoring instrument for a manufacturer standard ahead of a refresh — which earned itself within three weeks, when a reseller's savings claim turned out to be priced against the wrong baseline.
- Data Center Estate and Colocation ProgramSix tiered sites for the city and five external agencies — and a municipal IT department selling colocation against commercial carriers, and winning on engineering.
- Cecil Disaster Recovery Data CenterFacility and network design for the recovery site — power, cooling and carrier entry as well as what ran inside.
- E911 Caller Location and Backup Answering PointMaking the dispatcher see a room instead of a phone number — then building a second answering point so the dispatch center was not a single point of failure.
- Duval County Courthouse Technology ProgramA $12M IT program inside a $350M, 800,000 square foot capital facility, cut over live to a fixed date.