At-a-Glance
- Client/Industry: Confidential / holding
- Scope: Turnkey datacenter pod in Equinix DX1 (Dubai): compute/virtualization, storage, network, security, backup, and migrations
- Delivery window: 4 months
- Environment footprint: 2 racks, 19 devices, 44U, ~16kW nominal power planning baseline
- Key outcomes: New UAE DC pod deployed and integrated with existing infrastructure; critical services migrated with defined rollback; operations-grade As-Built produced for handover and auditability
Challenge
The client procured a turnkey solution to deploy new infrastructure in a UAE-based Equinix DX1 datacenter, while also upgrading and migrating existing systems into the new environment from the existing environment hosted in one of carrier-grade co-location DCs in Europe
The build had to fit inside a constrained colocation footprint across two adjacent cabinets, while interconnecting cleanly with existing racks via overhead and ladder racking routes to simplify topology.
The highest-risk part was not the installation. It was the cutover: moving production services into a new core, new storage, and a new security perimeter without turning email and shared data into a business outage.
Solution Overview
- Production-grade virtualized platform anchored by a 5-host VMware cluster on xFusion 2288H V6 servers (dual Intel Xeon Gold 6348, 1 TB RAM per host)
- All-flash primary storage (Huawei OceanStor Dorado 5000 V6) and separate backup storage (OceanStor 5310 with dedicated disk enclosures), plus an HPE MSL2024 tape library for long-term retention
- 32G Fibre Channel SAN (dual switches) with documented zoning and port-level visibility for predictable multipath behavior
- Redundant switching and uplinks with LACP aggregation and dual-path server connectivity to eliminate single points of failure at NIC, port, and switch levels
- Check Point firewalls (CPAP-SG6400 class) in HA with clear segmentation between management, server farm, and DMZ-style security zones
- Turnkey Virtualization and Microsoft AD/Exchange services deployment
Project Specifics
- Two-rack engineering, not “just rack space”: 19 devices spread across cabinets for maximum redundancy and simplicity
- Server resiliency by design: dual 10G links per host, redundant access paths, and LACP at the switching layer to keep workloads online during link/switch events
- SAN correctness matters: dual FC switching with documented port maps and zones, enabling consistent storage presentation and troubleshooting without guesswork
- Multi-tier backup architecture: disk-based primary and secondary copies plus tape copy, with defined retention and schedule (details available; ops-relevant, not “marketing”)
- Service migration with rollback: Exchange migration executed as a controlled sequence including snapshots, maintenance mode, VMware conversion steps, verification gates, and a fallback plan if synchronization failed
- File services moved with integrity controls: Windows Storage Migration Service approach included inventory, transfer, validation/UAT, and decommissioning steps
- Operational readiness baked in: patching and update management (WSUS) documented with monitoring and security controls to reduce drift after handover
Project Delivery
The team sequenced work to keep risk contained: rack-and-power readiness first, then core connectivity and security perimeter, then compute + storage bring-up, and only then service migrations with explicit verification gates and rollback paths. The Exchange plan is a good example: it formalizes checkpoints (connectivity validation, backups, controlled node migration, sync verification).
As-Built Documentation
The As-Built goes beyond diagrams. It records rack layouts, device placement and power planning, server connectivity tables, SAN switch evidence (device summaries, port details, zoning screenshots), and platform configuration specifics (vCenter/cluster build, backup topology, retention policy, and service migration procedures). That level of rigor shortens incident response, simplifies audits, and makes future expansion predictable.
Result
A new datacenter pod was delivered as a cohesive system: 5-host VMware compute, all-flash primary storage, FC SAN, redundant switching, Check Point HA security, Virtualization cluster with all necessary Microsoft services on top and a multi-tier Veeam-to-disk-to-tape backup strategy, with migrations executed using documented procedures and rollback planning
