Home / Infrastructure
04 / Infrastructure
Infrastructure strategy, hybrid and cloud, resilience and local AI, led by someone who has built the estates underneath.
I came up through the engineering, racking, cabling and commissioning real datacentre estates, and that is exactly what my architecture and leadership are now built on. Having done the physical work, I now lead the decisions above it: what to run where, what to modernise, and how to weigh commercial, operational and resilience trade-offs across physical, virtual and cloud.
That foundation spans government, NHS, policing, financial services, publishing, architecture, airports, MSPs and Crown Hosting environments, covering greenfield builds, zero-downtime migrations and infrastructure worth tens of millions, alongside several interim IT manager roles. Today I bring it to hybrid and cloud strategy, where sound decisions about what to move, keep or modernise depend on genuinely understanding what you are moving from.
The renewed interest in on-premises and hybrid infrastructure is a practical response to data sovereignty, AI inference economics and cloud licensing, and organisations increasingly need leaders fluent in both worlds.
Capabilities
Datacentre & infrastructure design
Own the design of datacentre and server-room estates, covering capacity, power and space strategy, resilience topology, and the standards facilities and engineering teams build to. Grounded in having done the build, now directed at architecture level.
Power & cooling
Set the capacity and resilience envelope: power budgeting, UPS and PDU topology (N+1, 2N) and thermal strategy, including the concentrated heat loads of GPU-dense AI hardware. These are the decisions that shape cost, uptime and headroom.
On-premises security & microsegmentation
Zero Trust does not stop at the cloud. I architect on-premises microsegmentation (VMware NSX, hardware-enforced boundaries), IAM and PAM, and network segmentation that blocks lateral movement across the estate, a layer cloud-native architects rarely cover.
Local AI infrastructure
As data sovereignty and inference economics pull AI back on-premises, I architect local GPU platforms, covering sizing, private inference endpoints, RAG on local data and converged stacks such as Nutanix GPT-in-a-Box, bringing private AI within reach of mid-market organisations.
Hardware platforms
I have racked, cabled, commissioned and migrated production compute across HPE ProLiant, Dell PowerEdge and Lenovo ThinkSystem, including hyperconverged installation and cluster expansion on Nutanix and VMware vSAN. That covers the unglamorous parts that decide whether a platform survives its first year: firmware and BIOS baselines, out-of-band management on iLO, iDRAC and XCC, disk and NIC layout, hardware compatibility lists, and the burn-in and handover evidence a regulated estate needs.
I also carry that through to the network and fabric layer, from Cisco datacentre and campus switching to NVIDIA Mellanox high-speed interconnect for GPU and inference nodes. The same person sizes the cluster, specifies the fabric, agrees the power envelope and then signs off the design, which is why my cloud decisions are grounded rather than theoretical.
Server & HCI platforms
Hyperconverged and traditional three-tier builds end to end: sizing, procurement, physical installation, cluster creation, node expansion and lifecycle. Nutanix and VMware vSAN clusters on HPE, Dell and Lenovo hardware, plus rack-and-stack, cabling and commissioning of the estates they sit in.
HPC, GPU & high-speed interconnect
High-performance and AI-adjacent compute: GPU node sizing, PCIe and thermal constraints, and the NVIDIA Mellanox InfiniBand and high-speed Ethernet fabrics that keep accelerators fed. My hands-on work here is local AI inference platforms and Mellanox fabric rather than large supercomputing clusters, and I architect HPC-pattern platforms on that basis.
Cisco & datacentre networking
Cisco Nexus datacentre switching and vPC topologies, Catalyst campus switching, ASA and Firepower firewalling, and routing across BGP, OSPF, WAN and MPLS. Designed, configured and cut over in live environments, including segmentation and firewall policy that holds up to accreditation review.
High availability datacentres
Designing out single points of failure across the whole stack: dual-path power and network, N+1 and 2N topology, redundant fabric and controller pairs, clustered and stretched platforms, and failover that has been tested rather than assumed. Delivered as greenfield builds and as zero-downtime migrations of live production workloads.
Storage, backup & recovery
I have installed, migrated and run primary storage and data protection platforms rather than only reviewing them: HPE Nimble, Alletra and 3PAR arrays, Cohesity, Veeam and Rubrik, and backup-to-disk targets including HPE StoreOnce, Dell EMC Data Domain and virtual tape libraries. That includes the migration work, moving data off ageing arrays and legacy tape onto disk and immutable platforms without losing recoverability along the way.
Above the tooling sits the strategy: 3-2-1-1-0, immutability and air-gapped copies, retention against regulatory requirement, recovery tiering by RPO and RTO, and clean-room restore for ransomware scenarios. Advising boards on backup strategy and infrastructure risk across Nutanix, Rubrik and Cisco estates is a current part of my consulting work.
Primary & secondary storage
SAN, NAS and hyperconverged storage design and installation, array migration and capacity planning, plus the fabric and multipathing underneath. Cohesity for consolidated secondary data, and HPE arrays as primary storage in production estates.
Backup, immutability & recovery
Veeam and Rubrik deployments, backup-to-disk architectures with deduplication and replication, immutable and air-gapped copies, and documented recovery testing. Designed so that the recovery path is proven before an incident, not discovered during one.
Vendor and technology matrix
A plain summary of platform coverage for readers and automated skills review. Hands-on means personally installed, configured or migrated in production. Architected means owning the design, sizing and assurance decisions.
| Domain | Vendors and platforms | Depth |
|---|---|---|
| Server hardware | HPE ProLiant, Dell PowerEdge, Lenovo ThinkSystem | Hands-on build, commissioning and architecture |
| Hyperconverged | Nutanix (AHV, Flow, GPT-in-a-Box), VMware vSAN and VCF | Hands-on installation and migration, NCP / NCM and VCP certified |
| Virtualisation | VMware ESX to vSphere, NSX, Nutanix AHV, Hyper-V | Twenty-five years, large-scale hypervisor migration |
| HPC and GPU | NVIDIA GPU platforms, local inference nodes | Hands-on inference platforms, architected HPC-pattern design |
| Networking | Cisco Nexus, Catalyst, ASA and Firepower; NVIDIA Mellanox InfiniBand and Ethernet | Hands-on configuration and cutover, BGP, OSPF, WAN and MPLS |
| Storage | HPE Nimble, Alletra and 3PAR; Cohesity | Hands-on installation, migration and capacity design |
| Backup and recovery | Veeam, Rubrik, HPE StoreOnce, Dell EMC Data Domain, VTL | Hands-on deployment, plus board-level backup strategy advice |
| Datacentre facilities | Crown Hosting, colocation, UPS and PDU, CRAC and CRAH, structured cabling | Hands-on build and greenfield design, high availability topology |
| Cloud | AWS, Azure, Oracle Cloud, Microsoft 365 | Architected migrations, cloud foundations, IAM and DR |
The short version for anyone reviewing these skills: I can specify, procure, rack, cable, commission and migrate physical datacentre hardware, and I can architect and govern the cloud platforms that sit alongside it. Both sides, with a delivery record behind each.
Virtualisation & hypervisor expertise
Twenty-five years across the VMware arc, from ESX through vSphere, vSAN, NSX and Cloud Foundation, delivering virtualisation estates for the NHS, the Metropolitan Police, financial services, and central and local government. Through the post-Broadcom shift I lead hypervisor strategy and large-scale migrations, including VMware to Nutanix AHV programmes and Hyper-V assessments in cost-constrained environments, architecting the patterns that move production workloads between platforms with no service disruption. VCP, NCP and NCM certified.
These are programme-level decisions as much as technical ones, spanning licensing models, hardware compatibility, patching and firmware strategy, and the operating practices infrastructure teams depend on. I pair that with workload-level security: NSX and Nutanix Flow microsegmentation, tenancy isolation and in-cluster Zero Trust enforcement across hybrid physical and virtual estates.