Requirements & scope
Record applications, workload, site and budget; agree essential specifications and acceptable alternatives.

Explore computer and AI infrastructure design, from requirements and architecture to bills of materials and acceptance.
Hardware and application reference imagerySix guides connecting requirements, configuration and acceptance. Open a guide for the complete design approach.
Build maintainable standard configurations for documents, multitasking, video meetings and fleet deployment.
Plan models, GPU memory, power and software as one system.
Choose platforms around business workloads, maintenance windows and recovery needs.
Separate active data, shared capacity and backup copies; design backwards from recovery needs.
Design switching, optics and operations from traffic patterns and topology.
Turn requirements, substitutions, costs and acceptance criteria into reviewable deliverables.
Record applications, workload, site and budget; agree essential specifications and acceptable alternatives.
Prepare architecture, BOM and compatibility lists; validate a pilot with representative workloads.
Accept against agreed criteria and retain configuration, warranty, backup and expansion records.
Start with weights, then account for context, concurrency and runtime overhead.
Start with traffic and topology rather than port speed alone.
Keep a verifiable record of configurations, test conditions and acceptance criteria.
Understand M.2, NVMe, DDR, USB-C and memory-card speed classes before creating a shortlist.
Read the selection guide and FAQ ↗Consider models, users, data and site conditions together to create testable configuration requirements.
Explore AI infrastructure ↗Model support, drivers and specifications change. Consult current vendor documentation when confirming a configuration.