Collect the inputs that shape the workload
| Input | What to provide |
|---|---|
| Active use | Concurrent activity, busiest periods and reconnect assumptions. |
| Product model | Instrument set, account/group requirements and relevant session behaviour. |
| Connected work | CRM, reporting, market-data, execution and other integration demands. |
| Data and growth | Retention, storage growth and the planning period. |
| Operation | Regions, access roles, monitoring, maintenance and recovery priorities. |
Label each input as measured, supplied by a provider or assumed. Unknowns should become discovery tasks, not hidden guarantees in a server quotation.
Translate requirements into acceptance criteria
Describe the workflows that must operate, the conditions under which they are tested and the records required to assess them. Include the routine jobs that can coincide with peak activity. Define who interprets a failed test and which organisation can correct the relevant component.
Check software permissions, version compatibility and supported deployment choices before sizing the infrastructure. More hardware does not resolve an unavailable interface, incompatible connector or missing authority to configure the environment.
Compare a complete scope
Ask each proposal to identify supplied resources, component locations, access, external dependencies and ongoing duties. Keep implementation, recurring hosting, integration work and third-party obligations visible as separate items, without assuming that one platform label includes them all.
Use the hardware test method for procurement evidence. For a new branded environment, attach this brief to the MT5 White Label project; for an existing installation, include the measured baseline and the specific operational problem to solve.