A manufacturer’s plant apps must run close to machines with

The scenario's requirements strongly indicate the need for an on-premises solution. The sub-5ms latency requirement can only be reliably met by placing compute resources physically close to the machinery, a core principle of edge computing. Public cloud models cannot guarantee this level of performance due to network distance. Furthermore, the mention of predictable loads and a "7-year asset plan" directly aligns with a traditional Capital Expenditure (CapEx) model, where hardware is purchased as a long-term asset and depreciated over its lifecycle. This financial and operational model is characteristic of traditional on-premises infrastructure procurement.

A. Pure public cloud: This model cannot meet the stringent <5 ms latency requirement due to the physical distance and network unpredictability between the plant and the cloud data center.
C. Multi-cloud bursting only: This strategy is designed for handling unpredictable, variable workloads by offloading peak demand to the cloud, which contradicts the "predictable loads" described in the scenario.
D. Desktop-as-a-Service: This is an end-user computing solution for delivering virtual desktops and is irrelevant to running industrial applications that interact directly with plant machinery.

1. HPE White Paper, "Edge computing: What it is and why it matters" (Document ID: a00056983enw, Page 3, "Why edge computing now?"): This document states, "For some use cases, such as in industrial automation... a response time of a few milliseconds is required. It is not feasible to send data to a centralized cloud for processing and get a response back within this time frame." This directly supports the need for an on-premises/edge deployment to meet the latency requirement.

2. HPE Solution Brief, "HPE solutions for the intelligent factory" (Document ID: a00055544enw, Page 2, "HPE solutions for the intelligent factory"): This brief discusses deploying compute on the factory floor for "real-time data processing and analytics at the edge." This reinforces that manufacturing scenarios with low-latency demands require local, on-premises infrastructure next to the operational technology.

3. HPE Financial Services White Paper, "Shift from traditional IT procurement to a flexible consumption model" (Document ID: a00058384enw, Page 2, "Traditional IT procurement"): This document describes the traditional CapEx model as involving "upfront payments" and "owning and managing assets." This aligns perfectly with the "7-year asset plan" mentioned in the question, confirming that a traditional procurement model is the intended fit.