I'm responsible for 37 locations, most of which use small mini-PC-based setups. Each site runs local workloads such as POS systems, inventory management, CCTV, guest Wi‑Fi, print services, and other site-specific applications. Connectivity can sometimes be unavailable for days, so at least some services need to keep running locally.
I'm looking for a platform that can use our existing hardware while making the environment easier to operate. Ideally, we could standardize every site, reduce configuration drift, deploy or update VMs and containers across locations, and provide as much high availability, resilience, monitoring, backup, and centralized management as practical.
VMware's pricing makes it unattractive, and Scale Computing looks interesting but may require specific hardware. I'm also considering Proxmox, VergeOS, Hyper‑V with Azure Arc, and similar options. What approaches or platforms have worked well for managing this kind of distributed environment?
5 Answers
Connectivity is a major architectural issue here. If links can be down for several days, design for the site operating in a degraded but usable state: local authentication or cached credentials, local DNS and DHCP, local application data, and a clear resynchronization plan. Consider LTE, satellite, or a second wired provider where feasible. Also involve a PCI-compliance specialist for any POS environment, since payment-card requirements can affect systems beyond the register itself.
Before choosing a hypervisor, decide which services truly need to run at every site. Centralize anything that can tolerate WAN latency and keep only connectivity-independent or latency-sensitive workloads locally. Use site-to-site VPNs with redundant internet connections where possible, and host critical centralized services in two locations rather than maintaining a full stack at all 37 sites.
Proxmox can handle the virtualization side and works well with existing hardware, but the bigger challenge is operations. You’ll need a consistent build standard plus your own approach to monitoring, backups, patching, remote access, and disaster recovery across the sites. Its newer centralized management tools may help, but don’t assume the hypervisor alone solves fleet management.
VergeOS is worth evaluating if you want something closer to Scale Computing while retaining more flexibility around hardware. It is KVM-based and designed to simplify clustered virtualization. Scale Computing may also support selected third-party hardware, but verify every model against its certified hardware list before designing around it.
For a mostly Microsoft environment, Hyper‑V combined with Azure Arc can provide a practical management path. It’s especially worth considering if you already use Microsoft identity, administration, and cloud services. Whatever platform you select, make every site as identical as possible so staff can troubleshoot one standard stack instead of 37 unique environments.
That’s the main reason I’m keeping local compute. The goal is to standardize the local stack and manage it centrally whenever the connection is available, without making the site completely dependent on the WAN.