PGEC Cuts Costs 20% and Builds Real Disaster Recovery with Proxmox VE

By implementing Proxmox Virtual Environment and Proxmox Backup Server, Prince George Electric Cooperative solved a rising VMware cost problem and an unreliable single-node disaster recovery site. The cooperative built a fully clustered, failover-ready environment and cut costs by roughly 20%, all managed as a service by Proxmox partner Richweb.

Prince George Electric Cooperative (PGEC) was founded in 1939 to bring electric power to rural south-central Virginia families that the cities had never reached. It grew from about 300 customers on 165 miles of line in 1940 to the 12,000-member, 1,300-mile system it operates today, serving homes, farms, and businesses across 6 counties from its headquarters in Waverly, Virginia (US). Through its subsidiary Ruralband, PGEC also runs a fiber-to-the-home broadband network, adding a second capital-intensive infrastructure business to the same lean IT footprint.

PGEC is a member-owned cooperative, not an investor-owned utility, part of a national model rooted in the US Rural Electrification Act of 1936. Like most of the nation's roughly 900 electric cooperatives, it's governed by a board elected from its own membership, and any margin it earns goes back to members rather than to shareholders. PGEC draws its wholesale power from Old Dominion Electric Cooperative (ODEC), one of 13 member-owned distribution cooperatives serving Virginia. That ownership structure means every infrastructure dollar has to be justified back to the membership.

Outgrowing VMware, and outgrowing DR

PGEC ran its production workloads on VMware vSphere 7.0.3. Staying current meant an upgrade to vSphere 8 and a shift to core-based licensing, which would have significantly increased the co-op's core count and its bill. For a nonprofit cooperative, that rising cost curve was hard to justify against the limited benefit it delivered members.

The more urgent problem was disaster recovery. PGEC's DR site was a single host running on local storage, with no shared storage and no realistic path to recovering the full production environment in a real failover.

When PGEC priced out a VMware-based DR build that could actually do the job, the combined cost of new licensing and new hardware sent the budget climbing fast. PGEC needed 2 things at once: a way off the VMware cost curve, and a DR site that could genuinely stand in for production if it had to.

Choosing Proxmox to reduce costs and strengthen resilience

PGEC brought in Richweb, an official Proxmox Partner, to solve both problems together rather than patch them separately. The goal wasn't to find a discount alternative to VMware. It was to land on an open platform that could carry the co-op's infrastructure forward without the licensing model working against it every time PGEC grew.

Instead of spending PGEC's budget on per-core licenses, Richweb redirected it into better hardware and a genuinely resilient architecture, on an open platform that reduces dependence on proprietary licensing models.

A production and DR design built on Proxmox Virtual Environment

Richweb rebuilt PGEC's infrastructure around 2 Proxmox Virtual Environment clusters, one for production and one for disaster recovery, along with Proxmox Backup Server.

The production cluster runs on a 3-node Dell PowerEdge R740 cluster, upgraded to 768 GB of RAM per host, with a Pure Storage X20R4 array presented over iSCSI and NFS. Onsite backups run through a local HVENS Vault appliance (a Dell R750) running Proxmox Virtual Environment and Proxmox Backup Server on ZFS.

At the disaster recovery site, a 3-node Dell PowerEdge R630 cluster with 512 GB of RAM per host replaces the old single-node setup with true clustering and shared storage. Backups replicate from production to DR over a 10G E-Line, carried on PGEC's own Ruralband network.

The HVENS Vault on the DR site, also a Dell R750, is split by purpose: an SSD and HDD tier holds the Proxmox Backup Server datastore, while a dedicated NVMe ZFS pool is setup as live restore storage. That split is what turns the DR site from a passive backup target into a site that can actually carry production: in a failover, VMs are restored straight from Proxmox Backup Server onto the NVMe pool and run on the R630 cluster. Richweb tests this failover process quarterly against a defined set of production VMs, with consistent successful results.

Migrating 33 VMs in 2 weeks

PGEC migrated 33 VMs in a two-week window, executing restores across just 5 nights to fit tight maintenance windows, using Proxmox Virtual Environment's built-in ESXi import wizard. Migrations mixed direct imports with VM-level rebuilds depending on each workload's complexity, sequenced by how critical the system was to daily operations.

As part of the same migration project, the organization retired 3 older Dell R720 hosts that had been running KVM for Ruralband. Those workloads were consolidated into the main clustered environment after the production hosts were expanded to 768 GB of RAM per node.

The results

• Cut licensing and operational costs by roughly 20% compared with staying on VMware.
• Migrated 33 VMs off VMware vSphere in 2 weeks, using Proxmox VE's built-in ESXi import wizard.
• 3-node production and DR cluster replacing the previous single-node DR architecture
• 4-hour RTO and 12-hour RPO for the defined DR environment, with quarterly failover testing to validate recovery readiness.
• 3 legacy KVM hosts retired and consolidated into the main production cluster.
• 24/7 monitoring of hardware and the Proxmox environment.
• Infrastructure patching, monitoring, and upgrade scheduling managed by Proxmox partner Richweb.

For a cooperative that has to justify every dollar to its membership, the savings and the closed DR gap mattered in equal measure. Rather than pouring more budget into VMware licenses, PGEC redirected that cost difference into hardware, storage, and recovery capacity.

More time for applications, less time managing infrastructure

The technology changes also reduced the operational burden on PGEC's internal IT team. With Richweb managing patching, monitoring, and upgrade scheduling, PGEC’s team is free to focus on applications rather than the platform underneath them. According to President and CEO Sarat Yellepeddi, the decision also held up on technical and cost fronts from the start:

This migration gives PGEC a platform built to grow, not just a fix for an aging DR site. As Ruralband's fiber footprint expands and PGEC's own virtual infrastructure grows with it, the clustered Proxmox VE environment scales without the licensing penalties that came with VMware. Richweb stays on as the managed partner behind it, so PGEC has ongoing technical support built into the relationship, not a one-time project that ends at go-live.

Sarat Yellepeddi
President and CEO, Prince George Electric Cooperative

About Prince George Electric Cooperative (PGEC)
Prince George Electric Cooperative, chartered in 1939 and headquartered in Waverly, Virginia, is a member-owned electric cooperative serving more than 12,000 members along 1,300 miles of power line across 6 counties. Through its Ruralband subsidiary, the cooperative also provides fiber-to-the-home internet service across its rural territory. Learn more: www.pgec.coop

About Proxmox Partner Richweb
Richweb, Inc. is an engineering-led IT solutions provider founded in 1995 and headquartered in Ashland, Virginia. We design, operate, and support critical IT infrastructure for organizations across Virginia and the Mid-Atlantic, delivering expertise in network engineering, managed IT services, Proxmox virtualization and support, private cloud hosting, and infrastructure modernization. Learn more: https://www.richweb.com/