Thailand-based agro-industrial conglomerate Sangfah Agri Product Co., Ltd. has modernized its virtualization and storage environment with a 6-node hyper-converged platform based on Proxmox Virtual Environment and Ceph. The new infrastructure has improved provisioning speed, operational resilience, and cost predictability while establishing a highly scalable foundation for upcoming AI-driven quality control initiatives.

Founded in 1991, Sangfah Agri Product Co., Ltd. is one of Thailand's leading agro-industrial enterprises, processing and exporting rice and tapioca products to international markets. With a team of approximately 200 employees driving an annual revenue of USD 275 million, their operations encompass industrial milling facilities, extensive warehouse networks, private shipping ports, a 9.9 MW biomass power plant, and an expanding enterprise IT services division providing internet and data center services.

With thousands of metric tons of agricultural products moving through its supply chain every day, uninterrupted access to critical operational systems is vital. Inventory management, logistics coordination, production planning, export documentation, and customer-facing applications all depend on a reliable and highly available infrastructure platform. As the business continued to expand, Sangfah's IT team recognized that its existing virtualization environment was becoming increasingly difficult to scale.

The challenge: Limited agility and rising licensing costs

As VMware transitioned through its post-Broadcom restructuring, Sangfah’s IT leadership recognized that their existing legacy virtualization environment was no longer sustainable. This shift, combined with inherent operational bottlenecks, created the following challenges:

• Agility barriers: Provisioning new workloads required days rather than minutes, delaying internal project delivery and limiting operational agility.
• Maintenance downtime: Routine infrastructure upgrades and host maintenance demanded extensive cross-departmental coordination, forcing the IT team into scheduled maintenance windows during weekends and late-night hours.
• Disaster recovery gaps: While standard backups were available, Sangfah required a more robust, automated replication strategy to guarantee rapid workload recovery in the event of a site-level disruption.
• Escalating VCF licensing costs: Following recent industry shifts, revised VMware Cloud Foundation (VCF) license calculations would have multiplied Sangfah’s historic infrastructure spend, making multi-year budget forecasting increasingly volatile. By moving to Proxmox VE, Sangfah gained robust enterprise functionality without a bloated license stack. Because they did not require the highly advanced, niche add-ons bundled into legacy VCF models, switching to the open platform allowed them to reduce infrastructure costs by 60–70% while maintaining premium performance.

Building a hyper-converged architecture with Proxmox VE and Ceph

With support from Proxmox Partner ReadyIDC, Sangfah deployed a 6-node hyper-converged Proxmox VE cluster. By implementing Ceph distributed storage directly within the cluster, compute and storage layers scale linearly together, eliminating separate storage systems. Because Proxmox VE natively includes enterprise features like high availability (HA) and live migration out of the box, Sangfah cut out the expensive, specialized software licensing required by legacy vendors – proving that critical enterprise requirements can be delivered entirely on an open platform.

Each node is provisioned with dual high-core enterprise CPUs, dense ECC RAM, and multiple NVMe SSDs dedicated to Ceph Object Storage Daemons (OSDs). To ensure network resilience, the architecture utilizes high-speed, dedicated 10 GbE links for the Ceph storage backend, leveraging an Ethernet cross-connect topology with high-availability (HA) network interface cards (NICs). This critical backend traffic is completely isolated from standard VM and management network traffic.

Post-deployment performance has been exceptional, delivering sub-millisecond storage latency and highly consistent IOPS under heavy, mixed production workloads. Crucially, the cluster has maintained 100% uptime since going live. By eliminating single points of failure, this architecture ensures continuous uptime for Sangfah’s supply chain and commercial operations.

Strengthening data protection with Proxmox Backup Server

Data protection formed a key part of the infrastructure modernization initiative. Sangfah integrated Proxmox Backup Server for centralized backup management, client-side deduplication, and automated verification. Resilience is further strengthened through disaster recovery synchronization to a Proxmox Backup Server at a separate location, ensuring that backup data is replicated off-site and ready for rapid restoration. By combining cluster-level HA with off-site replication, the company established a significantly stronger disaster recovery framework than its legacy environment.

The migration playbook: Risk mitigation

To reduce migration risk, prevent data loss, and avoid relying on expensive external integration consultants, Sangfah’s internal IT team built the new Proxmox-based environment in parallel with the existing virtualization platform. The team executed a strategic migration over a period of 5 months:

• Automation (IaC): Before migrating production workloads, the team deployed a Proof of Concept (PoC) environment. Utilizing Terraform and Ansible, they fully automated cluster orchestration, validated Ceph storage performance, and tested HA failover scenarios.
• Native import wizard: Once validated, workloads were migrated in stages based on business priority. The IT team utilized the built-in Proxmox VE ESXi import tool, allowing them to seamlessly ingest legacy workloads directly through the web console without complex multi-step conversions.
• Resource efficiency: Sangfah’s internal IT operations team of up to 20 engineers successfully migrated approximately 40 VMs over a 5-month period. By leveraging partner-led training, provided by ReadyIDC, the team ensured smooth knowledge transfer and handled the deployment completely in-house.

Wanna Manee-in further explains: “Our engineers came away confident in managing the platform independently, significantly reducing our long-term reliance on external support. This training provided by Proxmox partner ReadyIDC gave our team a shared vocabulary and established a unified baseline of best practices that continues to pay dividends in daily operational stability.”

Business and operational outcomes

By decoupling infrastructure growth from arbitrary licensing tiers, Sangfah has secured cost predictability for its long-term infrastructure planning forecasts. The transition has driven immediate operational transformations across the enterprise:

• Zero unplanned downtime: The Proxmox VE cluster has maintained zero unplanned outages since deployment
• Rapid automated provisioning: The timeline to deploy new environments dropped from days to minutes through the use of standardized templates and automated IaC pipelines.
• Elimination of weekend maintenance: Using live migration, the IT team can now migrate workloads away from individual nodes and perform hardware and software upgrades during regular business hours without disrupting users.

The combination of open standards, a strong feature set, predictable costs, and an engaged community means Sangfah is no longer planning around licensing cycles or vendor roadmaps. They are planning around their own business priorities. "That shift in mindset has been just as valuable as the technical and financial benefits," tells Wanna Manee-in.

Looking Ahead: Powering AI-Driven Quality Control

The open architecture of Proxmox VE is now serving as the foundation for Sangfah’s next digital initiative: an AI-powered computer vision system on the production line to automate real-time rice quality inspection.

By adding GPU-enabled compute nodes to the cluster, the company plans to analyze high-resolution production-line imagery in real time, instantly identifying discolored grains, fractures, and foreign material. Because Proxmox VE natively supports both robust VMs and lightweight Linux containers on a single platform, Sangfah expects to run heavy, GPU-accelerated training workloads side-by-side with lightweight edge inference services - all while keeping sensitive operational and proprietary data securely on-premises.

For Sangfah, migrating to Proxmox VE was far more than a tactical response to changing vendor licensing models. It was a strategic step toward establishing a resilient, sovereign infrastructure platform built to power both their current global supply chain and their next generation of AI breakthroughs.

Wanna Manee-in
Managing Director at Sangfah Agri Product Co.,Ltd.

About Sangfah Agri Product Co., Ltd.
Sangfah is a Thai agro-industrial and technology group operating five core divisions globally. Its portfolio spans large-scale agricultural processing, international export infrastructure with private shipping ports, and central storage facilities. Sangfah also runs a renewable biomass energy plant and an enterprise IT division providing data server services.

About ReadyIDC
Established in 2011, ReadyIDC is a premier cloud provider and ISP in Thailand, serving over 50,000 users. As a Proxmox Reseller, they deliver secure, high-performance hyper-converged infrastructure (HCI) across a resilient 400 Gbps Tier III network.