Combating Hardware Price Hikes: Six Enterprises Reuse Existing Servers to Drive Cost Efficiency

As demand for AI continues to rise, IT hardware prices, led by memory chips, have increased sharply, driving up server prices as well. According to IEEE Spectrum, if no new production capacity comes online, IT hardware prices may remain high until 2027 or even longer*.

*How and When the Memory Chip Shortage Will End: Despite new fabs and new technology, prices will stay high. IEEE Spectrum.

Therefore, for enterprises building new data centers or renovating existing ones, a pressing challenge is how to minimize the impact of rising hardware costs on IT infrastructure projects, ensuring high-performance and stable production workloads at lower hardware spending.

With broad compatibility, flexible delivery options, and heterogeneous cluster support, SmartX Enterprise Cloud Platform (SmartX ECP) enables enterprises to flexibly reuse existing servers. Enterprises can build a unified hyperconverged resource pool using servers of different brands, models, and components, avoiding high procurement costs for memory, chips, and other hardware while upgrading infrastructure, replacing VMware, and boosting cost efficiency.

Reuse Servers with SmartX ECP: Broad Compatibility, Flexible Delivery, and Heterogeneous Support

Wide hardware compatibility: SmartX ECP is not tied to specific hardware. It supports x86/ARM servers from multiple vendors, including ITAI servers, CPUs such as Hygon, Kunpeng, and Zhaoxin, and AI chips from Intel, AMD, Nvidia, Ascend, and others, flexibly meeting hardware reuse requirements. >>Use the SMTX OS Hardware Compatibility Validation Platform
Flexible delivery options: SmartX ECP can be delivered as pure software or as an appliance. It also supports modular combinations of infrastructure capabilities, including compute, storage, networking, security, disaster recovery (DR), Kubernetes services, management, and AI platforms, allowing users to choose based on their needs.
Unified management of heterogeneous clusters: Servers with different CPU frequencies, memory specifications, disk capacities, and other configurations can be deployed and managed in the same cluster, helping enterprises effectively integrate new and old devices while reducing operations and maintenance (O&M) workload and costs.

If existing servers were previously used to support VMware virtualization clusters, users only need to add a small amount of hardware, such as SSDs and 10GbE NICs, to upgrade to an HCI architecture. For servers originally used as VMware vSAN Ready Nodes, SmartX ECP provides full compatibility, enabling hardware reuse with almost no additional investment.

*When reusing hardware, users should still refer to the vendor’s compatibility list for server and component requirements, and keep hardware configurations within the same cluster as consistent as possible to avoid affecting overall cluster performance. In addition, although the distributed storage in SmartX ECP ensures reliability through a multi-replica mechanism, devices that have been in use for more than three years are generally not recommended due to the risk of simultaneous failures across multiple devices and potential procurement difficulties in the future.*

Industry Practices: Reusing Servers and Replacing VMware to Cut Software and Hardware Costs

Huashan Hospital Affiliated to Fudan University: Reusing vSAN Servers to Build a Heterogeneous ECP Cluster with a 60% Reuse Rate

When replacing VMware, Huashan Hospital planned to continue using eight domestic servers from its original vSAN cluster to save on new hardware procurement, while replacing only the hyperconverged software with a domestic alternative.

After learning about the hospital’s hardware reuse requirements, SmartX engineers conducted a detailed analysis of the existing servers. Because the eight servers had been purchased in three batches, their CPU, memory, and disk configurations varied, and some batches presented reliability risks. SmartX engineers provided detailed hardware renovation strategies for each batch and designed a phased, cyclic live migration plan to ensure that the migration would not affect upper-layer business operations.

In the end, Huashan Hospital only added a small number of SSDs, RAID cards, and 10GbE NICs to reuse the servers and build a heterogeneous ECP cluster using servers with different CPU models and frequencies. The cluster is managed centrally through CloudTower, achieving a 60% reuse rate for existing hardware such as x86 servers and storage controllers. Going forward, the hospital also plans to reuse Nutanix servers to build SmartX ECP clusters and host more business system VMs.

Learn more: Replacing VMware: Building a New Infrastructure for Core Healthcare Services via SmartX ECP

A Leading Automotive Wiring Harness Manufacturer: Reusing 10+ vSAN Servers to Run All Virtualized and Containerized Applications

When replacing VMware, a leading automotive wiring harness manufacturer fully reused more than 10 existing vSAN servers. By replacing only disks nearing the end of their lifecycle, the company built a SmartX ECP infrastructure cluster and seamlessly continued using its hardware.

Using the SMTX Migration Tool, the user migrated 200+ VMs from the VMware platform to SmartX ECP and completed the business switchover within a normal downtime window. The new platform now supports the stable operation of all production application systems, including MES, SAP, SRM, and big data platforms.

At the same time, the user introduced the SMTX Kubernetes Service (SKS), distributed firewall (Everoute), and SMTX File Storage (SFS) modulars of SmartX ECP to support containerized application systems and improve east-west traffic security in the virtualization, achieving both infrastructure capability upgrades and hardware cost savings.

A Semiconductor Packaging and Testing Solution Provider: Replacing VMware with Reused Servers and Avoiding the Purchase of 100+ Memory Modules

A semiconductor packaging and testing solution provider previously used “VMware virtualization + centralized storage” to support production systems. Due to external factors, the user would no longer be able to continue using overseas products such as VMware and urgently needed a feasible domestic alternative.

After multiple rounds of comparison and validation, the customer selected SmartX ECP to replace VMware and deployed it in factories in Wuhan and Shanghai. In phase two of the project, considering the cost of new servers, the user reused eight existing domestic servers and only added SSDs and 10GbE network cards to build an ECP cluster, avoiding the purchase of 100+ 64GB memory modules.

Based on reused servers, the ECP cluster stably supports key application systems across OA, production line, and DMZ environments, helping the user modernize and localize IT infrastructure at a lower cost.

Shenwan Hongyuan Securities: Reusing Servers to Build an Offsite DR Center and Reduce Costs by 50%

When building an off-site DR center for mission-critical systems, Shenwan Hongyuan Securities wanted to use existing servers to further reduce total costs.

With SmartX ECP’s support for reusing various types of servers, the user reused existing servers that had previously served as standby machines in the production environment. After simple upgrades such as adding disks and memory, the user adopted SmartX ECP software to build a 40-node cluster for the off-site DR environment. The cluster uses a 10 GbE network and requires no new dedicated SAN storage switches.

In addition, SmartX ECP’s support for heterogeneous clusters enables the user to use CPUs with different frequencies and disks with different I/O performance in the same cluster, meeting the performance requirements of different applications and reducing hardware procurement costs by about 50%.

Learn more: Shenwan Hongyuan Securities: An In-depth Transformation From Legacy Virtualization to Hyperconvergence

An Insurance Asset Management Institution: Reusing 10+ Servers and Centrally Managing ITAI and x86-Based Clusters

Based on SmartX ECP, an insurance asset management company has built a unified cloud foundation supporting both ITAI and non-ITAI systems through multiple project phases.

In phase one, the user built hyperconverged clusters in two forms, based on SmartX native virtualization ELF and integrated deployment with VMware ESXi, to support core production, Dev/Test, OA, ITAI business systems, and other scenarios.

In the later expansion, the user further reused more than 10 existing servers and added only a small number of hardware components to quickly build two ECP clusters, supporting dozens of VMs for DMZ internet services and dedicated-line interconnection services, significantly improving resource utilization.

Currently, multiple ECP clusters are centrally managed by CloudTower, reducing costs and improving efficiency across IT resources and O&M management.

A Leading Trust Institution: Reusing Physical Machines to Build a VM-Container Convergence Infrastructure for Higher Resource Utilization

Based on SmartX ECP (using SKS), a leading trust institution successfully deployed a hybrid-architecture Kubernetes cluster based on HCI VMs and reused physical machines.

The solution innovatively enables precise tiered utilization of heterogeneous resources: a hyperconverged VM cluster based on Hygon-CPU serves as the elastic layer for general containerized applications that require rapid scaling, while reused Intel-based physical machines are added to the container cluster as bare-metal nodes to supplement container resources and host customized containerized workloads.

Through SKS, the trust institution seamlessly integrates these two types of underlying compute resources, achieving a unified view and efficient scheduling for all Kubernetes workload clusters.