HCI Solution

Hyperconverged Infrastructure, Rebuilt for the AI Era

EasyStack HCI unifies compute, storage and networking into a single software-defined resource pool on the EOS Digital Native Engine — starting from just 3 nodes and scaling to thousands, replacing legacy three-tier infrastructure and VMware with production-grade simplicity.

Isometric illustration of a hyperconverged infrastructure platform: three glowing server nodes unified into a single cube with orbiting storage, compute, network and AI accelerator icons
Executive Summary

A production-grade HCI foundation — open, evolvable, AI-ready

Enterprise infrastructure is at a turning point. Traditional three-tier architectures — separate compute, storage and networking — are too complex, too slow to scale, and too costly to operate. At the same time, VMware licensing changes, data sovereignty regulations, and the rise of AI workloads are forcing organizations to rethink their virtualization and cloud foundations.

EasyStack Hyperconverged Infrastructure delivers a software-defined alternative built on the EOS Digital Native Engine. It unifies compute, storage and networking into a single, horizontally scalable resource pool — starting from just 3 nodes and scaling to thousands — and evolves seamlessly to private cloud and AI-ready infrastructure with optional EAF.

Replace legacy virtualization, cut TCO by up to 35%, deploy in 4 hours, upgrade with zero business interruption — and extend smoothly to AI when you are ready.

Replace three-tier & VMware

Consolidate legacy virtualization and siloed compute/storage/network into a single software-defined stack.

Up to 35% lower TCO

Hardware reuse, efficient resource pooling and per-node pricing — utilization rises from <20% to >45%.

99.99% availability

HA mode ≥99.9%; cloud mode ≥99.99% with automated fault detection, backup, snapshot and CDP built in.

AI-ready with optional EAF

Same platform extends to GPU · NPU · DCU pooling and Model-as-a-Service — no rip-and-replace later.

Why HCI Now

Four forces making HCI the default foundation

Legacy three-tier limits, VMware economics, sovereignty rules and AI workloads are converging at the same time.

The limits of traditional three-tier

Separate compute, storage and networking silos create complex procurement, slow provisioning, high operational overhead and poor utilization — often below 20%.

Days → weeksTraditional resource provisioning cycles
< 20%Typical utilization in siloed three-tier estates

VMware licensing shock

Broadcom's acquisition of VMware has driven price increases of 200% – 1,200%, eliminated perpetual licenses and imposed minimum bundle purchases. HCI delivers the same consolidation benefits as vSphere/vSAN with lower cost, open APIs and hardware-agnostic flexibility.

74%of IT leaders actively exploring VMware alternatives (Gartner)
35%of VMware workloads will migrate to alternative platforms by 2028

Data sovereignty & compliance

Indonesia's PDP Law, Thailand's PDPA and Europe's GDPR require data to remain within national borders or under strict jurisdictional control. HCI enables private, on-premises deployment with full sovereignty — without three-tier complexity.

On-premFull data residency with private hyperconverged deployment
PDP · PDPA · GDPRAligned with major regional data-protection regulations

AI & edge workloads demand new infrastructure

AI inference, real-time analytics and edge applications require low-latency, high-throughput infrastructure. Traditional CPU-only HCI cannot efficiently support GPU, NPU or DCU accelerators. EasyStack HCI is designed to evolve — start general-purpose, extend to AI-optimized with optional EAF.

GPU · NPU · DCUOptional EAF extension for heterogeneous AI accelerators
Edge → CoreLightweight edge deployment with centralized management

What three-tier silos cost you

The recurring pain points HCI eliminates on day one.

ProcurementComplex integrationMulti-vendor compute, storage and network sourcing cycles
ProvisioningDays to weeksSlow resource delivery blocks business agility
OperationsHigh overheadSpecialized skills required across silos
UtilizationBelow 20%Stranded capacity across separate pools
ScalingSpike painDifficulty absorbing traffic surges elastically
Architecture

One engine, five layers — the HCI stack

EOS Digital Native Engine underneath; compute, storage, network, management and observability unified above.

EOS — Digital Native Engine

The continuously iterating unified base and operating system dedicated to cloud platforms.

  • Microservices-based, fully symmetric distributed architecture
  • Manages over 150 microservices
  • OTA smooth upgrades — zero business impact, zero data migration, zero service interruption
  • Unified lifecycle management for all platform components
MicroservicesSymmetric HAOTA upgrade150+ services

Compute

KVM-based virtualization for VMs, containers and bare metal — unified in one pool.

  • KVM hypervisor with VM, container and bare-metal support
  • DRS multi-dimensional scheduling with 20+ resource metrics
  • Live migration, HA and fault self-healing for core components
  • Proactive balancing evolves from reactive to predictive placement
KVMVM · Container · Bare metalDRSLive migration

Storage

Distributed block, file and object storage — NVMe-optimized and hardware-agnostic.

  • Block, file and object interfaces in one software-defined stack
  • NVMe optimization for low-latency, high-IOPS workloads
  • Ceph, LocalPV, NFS and external SAN/NAS integration
  • Replication, erasure coding, snapshots, clones, CDP and DR
Block · File · ObjectNVMeCeph · LocalPV · NFSSnapshot · CDP

Network

Software-defined networking integrated with virtualization, containers and bare metal.

  • SDN with unified overlay for VMs, containers and bare metal
  • 10 / 25 / 40 / 100 / 200 GbE fabric support
  • RoCE v2, DPDK and SR-IOV for low-latency, high-throughput paths
  • Multi-tenant isolation with self-service networking
SDNRoCE v2DPDK · SR-IOVUp to 200 GbE

Management

Unified graphical interface with self-service portal and automated operations.

  • Single pane of glass for compute, storage, network and workloads
  • Self-service portal reduces VM delivery from days to minutes
  • 100+ built-in inspection items and one-click smooth upgrades
  • Open APIs compatible with OpenStack and AWS S3
Self-serviceOne-click upgradeOpenStack · S3 APIs100+ inspections

Observability

Unified monitoring across distributed, heterogeneous storage and compute pools.

  • Capacity, IOPS, throughput and latency in one view
  • Cross-backend visibility for heterogeneous storage pools
  • Visual operations cut daily inspection time by > 50%
  • Automated alerting with fault self-healing hooks
Unified monitoringCapacity · IOPSVisual opsSelf-healing
Deployment Modes

Start small. Scale to thousands.

01

3-node minimum

Start small with hyperconverged HA — production-grade from the very first cluster.

02

Scale-out

Add nodes linearly up to 20 in HCI mode; continue scaling to 100+ nodes in individual mode.

03

Cloud deployment

Extend to large-scale private cloud with ≥ 5 control replicas and ≥ 99.99% availability.

04

Edge

Lightweight deployment at remote sites with centralized full-lifecycle management.

Hardware compatibility — multi-generation, cross-ISA

Protect existing investments and avoid vendor lock-in.

CPU architecturesIntel Xeon (E5 v3/v4 → Sapphire Rapids gen 6), AMD EPYC 1–5 gen, ARM Kunpeng / Phytium, Hygon, Zhaoxin
AI acceleratorsOptional with EAF — NVIDIA GPU, Ascend NPU, Hygon DCU
Storage mediaNVMe, SSD, HDD — mixed-media pools with tiering
Network10 / 25 / 40 / 100 / 200 GbE, RoCE v2, DPDK, SR-IOV
Core Capabilities

Six capabilities that make HCI production-grade

Pooling, distributed storage, HA, intelligent scheduling, simple operations and openness.

Capability

What EasyStack HCI delivers

Unified Resource Pooling

Compute, storage and network pooled into a single logical resource domain — dynamic allocation lifts utilization from typical 15–20% to over 45%.

Distributed Storage

Block, file and object interfaces; NVMe optimization; replication, erasure coding, snapshots and clones; unified monitoring across heterogeneous backends.

High Availability & Data Protection

≥ 99.9% in HCI HA mode, ≥ 99.99% in cloud mode with ≥ 5 control replicas; automated fault detection with MTTR reduced to minutes; backup, snapshot, CDP and local/remote DR built in.

Intelligent Scheduling (DRS)

Built-in Distributed Resource Scheduler with a multi-dimensional awareness algorithm considering 20+ metrics — evolving from reactive to proactive VM placement.

Simple Operations

4-hour deployment from bare metal to usable cloud platform; 100+ built-in inspection items; one-click smooth upgrades; visual ops cut inspection and troubleshooting time by > 50%; fault self-healing.

Open & Compatible

No hardware binding, no ecosystem lock-in, no closed APIs. Compatible with OpenStack and AWS S3 APIs; supports third-party integration via standardized interfaces.

Migration & Deployment

Move off VMware — without the maintenance window

Block-level, CDP-like migration tools keep production VMs running during cutover.

VMware alternative, built in

EasyStack provides professional online / offline block-level migration tools utilizing CDP-like technology. This enables smooth migration of high-concurrency, production-grade core business VMs that would otherwise require downtime on other platforms.

  • Significantly reduces maintenance windows for core business systems
  • Supports multi-VM concurrent migration at production scale
  • Enables on-demand migration for different business systems
  • Rapid re-deployment of business systems after migration
  • Preserves familiar VM, cluster, datastore and network semantics
Illustration of six HCI deployment scenarios unified on a single hyperconverged foundation
Business Value & ROI

Quantified value — Forrester TEI validated

Based on the Forrester Total Economic Impact methodology and in-depth interviews with Asian customers.

$1.23M Annual average value per enterprise Based on a 100-physical-server deployment
400% Three-year ROI Payback period of only 8 months
~$1.2M Software license savings Over three years versus legacy virtualization
–35% Hardware procurement reduction Through reuse and efficient resource pooling
–40% Operations team size reduction Unified management plane and automation
99.99% Core business availability Proven across banking, securities and hosting
–70% Unplanned downtime Core business downtime reduction
–60% Known security vulnerabilities Smaller trusted computing base with KVM stack

Cost savings

One energy customer lifted server utilization from under 15% to over 45% after consolidating with EasyStack. Replacing VMware with self-developed compute virtualization eliminated significant licensing fees, and many customers reused legacy servers to preserve existing assets.

Productivity gains

"Previously we needed storage, networking and virtualization experts. Now our team only needs to be familiar with EasyStack — making hiring simpler and reducing labor costs." Visual operations and one-click inspection cut daily inspection and troubleshooting time by more than 50%.

Agility & risk reduction

Self-service portal reduces VM provisioning from days to minutes. Built-in backup, snapshot and CDP capabilities improve backup completion by nearly 100%. KVM-based stack shrinks the trusted computing base — one government customer reported ~60% fewer known vulnerabilities than their previous VMware solution.

Customer Evidence

Proven in production — across regions and industries

From cloud hosting providers to national-scale banking private clouds.

Malaysia Malaysia · Cloud Hosting

Shinjiru Technology Sdn Bhd

Deployed EasyStack ECF 621 as its unified hyperconverged cloud infrastructure, replacing legacy virtualization. Modular scalability and integrated high-performance storage absorb peak-season traffic spikes while centralized resource management cuts operational overhead.

99.99% availabilityHours → minutes provisioning–50% ops overhead
Indonesia Indonesia · Cloud Service Provider

Leading Cloud Service Provider

Reused legacy servers to migrate from Mirantis and completed 4 major-version smooth upgrades over 5 years. Operations cost reduced significantly while customer SLA remains consistently at the top of the market.

–60% ops cost99.99% SLA4 major upgrades
Singapore Singapore · Research

Advanced University Research Platform

Using EasyStack Cloud Platform since 2019 to support GPU-intensive AI research projects. Deployment cost dropped substantially while cluster utilization climbed to research-grade efficiency.

–40% deployment cost75% cluster utilizationGPU-intensive AI
China China · Banking

Large State-Owned Commercial Bank

Built one of the largest private clouds in China's banking industry, with over 20,000 servers. Uses a one-cloud-multi-chip architecture to manage Intel x86 and ARM servers, running core transaction systems, channel systems and data warehouses.

20,000+ serversOne-cloud-multi-chip99.99% availability
China China · Securities

Large Comprehensive Securities Firm

Smoothly evolved from traditional virtualization to private cloud. Deployment scale exceeds 100 physical nodes; completed a no-interruption major-version online upgrade involving 150+ microservices in 2024. Unified management of VMs, containers and bare metal supports high-frequency trading, ultra-fast market data and AI quantitative research.

100+ nodesZero-interruption upgradeVM · Container · Bare metal
China China · Energy

Large Energy Group Consolidation

Consolidated fragmented three-tier estates into a single hyperconverged platform. Server utilization rose from under 15% to over 45%, and legacy hardware was reused to protect existing investment while modernizing the infrastructure foundation.

15% → 45% utilizationLegacy hardware reuseUnified pool
Deployment Scenarios

Six scenarios, one hyperconverged foundation

From VMware replacement to AI-ready infrastructure — the same platform evolves with your business.

Illustration of six HCI deployment scenarios unified on one hyperconverged foundation

VMware Alternative

Migrate vSphere / vSAN workloads to an open, cost-effective HCI platform with block-level, CDP-like migration tools that keep production VMs running during cutover.

Block-level migrationMulti-VM concurrentDRS-style schedulingPer-node pricing
Illustration of six HCI deployment scenarios unified on one hyperconverged foundation

Private Cloud Foundation

Start with a 3-node HCI cluster and scale to a large-scale private cloud with ≥ 5 control replicas — the same operational model from day one to thousands of nodes.

3-node startScale-out to 100+99.99% availabilitySelf-service portal
Illustration of six HCI deployment scenarios unified on one hyperconverged foundation

Sovereign Cloud

Private deployment for government, finance and regulated industries — data stays inside national borders while meeting PDP Law, PDPA and GDPR requirements.

Data residencyPDP · PDPA · GDPRRegulated industriesOn-premises
Illustration of six HCI deployment scenarios unified on one hyperconverged foundation

Cloud Hosting Provider

Multi-tenant HCI with self-service portal, dynamic scaling for traffic spikes and centralized resource management — proven by Shinjiru Technology in Malaysia.

Multi-tenantSelf-service portalTraffic-spike elasticity–50% ops overhead
Illustration of six HCI deployment scenarios unified on one hyperconverged foundation

Edge & Remote Sites

Lightweight HCI deployment at remote and branch sites with centralized full-lifecycle management from a single control plane.

Lightweight footprintCentralized managementRemote sitesFull lifecycle
Illustration of six HCI deployment scenarios unified on one hyperconverged foundation

AI-Ready Infrastructure

Optional EAF extends the same HCI platform to GPU, NPU and DCU governance with Model-as-a-Service — protect investments and embrace the intelligent era without rebuilding from scratch.

GPU · NPU · DCUModel-as-a-ServiceOptional EAFSmooth evolution

Ready to modernize your infrastructure with HCI?

Talk to our solution architects about VMware replacement, 3-node pilots, sovereign deployment paths, or extending your HCI platform to AI-ready infrastructure with optional EAF.