Choose the architecture for the outcome you require
Explore storage and virtualisation by workload, technology or operational objective. The right product is easier to identify once access, availability, recovery and growth are understood.
What runs?
How is data accessed?
What must improve?
Start with what the organisation needs to run
Each workload creates a different mixture of latency, throughput, concurrency, retention and recovery requirements.
VMware, Hyper-V and eEVOS
Compare storage for established hypervisors with eEVOS as an independent virtualisation platform with integrated Backup & Disaster Recovery.
Transactional workloads
Design for durable writes, predictable tail latency, path resilience and recovery that respects database consistency.
Database storage guide →Kubernetes persistent data
Align CSI integration, access modes, volume lifecycle and data protection with the cluster’s failure domains.
Kubernetes storage guide →Data-intensive compute
Keep compute supplied through aggregate throughput, metadata planning and a network designed for parallel demand.
AI and HPC storage guide →Editing, ingest and rendering
Plan for sustained concurrent streams, a shared namespace, predictable collaboration and recoverable project data.
Media storage guide →Continuous video recording
Balance continuous writes, retention, VMS compatibility, failure behaviour and reliable evidence retrieval.
Surveillance storage guide →SMB and NFS
Choose file access around clients, identity, permissions, locking, namespace and service availability.
SMB and NFS guide →Enterprise and MSP S3
Separate application object storage from provider requirements such as tenancy, capacity quotas, billing and branded end-user access.
Understand the access model before comparing speed
File, block and object storage expose different semantics. The protocol must fit the application before performance optimisation begins.
Block, file or object?
Understand who owns the namespace, filesystem and data lifecycle.
Compare data models →NVMe over Fabrics
NVMe/TCP and NVMe/RDMA bring NVMe commands across a network fabric.
Explore NVMe-oF →RDMA
RoCE, RoCEv2 and InfiniBand can reduce transport overhead when the complete path is engineered correctly.
Explore RDMA →iSCSI
A mature block-storage choice with broad host knowledge and familiar Ethernet operations.
Explore iSCSI →Fibre Channel
Dedicated enterprise block fabrics with established zoning, multipathing and operational practices.
Explore Fibre Channel →SMB and NFS
Shared files, permissions and locking for Windows, Linux/Unix and supported application environments.
Compare SMB and NFS →S3-compatible storage
API-based object access for applications, protection workflows and multi-tenant services.
Ceph with eEKAS
Scale-out block, file and object services across multiple nodes and explicit failure domains.
Explore scale-out Ceph →A faster protocol cannot correct the wrong data model. Application support, consistency, host ownership and recovery behaviour remain selection criteria even when bandwidth and latency are important.
Focus the design on the operational outcome
A project may combine several objectives, but each one changes a different part of the architecture.
Performance
Reduce latency or increase throughput by examining the entire application-to-media path.
Compare block protocols →High availability
Define which component, node or service failures must be absorbed automatically.
Explore HA and resilience →Scale-out
Grow capacity and services across nodes without a fixed local two-node boundary.
Explore eEKAS scale-out →Backup & disaster recovery
Separate availability, replication, historical recovery and disaster recovery by event and objective.
Backup and DR guide →Cyber resilience
Protect clean recovery points across separate trust boundaries and define how verified services return after an incident.
Ransomware recovery guide →Simpler operations
Use graphical workflows and integrated lifecycle controls to reduce routine command-line dependency.
Use the selection guide →Use one decision path across every technology
The Architecture and Storage Selection Guide turns workload, access, availability, recovery and growth requirements into an architecture shortlist. euroNAS products appear only after those requirements are clear.
Different euroNAS products serve different operating models
These are practical examples, not a universal ranking. Selection should follow the workload, failure boundary, recovery objectives and scale model.
euroNAS Premium
Storage OS for one physical server or Virtual Storage Appliance.
- File, block and S3 services
- Scale-up within the selected server
- Snapshots and asynchronous replication options
HA Cluster
Automatic local storage service failover within a supported two-node design.
- Synchronous Mirror or dual-controller shared storage
- File, block and S3 services
- Separate asynchronous replication options
eEKAS
Graphically managed Ceph storage for multi-node scale-out services.
- Ceph block, file and S3
- Failure-domain-based protection
- S3 File Manager, capacity quotas and billing options
eEVOS
A virtualisation product rather than a storage product.
- VM lifecycle, availability and live migration
- Integrated Backup & Disaster Recovery
- Uses internal, external shared or Ceph-based storage
See how the operating models differ
Standalone storage
hosts
server or VSA
Defined scale-up boundaries and straightforward ownership.
Premium example →Two-node storage HA
node A
node B
Automatic service failover with synchronous local protection or shared storage.
HA architecture guide →Distributed storage
endpoints
and services
Scale-out block, file and S3 across explicit failure domains.
eEKAS architecture →Virtualisation platform
compute
storage model
Virtual machines and integrated recovery using internal or external storage.
eEVOS architecture →Validate the complete business service
The final design must be checked against current application support, measured load and operational recovery procedures.
Application, host, hypervisor, protocol, driver and supported topology.
Capacity growth, concurrency, read/write mix, block or object size, latency and bursts.
Device, path, node, rack, power and site events that must be tolerated.
RPO, RTO, retention, restore granularity and responsibility during an incident.
Bandwidth, redundant paths, zoning or VLANs, RDMA configuration and monitoring.
Expansion, maintenance, capacity reserve, skills, testing and lifecycle ownership.
Start with the requirement, then shortlist the platform
We can review the workload, access method, failure domains, recovery objectives and growth model for a business storage or virtualisation project.