INTRODUCTION TO STRATISTOR
StratiSTOR is a software-defined storage platform engineered to deliver massive scalability, high performance, and unified block, file, and object services—purpose-built to power the most demanding enterprise and data-centric workloads.
EXECUTIVE SUMMARY
StratiSTOR is the software-defined data foundation of StratiSYSTEM OS. It unifies enterprise block, file, object, and high-performance storage services in a distributed platform built to meet the changing demands of applications, virtualization, private cloud, AI, and HPC.
Instead of tying storage capabilities to a fixed proprietary appliance, StratiSTOR runs on supported industry-standard Intel and AMD servers with flexible NVMe, SSD, HDD, and external block-storage options. Organizations can choose the media, network, and protection model that fits each workload while maintaining a consistent operational environment.
Scale-out architecture, replication, erasure coding, checksumming, snapshots, automated recovery, and centralized observability provide the foundation for resilient data services. Supported high-speed networking, metadata acceleration, and parallel storage designs extend the platform to demanding throughput and latency requirements.
Deploy StratiSTOR as a dedicated storage cluster, integrate it with StratiSERV, or converge compute and storage through HyperSERV. The objective is simple: keep data accessible, protected, and scalable without repeatedly replacing the infrastructure beneath it.
SERVICE ARCHITECTURE
StratiSTOR makes multiple data-access services available from one software-defined storage architecture. Block services use supported iSCSI and NVMe over Fabrics interfaces; file services use SMB and NFS; object services use S3-compatible access. Additional supported transfer methods, including HTTPS and SFTP, address workload-specific needs.
These protocols share the StratiSYSTEM operating foundation while retaining their own access, performance, availability, and security configuration. This lets teams consolidate diverse storage requirements without assuming every application needs the same storage policy.
Data can be distributed across supported NVMe, SSD, HDD, and external or multipath block devices. Node and device placement, redundancy, metadata layout, caching, and network design can be matched to the required capacity, throughput, latency, and failure tolerance.
The same platform supports all-flash, hybrid, capacity-focused, disaggregated, and specialized high-performance designs. Storage architecture follows application requirements—not a fixed appliance configuration.
STORAGE FOR MODERN INFRASTRUCTURE
Enterprise storage now carries more than transactional applications and shared files. It must also supply virtual infrastructure, container platforms, backup repositories, AI models and datasets, research pipelines, and services distributed across locations.
StratiSTOR addresses those demands with a common data layer that can be scaled and optimized by workload while preserving consistent administration, protection, and lifecycle operations.
RAPID DATA GROWTH
Data growth spans databases, file repositories, backup retention, telemetry, analytics, media, and AI model artifacts. Capacity planning must allow new storage resources to be introduced without creating another isolated storage environment.
StratiSTOR expands by adding supported storage devices and nodes, increasing available capacity and aggregate service resources while keeping data under the same storage operating model.
INCREASING PERFORMANCE AND LATENCY REQUIREMENTS
NVMe media, dense virtualization, large file services, and GPU-powered applications place pressure on storage latency, IOPS, and sustained throughput. At 100GbE and faster, CPU, memory bandwidth, PCIe topology, and software data paths can matter as much as device speed.
StratiSTOR supports appropriately engineered NVMe, metadata, network, and multi-node configurations, including high-speed Ethernet and supported RDMA-oriented fabrics. Performance is sized and validated across the full system, not inferred from drive specifications alone.
AI AND GPU DATA PIPELINES
AI systems consume large training datasets, model files, checkpoints, intermediate data, and inference artifacts. These workloads require high-throughput movement of data between storage and compute, often with concurrent readers and writers.
StratiSTOR provides scalable data services for private AI and HPC infrastructure. Supported NVMe, high-speed networking, GPU-aware data paths, and complementary Lustre deployments can be selected where the application and qualified hardware benefit from them.
VIRTUALIZATION AND CLOUD-NATIVE CONSOLIDATION
Operating separate storage islands for hypervisors, containers, databases, files, and backup increases complexity and fragments capacity.
StratiSTOR supplies a common data layer for StratiSERV, HyperSERV, compatible third-party virtualization, Kubernetes, and private-cloud environments. Each service can use the protection, access model, and performance characteristics appropriate to its workload.
DISTRIBUTED AND MULTI-SITE INFRASTRUCTURE
Business-critical data must often survive more than a failed disk or node. Organizations may need resilience across racks, data centers, geographic sites, or remote locations.
StratiSTOR supports architectures using fault-domain-aware placement, replication, disaster recovery, and qualified synchronous protection. Multi-site designs account for quorum, network latency, recovery capacity, and workload-specific availability goals.
INFRASTRUCTURE LIFECYCLE AND HARDWARE INDEPENDENCE
Traditional storage refreshes often force a new hardware purchase and a complex data migration at the same time. As data volumes grow, those refresh cycles become expensive and operationally disruptive.
StratiSTOR separates the storage operating layer from individual server generations. Supported hardware can be added, expanded, or retired in stages, allowing the data environment to evolve without a wholesale platform reset.
ADVANCED FEATURES FOR SCALE, PERFORMANCE, AND AVAILABILITY
StratiSTOR combines distributed storage, workload-specific data services, configurable protection, and centralized operations. The features below can be selected and tuned to meet different performance, resilience, and capacity objectives.
DISTRIBUTED SCALE-OUT ARCHITECTURE
Data and services are distributed across multiple participating nodes rather than constrained to a fixed controller pair. Adding qualified devices or nodes can increase aggregate capacity, throughput, and failure-domain diversity while preserving the cluster operating model.
UNIFIED BLOCK, FILE, AND OBJECT SERVICES
Support application storage through iSCSI and NVMe-oF, enterprise file access through SMB and NFS, and S3-compatible object interfaces. Protocols share the underlying data platform but can be configured with service-specific access, protection, and availability policies.
For demanding file and object workloads, performance can be scaled through suitable node, storage-media, and networking configurations. Available SMB features, including multichannel where supported, can improve use of multiple client/server paths.
NVME AND HIGH-PERFORMANCE STORAGE
Build storage tiers around all-NVMe, SSD, HDD, hybrid media, or qualified external storage. High-performance designs can combine NVMe-oF, high-speed Ethernet, optimized queue and CPU placement, and low-latency device access.
METADATA ACCELERATION
Place supported metadata-intensive operations on dedicated high-performance devices to improve responsiveness for large namespaces, small-file workloads, and directory- or object-heavy applications. Metadata capacity, endurance, and redundancy must be designed with the same care as bulk data.
SDCACHE
SDCache provides configurable cache acceleration and a supported device-presentation layer for selected storage configurations. Memory-based or disk-based options can be chosen according to locality, performance targets, and the protection of underlying data.
Cache sizing, write policy, and resource reservations are workload-specific; write-back designs require appropriate power-loss and recovery safeguards.
HIGH-SPEED NETWORKING AND ACCELERATED DATA PATHS
Supported Ethernet and fabric options range from conventional clustered networking to 100/200/400GbE and other qualified high-speed designs. Traffic separation, multipathing, NUMA locality, and switch configuration help deliver predictable storage performance.
StratiSYSTEM also supports selected SR-IOV, DPDK, VPP, VLAN, overlay, QoS, and hardware-offload configurations. The actual storage data path depends on protocol support and the validated NIC, driver, and platform combination.
CONFIGURABLE DATA PROTECTION
Select replication or erasure-coding policies to balance resiliency, usable capacity, write overhead, and recovery behavior. Data placement can reflect drive, node, rack, and site fault domains in a qualified cluster design.
Capacity efficiency options, including supported data compression, should be evaluated together with workload type, CPU overhead, and protection settings.
DATA INTEGRITY AND SELF-HEALING
Checksums, background scrubbing, consistency checks, and redundant data placement help detect silent corruption and support repair when a valid protected copy or erasure-coded reconstruction is available.
Distributed recovery restores redundancy following supported device or node failures. Operators retain visibility into scrub findings, recovery activity, placement balance, and media warnings; not every error can be corrected automatically.
SNAPSHOTS AND RECOVERY
Point-in-time snapshots support recovery, testing, rollback, and protection workflows for supported storage services. Snapshots complement—but do not replace—independent backups and a tested disaster-recovery plan.
ENCRYPTION AND SECURITY
StratiSTOR supports encryption at rest and secure access configurations where implemented for the selected service. Role-aware administration and protected management networking help control access to data and storage operations.
Local infrastructure services and offline software repositories also support restricted or disconnected StratiSYSTEM deployments, subject to the chosen service dependencies.
QUALITY OF SERVICE
Supported service and network controls can help manage bandwidth, I/O demand, and resource contention. QoS should be validated under representative mixed workloads to protect latency-sensitive applications.
MULTI-SITE AVAILABILITY AND DATA PROTECTION
StratiSTOR extends beyond a single node or data center through configurable protection and placement policies. Qualified multi-site architectures can address disaster recovery, geographic resilience, workload locality, and site-level fault tolerance.
SYNCHRONOUS MULTI-SITE PROTECTION
In a qualified synchronous design, writes can be acknowledged only after the active cross-site durability policy is satisfied. This supports zero RPO for acknowledged writes while full synchronous protection remains in effect; it does not imply zero recovery time or uninterrupted access.
Inter-site round-trip latency, bandwidth, network redundancy, quorum, and capacity headroom are central to the design. Degraded operation after a site failure may have different protection characteristics until the site is restored and synchronized.
SITE FAILURE PROTECTION
Fault-domain-aware placement can distribute protected copies or fragments across independent locations. Two-data-site designs may use a separate quorum or tiebreaker location to prevent competing writable partitions during supported failure scenarios.
MULTI-SITE ACCESS
Supported services can be designed for access from more than one protected location where application behavior, routing, quorum, and data consistency permit. Active/active access must be validated for the specific protocol and site topology.
OPERATIONS AND OBSERVABILITY
StratiSTOR integrates storage monitoring with the broader StratiSYSTEM management environment. Operators can review node and device health, cluster capacity, storage-service status, utilization, recovery activity, and relevant performance indicators from consistent administrative workflows.
Integrity and maintenance visibility includes background scrubbing, device warnings, inconsistent placement groups, and repair or recovery state. Placement optimization and balancing workflows can help redistribute load and capacity as cluster conditions change.
Web, CLI, and supported API workflows allow administrators and service providers to standardize provisioning, monitoring, maintenance, and troubleshooting without treating every node as an isolated storage appliance.
NON-DISRUPTIVE GROWTH AND LIFECYCLE MANAGEMENT
Storage infrastructure must be able to change without forcing applications to change with it. StratiSTOR supports qualified online expansion, rolling maintenance, and node replacement when the configured redundancy and cluster health permit.
New devices and hardware generations can be introduced incrementally, with data rebalanced or recovered across available resources. Maintenance planning must still account for remaining failure tolerance, client workload, and recovery bandwidth.
Expand capacity. Refresh the hardware. Preserve the data platform.
COMMON STRATISTOR WORKLOADS
StratiSTOR is designed for a broad range of enterprise and high-performance data needs. The storage media, protocol, protection, and network design can be tailored to each workload instead of prescribing a single configuration for all of them.
ENTERPRISE VIRTUALIZATION
Deliver shared block or file storage for StratiSERV, HyperSERV, and supported third-party hypervisors, with resilient data services and capacity growth independent of compute.
DATABASES AND TRANSACTIONAL APPLICATIONS
Combine suitable NVMe or SSD media, low-latency network paths, and workload-specific redundancy for transactional databases and business-critical applications.
AI TRAINING AND INFERENCE
Store and serve training datasets, checkpoints, model weights, inference data, and AI artifacts using scalable file, object, and high-performance storage services. Qualified GPU-oriented and Lustre configurations can be incorporated where required.
BACKUP AND RECOVERY
Build capacity-efficient backup repositories using protected storage services and suitable media, with integrity checks, throughput scaling, and growth matched to retention requirements.
FILE SERVICES AND COLLABORATION
Provide SMB and NFS services for shared enterprise files, engineering data, home directories, media, and collaboration workloads. High-speed networking and supported SMB multichannel deployments can address demanding client access patterns.
OBJECT AND APPLICATION DATA
Serve S3-compatible object workloads for application content, backup targets, analytics, cloud-native data, and archives, with protection and service topology configured for the intended access pattern.
ANALYTICS, MEDIA, AND RESEARCH
Support large-file movement, research data, rendering, simulations, and analytics with sustained throughput from qualified scale-out storage and network architectures.
DEPLOYMENT MODELS
Deploy storage in the form that matches the site: compact local services, highly available clusters, dedicated storage systems, converged compute, or multi-site infrastructure. The StratiSYSTEM operating model remains consistent across supported architectures.
SINGLE-NODE DEPLOYMENTS
Use a compact node for labs, edge, development, or workloads that explicitly accept the absence of node-level high availability.
MULTI-NODE STORAGE CLUSTERS
Distribute services and protected data across multiple nodes for enterprise storage, databases, virtualization, backup, and private cloud. Cluster topology determines actual fault tolerance and maintenance flexibility.
LARGE-SCALE STORAGE CLUSTERS
Aggregate devices, nodes, network bandwidth, and processing capacity for enterprise-scale file and object repositories, backup platforms, analytics, AI, and research workloads.
TIERED AND DISAGGREGATED ARCHITECTURES
Keep storage separate from compute when data growth, performance, cost, or availability requirements justify independent scaling. Mix supported media and node roles around the selected architecture.
HYPERCONVERGED INFRASTRUCTURE
Combine StratiSTOR with StratiSERV through HyperSERV to deliver storage and virtual-machine or container services on shared nodes with coordinated operations and protection.
HYBRID INFRASTRUCTURE
Integrate dedicated StratiSTOR clusters, HyperSERV systems, supported third-party block resources, and specialized performance tiers where different workloads require different physical designs.
FEDERATED AND MULTI-SITE ENVIRONMENTS
Extend protected data services across locations for disaster recovery, geographically distributed workloads, or qualified synchronous operation while accounting for network behavior, quorum, and data consistency.
DESIGNED TO EVOLVE WITH THE DATA
Enterprise data should outlive the hardware that stores it. StratiSTOR is designed around the ability to modernize servers, drives, networking, and capacity over time without abandoning the storage software and operational model.
As workloads shift from traditional applications to containers, AI, and distributed data services, StratiSTOR can add the storage capabilities and physical resources those workloads require.
Hardware changes. Workloads evolve. The data foundation remains.
CONCLUSION
StratiSTOR brings block, file, object, and high-performance data services together on a software-defined platform built for scale, resilience, flexible hardware, and unified operations.
From enterprise virtualization and databases to Kubernetes, private cloud, AI, backup, HPC, and multi-site infrastructure, StratiSTOR supplies a common storage foundation with workload-specific deployment choices.
Choose the workload. Design the data path. Scale the platform without starting over.