Features

Core Object Storage Capabilities

FeatureDescription
S3 compatibilityS3-compatible API: PUT, GET, DELETE, LIST, HEAD, versioning, multipart, and tagging.
Scale-out object storageDisaggregated, erasure-coded object store; 40-80 PB usable per cluster, and beyond through federation.
Erasure codingErasure coding across nodes and drives, with configurable node- and drive-level resilience.
DurabilityDesigned for >12 nines object durability.
VersioningFull object versioning with listing and rollback.
Object lockingCompliance and Governance mode locking for immutability and retention.
Object capacityBounded by metadata-drive capacity - each object consumes a fixed metadata footprint, so provision metadata drives to your expected object count. See Hardware Sizing.
Space limitsEffectively unlimited through multi-cluster federation; constrained only by physical resources.

Operations & Management

FeatureDescription
High availabilityMulti-node active/active clusters (three or more nodes) with no single point of failure; survives node, drive, and rack failures.
Cluster managementAdd and remove nodes, auto-discovery, and background rebalancing.
Power-failure resilienceTested for data integrity under sudden full power loss.
Monitoring & metricsBuilt-in dashboards and alerting, plus an OpenMetrics-compatible metrics endpoint for your own tooling.
Remote monitoringOptional managed monitoring of your clusters via one.exaba.com.
Usage & billingPer-tenant usage tracking and metering, with reporting for billing.
Web console / UIConsole or web-based UI for management and observability.
Ecosystem integrationsCertified for Veeam Ready; works with Commvault, Acronis, rclone, and other S3-compatible backup and data tools.
Certificate managementBuilt-in certificate issuance and renewal for TLS endpoints.

Security & Compliance

FeatureDescription
EncryptionRSA 2048/3072/4096, AES-256, SSE-S3 / SSE-KMS / BYOK, FIPS-aware (runs under FIPS-mode Linux kernels).
Zero trustMutual TLS on every internal service-to-service call.
Tenant isolationStrong cryptographic and network isolation per tenant.
Auditable binariesReproducible builds, signed and auditable crates.
Remote syslog / SIEMLog forwarding to SIEM and central log platforms.
Reduced attack surfaceUserspace data path reduces exposure to kernel-level vulnerabilities.

Performance & Scale

FeatureDescription
NetworkingNVMe-oF data path over RDMA (RoCEv2 or iWARP), or TCP for sites without an RDMA fabric; 10/25/100/200 GbE Ethernet.
Throughput10+ GB/s of object throughput per node; aggregate throughput scales as nodes are added.
StreamingUpload and download streams with low RAM usage.
NVMe / flashWrite patterns optimized for flash and NVMe endurance.
Hybrid arraysNVMe metadata with SATA/SAS bulk HDD configurations.
Node limitsClusters start at three nodes (five or more recommended) and scale to around 20 per cluster; larger footprints via multi-cluster federation.
Metadata mirroringMetadata replicated across multiple drives and nodes.

Extensibility & Developer Features

FeatureDescription
Management APIsREST APIs for programmatic control of the entire system: clusters, nodes, tenants, buckets, keys, and policies, alongside the S3 data API.
Client SDKsNative Rust SDK and RESTful HTTP; more languages planned.
Written in RustMemory-safe, async-enabled, and built on SPDK for performance.

Platforms & Multi-Tenancy

FeatureDescription
Multi-platformx86_64, ARM, and Power; Docker for evaluation only.
Multi-tenantIAM with JWT and OAuth 2.0 / OIDC login compatibility.
Global namespaceCentralized namespace management for accounts and buckets.
Cluster peeringJoin clusters globally via namespace peering.
Air-gap modePull-only, outbound-only mode with no exposed listeners.
KMS integrationBuilt-in KMS with TPM 2.0 / YubiKey / customer-HSM vault custody and TLS certificate management.