HPE ATP Storage Solutions key terms

HPE's disaggregated, multi-protocol storage platform in which compute nodes and capacity enclosures are separate, connected by a high-speed fabric so capacity and performance scale independently.

HPE's high-end, mission-critical storage array family known for its always-on, self-healing architecture and application-aware quality-of-service controls, now succeeded by Alletra MP B-series.

HPE's midrange storage family built around adaptive flash and predictive analytics, later folded into the Alletra line; its dHCI variant combined storage with compute management.

HPE's entry-level, affordable SAN array family aimed at small and mid-sized businesses, offering dual controllers, wide drive support, and simple web-based setup.

HPE's purpose-built backup appliance family that deduplicates data at the target, supports replication between sites, and integrates with backup software through Catalyst and CIFS/NFS interfaces.

HPE's tape storage portfolio, including LTO tape drives, autoloaders, and libraries used for long-term retention, offline archives, and air-gapped protection against ransomware.

HPE's as-a-service consumption model in which infrastructure is delivered and managed on a pay-per-use or subscription basis, often with a cloud operations portal for metering and lifecycle management.

The cloud-hosted management portal used to view, subscribe to, and operate HPE GreenLake services, including storage health, capacity, and warranty information across registered devices.

HPE's cloud-based predictive analytics engine that collects telemetry from arrays, detects anomalies, recommends fixes, and provides cross-customer benchmarking to prevent performance and availability issues.

A design where compute, storage capacity, and networking are separate resource pools that can be scaled or replaced independently, rather than fixed together in one chassis.

An approach where additional nodes or enclosures are added to increase capacity and performance linearly, distributing data and workload across the growing set of resources.

An approach where a single system is made more powerful by adding controllers, drives, or memory within the same architecture, rather than adding independent nodes.

A protocol that carries the low-latency NVMe command set over a network fabric such as Fibre Channel or RDMA-capable Ethernet, letting hosts access remote flash with local-like latency.

A host-to-storage protocol designed for non-volatile memory, using deep parallel queues and low overhead to deliver far higher throughput and lower latency than legacy SAS or SATA interfaces.

A serial SCSI interface used to connect drives and enclosures inside a storage array, supporting dual paths for redundancy and high per-drive throughput.

A lossless, high-speed network standard used primarily for block storage SANs, with zoning and fabric services that isolate traffic and provide deterministic delivery.

A protocol that transports SCSI block commands over standard TCP/IP networks, allowing hosts to access SAN storage using existing Ethernet infrastructure instead of dedicated Fibre Channel hardware.

A Fibre Channel fabric control that restricts which host initiators and array target ports can communicate, reducing interference and preventing unauthorized access to LUNs.

A logical unit number identifying a block storage volume presented to a host, which the operating system sees as a disk device after masking and path configuration.

The array-side configuration that maps a volume only to specified host initiators, so other hosts on the same fabric cannot see or write to that volume.

A method of combining multiple physical drives into one logical unit using striping and parity or mirroring, improving performance and protecting against drive failure.

A RAID level that stripes data with two independent parity blocks, allowing the array to survive two simultaneous drive failures while still serving reads and writes.

A drive kept idle and ready in an array so it can automatically replace a failed member of a RAID set and begin rebuilding without administrator action.

The process of reconstructing lost data onto a replacement drive from parity or mirror copies, during which performance may degrade and the array is more vulnerable.

A capacity technique where volumes are created larger than the physical space initially allocated and consume storage only as data is actually written.

A capacity technique where the full size of a volume is reserved on physical media at creation, guaranteeing space but reducing flexibility and usable capacity.

The process of estimating current and future storage needs from workload growth, retention rules, replication copies, and headroom targets so the design avoids running out of space.

Input/output operations per second, a measure of how many read or write requests a storage system can complete in a second, heavily influenced by access size and randomness.

The time a single storage request takes to complete, usually measured in milliseconds or microseconds, and the metric users feel most directly in interactive applications.

The amount of data transferred per unit of time, typically in megabytes or gigabytes per second, which matters most for large sequential reads and writes.

Host software that presents multiple physical paths to the same volume as one device, providing failover if a path fails and load balancing across active paths.

A stretched-cluster feature that keeps two arrays synchronized so a host can transparently fail over to the remote array with the same LUN identity if the primary site fails.

Copying data from one storage system to another, either synchronously for zero data loss or asynchronously for distance and performance flexibility, to support disaster recovery.

A point-in-time logical copy of a volume that consumes little space initially and is used for fast rollback, testing, or as a source for backups.

A technique that stores only one unique copy of repeated data blocks and replaces duplicates with references, greatly reducing backup and archive capacity needs.

An HPE StoreOnce interface that lets supported backup applications send data directly to the appliance with client-side deduplication awareness and store-level status reporting.

A physical or logical separation that keeps backup copies offline or unreachable from production networks, protecting them from ransomware and destructive attacks.

Linear Tape-Open, a widely used open tape standard with generations that increase capacity and speed, commonly used for long-term retention and offline archives.

A delivery model where storage capacity and management are consumed on subscription or metered terms instead of purchased outright, shifting spending from capital to operating budget.

The automated gathering of configuration, performance, and health data from devices and sending it to a cloud analytics service for monitoring and predictive support.