NetApp ONTAP glossary

A

aggregate

The physical storage pool: one or more RAID groups of disks (or an all-flash pool) formatted as a single WAFL filesystem. Aggregates own the raw space; volumes are created inside them. See WAFL deep-dive.

AFF

All-Flash FAS — NetApp's all-NVMe/all-SSD platform line. Same ONTAP, no spinning disks, designed for low latency and high IOPS.

ASA

AFF SAN Array — an AFF system locked into SAN-only mode (no NAS protocols, no SVMs in the usual sense). Simplifies FC/NVMe-over-FC deployments; storage is presented as LUNs only.

AUSO

Automatic Unplanned Switchover — the MetroCluster feature that lets the surviving site take over automatically after a site failure, with the mediator's vote.

AutoSupport

ONTAP's built-in health telemetry: periodic and event-triggered messages sent to NetApp (or your own mail server) so support can see the system's state before you open a case.

adaptive QoS

A QoS mode where performance ceilings scale automatically with storage size, rather than being fixed numbers you set once.

B

broadcast domain

The grouping of network ports that share an L2 broadcast domain; ONTAP uses these to place LIFs, manage failover groups, and allocate IPspace subnets.

break (SnapMirror)

Stopping replication so the destination can be written to independently — the first step of an unplanned failover. Re-establish with a resync afterwards.

C

cluster

The whole ONTAP system: two or more nodes (up to 24 in a cluster) that share management, SVMs, and a unified namespace.

cluster peering

A trust relationship between two clusters that enables cross-cluster features: SnapMirror, SVM DR, FlexCache, MetroCluster.

CIFS

The older name for SMB file sharing, still used in ONTAP for SMB server objects (vserver cifs). SMB 3.x is the modern protocol version.

consistency point (CP)

The moment WAFL atomically flips its root pointer, committing batched writes to disk. The filesystem is always consistent because states change only at CPs.

constituent

One of the FlexVols that make up a FlexGroup. Files are distributed across constituents; each lives on its own aggregate.

copy-on-write (CoW)

WAFL's rule that changed blocks are written to new locations rather than overwritten in place — what makes snapshots and clones nearly free. See the deep-dive.

cloud tiering (FabricPool)

Moving cold blocks of a volume to object storage (S3, Azure Blob, Google Cloud) while hot blocks stay on local flash. The volume still looks and behaves like one filesystem.

compression

Storage efficiency: compressing data blocks to reduce physical footprint. ONTAP does inline (at write time) and post-process (background) compression.

D

deduplication

Storage efficiency: replacing duplicate blocks with pointers to a single shared block. Most effective for virtual machine disks and copies of the same data.

destination volume

The replica side of a SnapMirror relationship — read-only during normal replication, writable once you break the relationship.

disk ownership

Which node in the HA pair owns a given disk. Disks can be moved between partners; ownership is what lets takeover/giveback work.

data protection (DP)

The volume type used for SnapMirror/SnapVault destinations. DP volumes exist only to hold replicas and aren't directly mounted.

disaster recovery

The set of processes and technologies (SnapMirror, MetroCluster, SVM DR) that keep data available after a site or system loss. See MetroCluster.

disk shelf

The physical enclosure holding disks, attached to nodes via SAS or NVMe cabling. Shelves provide power, cooling, and the disk-to-node connections.

E

EMS

Event Management System — ONTAP's event bus. Everything notable becomes an EMS event you can see with event log show, filter, and alert on.

export policy

The NFS access-control list attached to volumes and qtrees: rules saying which clients can mount with which permissions. See the NFS guide.

storage efficiency

The umbrella term for deduplication, compression, compaction, and thin provisioning — all the features that make data take less physical space.

encryption (NVE/NVMe)

NetApp Volume Encryption (NVE) encrypts each volume with its own key; NetApp Aggregate Encryption (NAE) encrypts at the aggregate level. Both are software-based, keyed via ONTAP's key management.

F

FAS

NetApp's hybrid platform line — Flash + spinning disk (or all-flash configurations). The general-purpose ONTAP systems. AFF is the all-flash member of the family.

FSx for ONTAP

Amazon's managed ONTAP service: real ONTAP HA pairs run by AWS inside your VPC, with the same SVMs, volumes, snapshots and SnapMirror as on-prem. See the FSx guide.

FabricPool

NetApp's cloud tiering feature — see cloud tiering.

storage failover (SFO)

The HA mechanism where a partner node takes over a failed node's storage (takeover) and later returns it (giveback). Lossless thanks to the NVRAM mirror.

FlexCache

A read cache volume that mirrors a subset of a source volume onto other nodes, so remote sites serve their own reads instead of hitting the origin.

FlexClone

An instant, writable copy of a volume (or file/LUN) that shares blocks with its parent snapshot until it writes its own. See the deep-dive.

FlexGroup

ONTAP's scale-out NAS volume: one namespace striped across many constituent FlexVols on different nodes. See the FlexGroup deep-dive.

FlexVol

The standard ONTAP volume: a flexible, resizable logical container in an aggregate, with its own snapshots, quota rules, and settings.

FPolicy

An external policy engine hook: ONTAP sends file operations to a partner server (e.g. for anti-virus or monitoring) and can block or allow them.

Fibre Channel (FC)

The classic SAN transport. ONTAP uses FC for LUNs and NVMe-over-FC, and FC (or IP) between MetroCluster sites.

Flash Cache

A read cache card in a hybrid system that accelerates reads of hot data without moving it to flash permanently.

G

giveback

Returning storage from the takeover partner to the original node after maintenance or a failure — the second half of storage failover.

space guarantee

A volume's promise of capacity: volume (reserve the full size), none (thin — reserve nothing), or snapshot (reserve data, not snapshot space).

H

HA pair

Two nodes connected by an HA interconnect, each able to take over the other's storage with no data loss. The fundamental high-availability unit of ONTAP.

hybrid aggregate

An aggregate mixing flash and spinning disk, letting ONTAP tier hot data to flash automatically within the aggregate.

I

inode

The on-disk structure describing a file or directory: its blocks, ownership, permissions, timestamps. Running out of inodes means running out of files, even with free capacity.

iSCSI

SAN protocol carrying SCSI over IP networks. ONTAP serves LUNs over iSCSI; needs no special host hardware.

HA interconnect

The dedicated link(s) between HA pair nodes carrying NVRAM mirror traffic and heartbeat — the backbone of takeover.

IPspace

A fully isolated network domain within a cluster, letting separate tenants share the hardware with completely separate LIFs, routing, and broadcast domains.

J

junction path

The mount point of a volume in an SVM's namespace, e.g. /data/projects. The junction is what makes volumes visible to NAS clients.

K

Kerberos

The authentication protocol used for secure NFS (krb5/krb5i/krb5p) and SMB. ONTAP authenticates users against Active Directory/KDC via its Kerberos service.

key management

Where encryption keys live: ONTAP's onboard key manager (OKM) or an external KMIP server. Lose the keys, lose the data.

L

LIF (logical interface)

A network address (IP + port) that clients connect to. LIFs are movable between ports and nodes — the trick that makes failover and load balancing invisible to clients.

LUN

A block device carved from a volume and presented over FC/iSCSI/NVMe. LUNs are what SAN hosts (VMware, Windows, Linux) see as disks.

LDAP

A directory protocol ONTAP can use (with or instead of NIS/AD) to resolve Unix users and groups for NFS permissions.

latency

Response time of an I/O operation. ONTAP keeps client-visible latency low by acknowledging writes from NVRAM and batching disk flushes.

M

MetroCluster

NetApp's stretch-cluster DR solution: two sites, synchronous mirroring, automatic failover. See the MetroCluster guide.

mediator

A small service (in MetroCluster IP configurations) that votes during split-brain situations so the surviving site can safely take over.

SnapMirror

ONTAP's replication engine: asynchronous or synchronous copies of volumes/consistency groups to another location, using snapshots as the transfer unit.

SnapVault

Disk-to-disk backup flavor of SnapMirror: keeps many point-in-time snapshots on the destination with different retention than the source.

N

namespace

The tree of junction paths clients see from an SVM root — the client-visible view of all the SVM's volumes.

NFS

The Unix file-sharing protocol. ONTAP supports NFSv3, NFSv4.0/4.1 (with pNFS), and exports with granular export policies.

NVRAM/NVMEM

The non-volatile write log that makes acknowledged writes durable and takeover lossless. See the NVRAM deep-dive.

NVMe

The modern storage protocol/transport: NVMe over Fabrics (FC or TCP) for SAN, and the media type inside AFF systems.

node

One controller (hardware + ONTAP instance) in a cluster. Nodes come in HA pairs and share the cluster's SVMs and volumes.

NDMP

The protocol ONTAP uses for tape-based backup: dump/restore streams to NDMP-capable backup software, often from snapshots for consistency.

NIS

Legacy Unix directory service; ONTAP still supports it for name services, though AD/LDAP is the modern path.

O

ONTAP

NetApp's storage operating system, running on FAS/AFF systems, ONTAP Select (software), and in the cloud (FSx for ONTAP, Azure NetApp Files, Google Cloud NetApp Volumes).

ONTAP Select

ONTAP as software on your own commodity servers/hypervisors — for labs, edge, and software-defined storage.

object store server

ONTAP's built-in S3 service (9.8+): an SVM can present buckets to S3 clients. See the S3 guide.

P

parity

The redundancy math in RAID: parity blocks let the RAID group reconstruct data after a disk failure. RAID-DP uses double parity; RAID-TEC triple.

partition

A slice of a disk with defined ownership. Root-data partitioning splits a disk so each node of an HA pair gets a piece — the foundation of 2-node clusters.

snapshot policy

The named schedule + retention set that drives automatic snapshots on a volume (e.g. 12 hourly, 7 daily, 4 weekly).

QoS

Quality of Service: performance ceilings (and floors, with adaptive QoS) applied per volume or workload so noisy neighbors can't starve others.

performance service level

The adaptive-QoS wrapper (Extreme/Performance/Value) that scales IOPS limits with storage size.

Q

qtree

A sub-volume directory tree that can have its own quotas, export policies, and security style — a way to carve a volume into independently-managed pieces.

quota

A limit (space and/or file count) applied to users, groups, or qtrees. Quotas enforce capacity discipline per tenant.

R

RAID-DP

NetApp's double-parity RAID: survives any two disk failures in a RAID group, and is the default protection level for data aggregates.

RAID-TEC

Triple-parity RAID for very large or high-risk RAID groups: survives three concurrent disk failures.

RAID group

A set of disks (data + parity) that RAID protects as a unit. Aggregates contain one or more RAID groups.

root aggregate

The aggregate holding the node's operating system and root volumes — the system's boot storage, separate from data aggregates.

volume rebalance

Moving files between FlexGroup constituents to even out distribution. Background, throttled, online. See FlexGroup deep-dive.

RPO / RTO

Recovery Point Objective — how much data you can afford to lose; Recovery Time Objective — how fast you must be back. DR design targets, not ONTAP features.

snapshot reserve

The percentage of a volume (default 5%) set aside so snapshots don't crowd out live data.

S

SVM (vserver)

Storage Virtual Machine — the tenant/namespace unit of ONTAP: its own protocols, LIFs, volumes, and security. One cluster hosts many SVMs.

SVM DR

Disaster recovery for a whole SVM: replicate all its volumes to a peer cluster and fail over the namespace.

SMB

The Windows file-sharing protocol (formerly CIFS). ONTAP supports SMB 3.x with encryption, multichannel, and continuous availability. See the SMB guide.

S3

The object-storage API. ONTAP 9.8+ speaks S3 natively; see S3 on ONTAP.

snapshot

A point-in-time, read-only image of a volume, nearly free thanks to copy-on-write. See the deep-dive.

spare disks

Disks held in reserve so RAID can rebuild automatically when one fails — not counted in usable capacity.

SyncMirror

Mirroring at the aggregate level: every write goes to two disk pools simultaneously, used by MetroCluster for zero-RPO protection.

shelf

See disk shelf.

System Manager

ONTAP's built-in GUI — the browser-based management interface that mirrors most CLI functions.

SnapCenter

NetApp's application-aware backup/orchestration software (separate product) for databases and VMs using ONTAP snapshots.

T

takeover

One HA node assuming the failed partner's storage, replaying its NVRAM log and serving its volumes. See NVRAM deep-dive.

thin provisioning

Volumes with space-guarantee none: capacity is claimed virtually and consumed only as data is written. Watch oversubscription.

tiering

Moving blocks between performance tiers (flash ↔ spinning ↔ cloud) automatically based on access temperature.

transition

7-Mode to cDOT/ONTAP migration (now "ONTAP 9 conversion") — moving legacy 7-Mode systems to clustered ONTAP.

U

user ID mapping

Translating between Windows SIDs and Unix UIDs/GIDs so mixed NFS+SMB access sees consistent ownership. ONTAP does this via AD/LDAP/NIS lookup.

unified networking

Serving NAS, SAN, and replication over shared Ethernet ports with VLANs/IPspaces — ONTAP's default model on modern platforms.

V

volume

The logical container of files/LUNs inside an aggregate: has its own space guarantee, snapshots, quotas, and settings. FlexVol is the standard type.

vserver

See SVM.

Vscan

ONTAP's anti-virus framework: file writes are scanned by external AV servers connected over the cluster, blocking or quarantining on detection.

VLAN

Virtual LAN tagging used to split physical ports into multiple networks for different SVMs, IPspaces, or tenants.

W

WAFL

Write Anywhere File Layout — ONTAP's filesystem. Writes land anywhere, metadata is copy-on-write, and state changes atomically at consistency points. See the WAFL deep-dive.

WAFL reserve

The internal space ONTAP keeps for filesystem metadata. In ONTAP 9 it's managed automatically — don't try to reclaim it.

WWPN/WWNN

World Wide Port/Node Names — the unique identifiers of FC ports (and the NVMe equivalents) used for SAN zoning and host mapping.

write-anywhere

WAFL's allocation strategy: new data goes to any free block, enabling write batching, full-stripe RAID writes, and CoW snapshots.