DMCP key terms

A disk array configuration that stores two independent parity blocks across member drives, allowing the array to keep running through the failure of any two disks simultaneously.

A mirrored-and-striped layout that combines RAID 1 mirroring with RAID 0 striping, giving fast reads and fast rebuilds at the cost of losing half the raw capacity.

An idle drive kept in or attached to an enclosure that automatically takes over for a failed member disk, restoring redundancy without waiting for a technician to arrive.

The process of reconstructing data onto a replacement or spare drive after a disk failure, during which the array runs without full protection and performance may degrade.

A point-in-time logical copy of a volume or file system that preserves the state of data for rollback or backup without duplicating the entire dataset immediately.

Continuous or scheduled copying of data from one storage system to another, often at a remote site, so that a usable copy exists if the primary copy is lost.

Recovery point objective: the maximum amount of data, measured in time, that an organization is willing to lose if a disruption occurs before recovery completes.

Recovery time objective: the target duration within which a service must be restored after an outage, driving decisions about standby capacity and automation.

A SAN access control method that maps specific logical volumes to specific host initiators, preventing unauthorized servers from seeing or writing to those volumes.

World Wide Name: a unique 64-bit identifier burned into a Fibre Channel or SAS device, used by fabrics and arrays to identify initiators and targets.

Allocating storage capacity to volumes on demand rather than up front, so physical space is consumed only as data is actually written, requiring careful capacity monitoring.

A protocol that carries SCSI commands over TCP/IP, letting hosts use standard Ethernet infrastructure to reach block storage targets instead of a dedicated Fibre Channel fabric.

A Fibre Channel fabric configuration that groups ports into allowed communication sets, isolating traffic and limiting which initiators can reach which targets.

Ethernet frames larger than the standard 1500-byte payload, commonly 9000 bytes, which reduce per-packet overhead but must be configured consistently end to end.

A logical Layer 2 network segment created by tagging frames, allowing one physical switch fabric to carry isolated broadcast domains for different purposes or tenants.

Bonding multiple physical network links into one logical channel to increase throughput and provide failover, requiring matching configuration on both ends of the connection.

Designing host-to-storage connectivity with more than one independent physical route, so a failed adapter, cable, or switch does not sever access to data.

Host software that presents several physical paths to the same storage device as one logical device, providing load balancing and continued access when a path fails.

A standardized out-of-band management interface for servers that exposes power control, sensor readings, and event logs independently of the operating system.

Baseboard management controller: the dedicated chip on a server motherboard that provides remote console, power, and health monitoring even when the host is powered off.

Administering hardware through a separate management network rather than the production data path, so devices remain reachable when the operating system or network is down.

A processor and memory architecture in which each CPU socket has local memory with faster access than remote memory, influencing how workloads and virtual machines should be placed.

Memory that detects and corrects single-bit errors using extra parity bits, reducing the risk of silent data corruption in long-running server workloads.

Software that creates and runs virtual machines by abstracting physical CPU, memory, storage, and network resources and scheduling them among guests.

Assigning more virtual processors to virtual machines than there are physical cores, which raises consolidation density but can cause contention when workloads peak together.

Moving a running virtual machine between hosts with little or no interruption to its workload, typically used for maintenance, balancing, or evacuation of failing hardware.

A preconfigured, non-running virtual machine image used as the source for new deployments, ensuring consistent configuration and speeding up provisioning.

A group of hosts managed as a unit so that workloads can be placed, restarted, or migrated automatically when a member fails or is taken offline.

Defining servers, networks, and storage in version-controlled machine-readable files so environments can be provisioned repeatedly and reviewed like application code.

The property that running an automation task repeatedly produces the same end state, making reruns safe after partial failures or interrupted deployments.

Coordinating multiple automated tasks and systems in a defined sequence, handling dependencies, retries, and rollback across servers, storage, and network components.

An approved, documented set of firmware and driver versions applied across a hardware fleet so that known interoperability and stability issues are avoided.

A design that removes single points of failure through redundancy and automatic failover, targeting continuous service rather than merely fast recovery.

Automated checks that compare device and host configurations against a security benchmark or policy, producing findings that must be remediated or formally accepted.

Granting each account only the permissions required for its tasks, limiting the blast radius if credentials are stolen or a user makes a mistake.

An authorization model that assigns permissions to roles rather than individuals, so users inherit rights by role membership and access scales with the organization.

Protecting stored data by encrypting volumes or drives so that media removed from the array or stolen from a site cannot be read without the key.

The lifecycle handling of cryptographic keys, including generation, storage, rotation, escrow, and revocation, which determines whether encrypted data remains recoverable.

Calculating required compute, memory, storage, and bandwidth from measured workload demand plus headroom, so a design meets performance targets without excess cost.