DMCP flashcards

32 cards. Read the term and the definition here. The deck at the top of this page is optional.

RAID 5 usable capacity

Usable capacity equals the number of data drives times the smallest drive size, i.e., (N−1) × disk size. One drive's worth of capacity is consumed by distributed parity, and the array survives one drive failure.

RPO versus RTO

Recovery Point Objective is the maximum tolerable data loss measured in time; Recovery Time Objective is the maximum tolerable service downtime. Backup frequency is set by RPO, restore method by RTO.

Differential versus incremental backup

A differential backup captures all changes since the last full backup, so restore needs only the full plus the latest differential. An incremental captures changes since the last backup of any type, so restore needs the full plus every incremental since.

Snapshot limitations

Snapshots are point-in-time metadata copies, usually copy-on-write, that are fast and space-efficient. They typically live on the same array as the source, so they do not protect against array loss or site failure.

3-2-1 backup rule

Keep at least three copies of data, on two different media types, with one copy stored offsite. This protects against media failure, local disasters, and site-wide outages; immutable or air-gapped copies add ransomware resilience.

Rack unit height

One rack unit (1U) equals 44.45 millimeters, or 1.75 inches, of vertical space in a standard 19-inch rack. Server heights are quoted in U, so a 2U server occupies 88.9 millimeters.

NUMA architecture

Non-Uniform Memory Access means each CPU socket has local memory that it accesses faster than memory attached to another socket. Latency-sensitive workloads should be pinned within a single NUMA node to avoid slower remote accesses.

ECC memory

Error-Correcting Code memory detects and corrects single-bit memory errors and detects many multi-bit errors. It is standard in servers because uncorrected memory errors can cause crashes or silent data corruption.

Out-of-band management controller

A baseboard management controller such as iDRAC, iLO, or XCC provides power control, remote console, and hardware health monitoring independent of the operating system. It requires its own network connection and credentials.

PSU redundancy configurations

Redundant power supplies are described as 1+1 (one spare) or 2+2 (two spares). Each path should connect to a different power distribution unit or feed so a single PDU or circuit failure cannot take down the server.

Usable hosts per subnet

Usable host addresses equal 2^h − 2, where h is the number of host bits. For example, a /26 leaves 6 host bits, giving 64 addresses minus network and broadcast, so 62 usable hosts.

VLAN purpose

A VLAN divides a physical switch into separate Layer 2 broadcast domains. Traffic between VLANs must be routed by a Layer 3 device or inspected by a firewall, which improves segmentation and security.

Link aggregation

Link aggregation, standardized as 802.3ad with LACP, bundles multiple physical links into one logical link for increased bandwidth and redundancy. Member links must match in speed and duplex and be configured on both ends.

iSCSI port number

iSCSI uses TCP port 3260 to carry SCSI commands over an IP network. It is a block storage protocol commonly deployed on 10, 25, or 100 Gigabit Ethernet with dedicated VLANs or separate physical networks.

TCP versus UDP

TCP is connection-oriented and guarantees ordered, retransmitted delivery, suiting web, file transfer, and iSCSI. UDP is connectionless with no delivery guarantee, suiting DNS, DHCP, VoIP, and streaming where low latency matters more than reliability.

Type 1 hypervisor

A Type 1 hypervisor runs directly on server hardware without an underlying host operating system. Examples include VMware ESXi, Microsoft Hyper-V, KVM, and Xen; production virtualization platforms use this type.

VM resource shares

Shares determine relative priority when a resource such as CPU or memory is contended. Reservations guarantee a minimum, limits cap a maximum, and shares only matter under contention, so they do not reserve capacity.

Cloud service models

IaaS provides virtual machines, storage, and networking; PaaS provides a managed runtime or database; SaaS provides a finished application. The provider's patching responsibility increases as you move from IaaS toward SaaS.

Live migration requirements

Live migration moves a running virtual machine between hosts with minimal downtime. It requires shared storage or concurrent storage migration, compatible CPU features across hosts, and sufficient network bandwidth for memory transfer.

Containers versus virtual machines

Containers share the host operating system kernel, making them lightweight with fast startup. Virtual machines include a full guest operating system, providing stronger isolation but requiring separate patching and more resources.

Idempotent automation

An idempotent automation task produces the same end state no matter how many times it runs. Configuration management playbooks and manifests should be idempotent so repeated executions do not create duplicate or conflicting changes.

Agentless configuration management

Ansible is agentless: it connects over SSH or WinRM and pushes modules to managed hosts without installing persistent software. Puppet and Chef typically rely on agents installed on each managed node.

ITIL incident versus problem

An incident is an unplanned interruption or degradation that must be restored quickly. A problem is the underlying cause of one or more incidents, investigated to prevent recurrence through root-cause analysis and permanent fixes.

Infrastructure as Code state

Declarative IaC tools such as Terraform record provisioned resources in a state file. The state must be stored securely with locking enabled so concurrent runs cannot corrupt it or create duplicate infrastructure.

CIA triad

Confidentiality protects data from unauthorized disclosure, integrity protects it from unauthorized modification, and availability ensures authorized access when needed. Security controls are classified as preventive, detective, or corrective against these goals.

Multi-factor authentication

MFA requires two or more different factor types: something you know, something you have, or something you are. Two passwords or two biometric checks do not qualify because they use the same factor category.

RBAC versus ABAC

Role-based access control grants permissions through assigned roles. Attribute-based access control evaluates attributes such as department, time, or device posture, enabling finer-grained and context-aware authorization decisions.

NIST SP 800-88

NIST Special Publication 800-88 provides guidelines for media sanitization, defining clear, purge, and destroy actions. It is used to verify that data is unrecoverable before equipment is reused, resold, or disposed of.

Availability of serial components

For components in series, overall availability is the product of individual availabilities. Two serial components at 0.99 each yield 0.9801, so adding dependencies reduces end-to-end availability unless redundancy is introduced.

Functional versus non-functional requirements

Functional requirements describe what a solution must do, such as serving files or running a database. Non-functional requirements describe qualities such as performance, availability, security, and cost, and should be stated as measurable targets.

Total cost of ownership

TCO sums acquisition, licensing, power and cooling, rack space, support contracts, staff time, and migration or decommissioning costs over the solution's life. Purchase price alone understates the real cost of ownership.

Layered troubleshooting method

Troubleshooting proceeds by defining the problem, gathering evidence, forming a hypothesis, and testing the simplest or most likely cause first. Isolating by OSI layer or component narrows the fault before a fix is applied and verified.