Plain-English IT glossary.

These terms describe how devices connect and how traffic moves between them. Use them to discuss coverage, capacity and access boundaries before choosing equipment.

Related service: Networking Solutions.

VLAN

A VLAN is a virtual local area network that separates a switched network into distinct logical broadcast domains.

It helps keep guest devices, staff computers and cameras in separate groups on shared switches. Communication between groups still needs routing and access rules; creating VLANs alone does not define who can reach sensitive systems.

SD-WAN

SD-WAN is software-defined wide area networking, an approach that centrally manages traffic policies across connections between sites.

It can steer business applications over different internet or private links according to measured conditions. Ask which failures the design handles and how traffic behaves when every available connection is congested or unavailable.

PoE

PoE is Power over Ethernet, a method of delivering electrical power and data over compatible Ethernet cabling.

It can power access points, desk phones and IP cameras without separate sockets beside each device. Check both the power required by each device and the total power budget available from the switch.

Cat6A

Cat6A is an augmented Category 6 copper cabling specification that supports 10 Gigabit Ethernet over channels up to 100 meters when correctly installed.

It matters when planning cable routes that should support faster links later. Cable, connectors and installation quality all affect performance, so request channel test results rather than relying on the label printed on the cable.

Single-mode fiber

Single-mode fiber is optical fiber designed to carry light in one propagation mode, supporting long-distance data links with suitable transceivers.

It is useful between buildings or across longer site routes where copper distance limits become a problem. Match the optics, connectors and fiber specification, and include spares and access for repairs in the design.

Wi-Fi heat map

A Wi-Fi heat map is a floor-plan visualization of wireless measurements or predicted coverage, such as signal strength.

It helps identify weak areas and guide access point placement. A prediction is different from an on-site survey, and strong signal alone does not prove adequate capacity when meeting rooms fill with active users.

Firewall

A firewall is a system that allows or blocks network traffic according to configured security rules.

It controls which connections can enter, leave or move between network segments. Buyers should ask who maintains the rules, how exceptions are approved and which logs will be available when investigating a failed connection.

NAT

NAT is Network Address Translation, a process that changes IP address information as traffic passes through a network device.

It commonly lets many private devices share a public internet address. It can complicate inbound connections and voice troubleshooting, and it should not be treated as a replacement for an explicit firewall policy.

BGP

BGP is Border Gateway Protocol, a routing protocol used to exchange network reachability information between autonomous systems and within some large networks.

It matters when discussing multiple internet providers or complex routing policies. Having BGP does not by itself guarantee failover; address ownership, provider arrangements and route filtering must also fit the proposed design.

These terms explain hosting boundaries, storage and recovery. Compare them against the applications you run and the interruption your business can tolerate.

Related service: AWS Cloud Services.

Landing zone

A landing zone is a prepared cloud foundation with agreed account structure, identity controls, networking and governance for deploying workloads.

It gives teams a consistent starting point before applications arrive. Ask which controls are actually configured, who owns changes and how logging and cost allocation will work as additional environments are created.

IAM

IAM is Identity and Access Management, the policies and mechanisms that determine who or what can access resources and perform actions.

It covers people and application identities, not only login screens. Review permissions by job and workload, reduce unnecessary access and define what happens when an employee leaves or an application credential is exposed.

Region

A region is a geographic area in a cloud provider’s infrastructure where services and resources can be deployed.

Region choice affects latency, service availability and where data is stored. Check the location of backups and connected services too, because choosing one primary region does not automatically keep every copy of data there.

Availability zone

An availability zone is an isolated infrastructure location within a cloud region, designed to limit the effect of failures in other zones.

Using multiple zones can reduce dependence on one location, but applications must support that arrangement. Check database replication, network paths and failover behavior rather than assuming that extra virtual machines provide application resilience.

RTO

RTO is Recovery Time Objective, the target elapsed time for restoring a system or business process after disruption.

It turns recovery discussions into a requirement that can be tested. Include the time needed to obtain access, restore dependencies and confirm normal operation; copying data back is only part of getting staff working again.

RPO

RPO is Recovery Point Objective, the target maximum age of data that a business can accept losing after disruption.

It guides backup and replication frequency. A daily backup may leave almost a day of changes at risk, so compare the target with the actual recovery points available and test whether those copies can be restored.

3-2-1 backup

3-2-1 backup is a guideline to keep three copies of data, on two different storage media, with one copy off-site.

It reduces dependence on one device or location. The count alone does not prove protection against ransomware, so consider isolated or immutable copies, separate administrative access and restore tests alongside the storage arrangement.

Hypervisor

A hypervisor is software that creates and runs virtual machines while allocating the underlying computer’s hardware resources between them.

It lets several server workloads share physical hardware. That also creates a shared point of failure unless the design includes suitable recovery or clustering, and capacity planning must account for memory and storage as well as processors. See On-Premise Servers & Data Centers.

NAS vs SAN

NAS vs SAN compares network-attached storage, which typically provides shared files, with a storage area network, which provides block storage to hosts.

The distinction affects how applications access and manage data. Choose from workload needs, administration skills and recovery requirements; neither approach is automatically faster or safer simply because of its name. See On-Premise Servers & Data Centers.

RAID

RAID is Redundant Array of Independent Disks, a set of techniques that combines drives for performance, fault tolerance or both, depending on the level.

Some levels tolerate drive failure, while RAID 0 provides no redundancy. RAID does not protect against accidental deletion or ransomware, so a server still needs independent backups and a tested replacement procedure. See On-Premise Servers & Data Centers.

These terms describe controls and assessments that reduce exposure. Ask what each control covers, what evidence it produces and who acts when it identifies a problem.

Related service: Cybersecurity Services.

MFA

MFA is Multi-Factor Authentication, a login method that requires evidence from at least two distinct categories of authentication factors.

A password plus another password is not two factors. MFA reduces the damage from stolen passwords, but methods differ in phishing resistance, and recovery procedures need protection so attackers cannot bypass the normal login checks.

EDR

EDR is Endpoint Detection and Response, technology that records endpoint activity and helps detect, investigate and respond to suspicious behavior.

It can help isolate an affected laptop and trace what happened. Its value depends on coverage and response ownership; alerts that nobody reviews will not resolve an incident or restore an affected business application.

SPF

SPF is Sender Policy Framework, an email authentication mechanism that publishes which servers may send mail for a domain used in the SMTP envelope.

It helps receiving systems check the sending source. It does not authenticate the visible From address by itself, and forwarding can affect results, so review it together with DKIM and DMARC when configuring business email.

DKIM

DKIM is DomainKeys Identified Mail, an email authentication mechanism that uses a digital signature to associate a message with a signing domain.

Receivers can check that signed content has not changed and validate the signature using a public key. Configure every legitimate sending service and manage signing keys; a valid signature alone does not prove that a message is harmless.

DMARC

DMARC is Domain-based Message Authentication, Reporting and Conformance, a mechanism that checks alignment of the visible From domain with authenticated SPF or DKIM domains.

It lets domain owners publish handling policies and request reports. Inventory legitimate senders before enforcing rejection, because an incomplete setup can cause genuine invoices, support messages or application notifications to fail authentication.

Zero trust

Zero trust is a security approach that grants access through explicit verification and policy rather than implicit trust based on network location.

It encourages limited access and decisions informed by identity, device and resource context. Buying a product with this label is insufficient; map actual workflows and decide how access is reviewed and withdrawn over time.

Vulnerability scan

A vulnerability scan is an automated assessment that checks systems for known weaknesses, exposed services or insecure configurations.

It provides findings for investigation and remediation, but results can include false positives and miss issues outside its scope. Ask whether credentials are used, which assets are covered and how fixes will be verified.

Penetration test

A penetration test is an authorized security assessment that attempts to exploit weaknesses within an agreed scope to demonstrate their impact.

It can reveal attack paths that a scan alone does not explain. Agree boundaries and handling procedures beforehand, then require actionable findings and retesting; a completed test is evidence about that scope and time, not permanent safety.

Ransomware

Ransomware is malicious software that blocks access to systems or data, commonly through encryption, and demands payment.

Attacks may also involve stolen data and threats to publish it. Planning should cover identity protection, containment and independent recovery copies, with clear responsibility for decisions when normal email or business systems are unavailable.

These terms help separate application requirements from implementation choices. Define the user task and acceptance criteria before choosing a development approach or model.

Related service: AI Solutions.

API

An API is an Application Programming Interface, a defined way for one software component to request data or actions from another.

It allows systems such as payroll and attendance software to exchange information. Ask about authentication, limits and error handling, because an available API does not automatically make an integration complete or reliable. See Custom Software Development.

MVP

An MVP is a Minimum Viable Product, the smallest usable product that can test a specific business assumption with real users.

It should focus on a useful workflow and a clear learning goal. Agree what evidence will justify further development while retaining essential security and data handling; reducing scope should not mean accepting unusable or unsafe behavior. See Custom Software Development.

Cross-platform app

A cross-platform app is an application built to run on multiple operating systems while sharing some or most of its code.

Shared code can reduce duplicated work, but platform differences still need testing. Check device features, offline behavior and accessibility on each target platform before assuming that one build will behave identically everywhere. See Mobile App Development.

RAG

RAG is Retrieval-Augmented Generation, an approach that retrieves relevant information and supplies it as context to a generative model.

It can ground answers in company documents without retraining the model for every update. Retrieval quality, document permissions and answer evaluation still matter; supplying a source does not guarantee that the generated answer uses it correctly.

LLM

An LLM is a Large Language Model, a machine learning model trained on large amounts of text to model and generate language.

It can draft, classify or extract information, but fluent output is not proof of factual accuracy. Test representative tasks and exceptions, and decide which outputs require human review before they affect customers or business records.

OCR

OCR is Optical Character Recognition, technology that converts text in scanned documents or images into machine-readable text.

It helps move information from paper into searchable records or application workflows. Accuracy depends on image quality and layout, so verify critical fields such as invoice totals and employee identifiers rather than accepting every extracted character.

Fine-tuning

Fine-tuning is additional training of an existing model on selected examples to adjust its behavior for a task or domain.

It may help with consistent formats or specialized behavior when suitable examples exist. It is not a substitute for retrieving changing facts, and the training data and separate evaluation set need careful review before deployment.

Hallucination

A hallucination is model-generated content that is incorrect, unsupported or inconsistent with the information the model should use.

It matters when an answer sounds convincing enough to be accepted without checking. Use source checks and task-specific evaluation, and let the application decline uncertain requests or route them to a person when consequences justify review.

These terms describe recording equipment and video analytics. Start with what operators need to see, retain and investigate, then test performance under actual site conditions.

Related service: CCTV & AI Video Surveillance.

NVR

An NVR is a Network Video Recorder, a device or system that receives and stores video streams from network cameras.

It centralizes recording, playback and export. Check supported camera streams, recording bandwidth and usable storage, then verify what happens if a disk fails or the connection between a camera and the recorder is interrupted.

VMS

A VMS is a Video Management System, software used to manage cameras, recordings and operator access across a surveillance installation.

It matters when several operators or sites need coordinated monitoring. Compare camera compatibility, permissions and export workflows, and clarify licensing before assuming that adding cameras or analytics requires only extra storage.

ANPR

ANPR is Automatic Number Plate Recognition, technology that detects vehicle plates in images and converts their characters into searchable text.

It can support vehicle searches or access workflows, but accuracy depends on plate format, speed and lighting. Test the actual entry lanes and retain a review path for ambiguous readings before connecting recognition to an automatic action.

IP camera

An IP camera is a digital camera that sends video over an Internet Protocol network.

It can use the same network technologies as other business devices, which makes cabling and access design important. Keep administration restricted and confirm recording compatibility; a camera being network-connected does not require exposing it directly to the internet.

Bitrate

Bitrate is the amount of data transmitted or stored per second, commonly measured in bits per second.

For CCTV, it affects network demand and recording storage. Higher resolution does not determine bitrate alone: motion, frame rate and compression settings matter, so estimate retention from realistic scenes and confirm the image detail remains useful.

H.265

H.265 is a video compression standard, also called High Efficiency Video Coding, that encodes video to reduce the data needed for storage or transmission.

It can reduce storage demand compared with older encoding at comparable visual quality, depending on content and settings. Check recorder and playback compatibility, and test exported footage on the systems investigators will actually use.

Line crossing

Line crossing is a video analytics rule that triggers an event when a detected object crosses a configured virtual line in the camera view.

It can flag movement through a doorway or restricted boundary. Set direction and object filters carefully, then test shadows, crowds and changing light so operators understand likely false alerts and missed detections.

PTZ

PTZ means Pan, Tilt and Zoom, camera functions that change viewing direction and magnification remotely or through configured movement.

It lets operators examine different areas from one camera position. A camera looking elsewhere may miss an event, so assess whether fixed cameras are also needed for continuous coverage of entrances or other critical views.

These terms connect telephony and staffing systems with daily operations. Test the customer journey and employee workflows, including exceptions, before accepting the configuration.

Related service: Call Center & BPO Setup.

ACD

ACD is Automatic Call Distribution, a telephony function that routes incoming calls to agents or queues according to configured rules.

It can match callers to a team or skill group and handle busy periods through queues. Agree overflow and unavailable-agent behavior so callers have a defined path when the preferred team cannot answer.

IVR

IVR is Interactive Voice Response, a telephone system that accepts keypad or spoken input to guide callers through menus or automated tasks.

It can collect a caller’s intent before routing the call. Keep menus short and test invalid input, silence and transfer failures, because a correct technical configuration can still leave callers unable to reach the right team.

Predictive dialer

A predictive dialer is an outbound calling system that estimates agent availability and starts calls ahead of agents becoming free.

It can reduce agent waiting time, but inaccurate pacing can leave answered calls without an available agent. Review dialing controls, consent requirements and applicable calling rules for the destinations involved before enabling automated campaigns.

SIP trunk

A SIP trunk is a service that connects a business phone system to a telephony provider using Session Initiation Protocol signaling over an IP network.

It carries calling service without requiring a separate traditional circuit for each line. Confirm concurrent call capacity, number arrangements and failover, and test audio paths as well as successful call setup.

AHT

AHT is Average Handle Time, the average time spent handling a contact, commonly including conversation, hold time and after-contact work.

It helps estimate staffing and compare workflows when measurement rules are consistent. Using it as the only performance target can encourage rushed calls, so review resolution and service quality alongside the time recorded.

WFM

WFM is Workforce Management, the planning and scheduling of staff to match expected workload and operating requirements.

In a call center, it connects demand forecasts with shifts, skills and breaks. A useful plan accounts for absence and non-call activities, and supervisors need a process for adjusting schedules when actual demand differs from the forecast.

HRMS

An HRMS is a Human Resource Management System, software that manages employee records and HR processes such as leave, attendance and payroll-related workflows.

It gives HR teams a controlled place to maintain employee information. Confirm approval rules, access permissions and reporting needs, and establish which system owns each field before importing data or connecting attendance devices. See HRMS Solutions.

Payroll parallel run

A payroll parallel run is a validation exercise that calculates payroll in an existing process and a proposed system for the same period and compares the results.

It exposes differences in rules or imported data before the new system becomes authoritative. Investigate exceptions employee by employee, document approved corrections and agree acceptance with the people responsible for payroll. See HRMS Solutions.