150 Terms

The ERP & On-Prem AI Glossary

Straight definitions for the terminology that shows up in Infor SyteLine, Infor LN, Baan, manufacturing operations, defense compliance, and on-prem AI - written by consultants who work in these systems daily, not by a marketing team.

Infor SyteLine

AHL (Automatic Handling Logic)

AHL (Automatic Handling Logic) is the declarative rule layer in SyteLine and its Mongoose framework that automatically performs standard record behavior - defaulting values, validating input, and cascading updates - when data or form events fire, without a developer writing custom code.

Infor SyteLine

IDO (Intelligent Data Object)

An IDO (Intelligent Data Object) is the middle-tier business object in SyteLine and the Mongoose framework that exposes database tables as named properties, methods, and collections, so forms, mobile apps, and REST clients all read and write data through one governed layer.

Infor SyteLine

Mongoose Framework

Mongoose is Infor's metadata-driven application development framework that SyteLine and CloudSuite Industrial are built on. It renders forms, exposes data through IDOs, runs an event system, and lets customers extend or build applications without altering delivered source code.

Infor SyteLine

BOD (Business Object Document)

A BOD (Business Object Document) is the standardized XML message format Infor applications use to exchange data through ION. Each BOD pairs a verb such as Sync or Process with a noun such as ItemMaster, based on the open OAGIS integration standard.

Infor SyteLine

Infor ION

Infor ION is the integration backbone of Infor OS. It routes BOD messages between Infor applications such as SyteLine, LN, and M3 and third-party systems, and adds document flows, workflow, event alerts, and a monitored, replayable message trail.

Infor SyteLine

APS (Advanced Planning and Scheduling)

APS (Advanced Planning and Scheduling) is SyteLine's memory-resident planning engine that schedules production against finite material and capacity constraints, producing realistic operation start and finish dates and enabling capable-to-promise responses during order entry.

Infor SyteLine

DataView

A DataView in SyteLine is a user-configurable grid that reads data through an IDO, letting users choose columns, apply filters, group and total rows, and export to Excel - producing ad hoc analysis without a developer or a formal report.

Infor SyteLine

SyteLine Event System

The SyteLine Event System is the framework's rule engine that detects defined business events - a record saved, a status changed, a threshold crossed - and runs configured handlers that send notifications, request approvals, execute SQL, or call custom logic.

Infor SyteLine

Form Personalization

Form personalization in SyteLine is the ability to modify a form's appearance and behavior - hiding fields, relabeling, reordering, setting defaults - at user, group, or site level, stored as a layer above the delivered form so the base object is never altered.

Infor SyteLine

Item Master

The item master in SyteLine is the central record that defines every part the business buys, makes, stocks, or sells - carrying identity, units of measure, planning parameters, costing method, product code, and quality attributes that drive nearly every downstream transaction.

Infor SyteLine

Job Order

A job order in SyteLine is the production work order that authorises manufacturing a quantity of an item. It carries its own copy of the bill of material and routing, collects material, labor, and overhead costs, and moves through a defined status life cycle.

Infor SyteLine

Estimate Job

An estimate job in SyteLine is a non-production job used to cost a potential order. It carries materials and routing operations like a real job so SyteLine can calculate a full cost, but it creates no demand, consumes no inventory, and is ignored by planning.

Infor SyteLine

Current Operations

Current Operations in SyteLine is the shop-floor-facing view of job operations currently available or in process, filtered by work center or resource, from which operators report quantity complete, scrap, and labor time against the job.

Infor SyteLine

Work Center

A work center in SyteLine is the production location where a routing operation is performed. It carries the labor, machine, and overhead rates used to cost that operation and links to the resource group that APS uses for finite capacity scheduling.

Infor SyteLine

Resource Group

A resource group in SyteLine is a set of interchangeable machines, cells, or people that operations schedule against. It carries the capacity, calendar, and constraint data that APS uses to place work into finite time windows.

Infor SyteLine

Backflush

Backflush in SyteLine is the automatic deduction of material and posting of labor when an operation or job is reported complete. Instead of issuing each component manually, the system relieves inventory at standard quantity based on the quantity reported.

Infor SyteLine

Phantom BOM

A phantom BOM is a subassembly that exists in the bill of material for engineering clarity but is never stocked or built as its own order. Planning blows straight through it, passing demand directly to its components on the parent job.

Infor SyteLine

Product Code

A product code in SyteLine is a required grouping assigned to every item that determines which general ledger accounts inventory, cost of sales, and revenue transactions post to, and serves as the primary dimension for product-line reporting.

Infor SyteLine

SyteLine Replication

SyteLine replication is the built-in mechanism that copies data between sites and databases in a multi-site installation. Replication categories define what moves, replication rules define where it goes, and the framework queues and delivers changes between site databases.

Infor SyteLine

Multi-Site (SyteLine)

Multi-site in SyteLine is the architecture that lets one installation run several plants, warehouses, or legal entities as distinct sites - each with its own inventory, planning, and transactions - while sharing master data and consolidating financially.

Infor SyteLine

CloudSuite Industrial

CloudSuite Industrial (CSI) is the cloud-delivered edition of Infor SyteLine, hosted by Infor on AWS alongside the Infor OS platform. It uses the same Mongoose application but adds managed updates and restricts customization and direct database access.

Infor SyteLine

QCS (Quality Control System)

QCS (Quality Control System) is SyteLine's quality management module. It adds inspection plans and specifications, test result recording, nonconformance and corrective action tracking, and supplier quality control to the core ERP transaction flow.

Infor SyteLine

AES (Application Event System)

AES (Application Event System) is the Mongoose framework component in SyteLine that detects defined application events, evaluates conditions, and executes configured handler actions such as notifications, prompts, SQL statements, and IDO method calls.

Infor SyteLine

FormSync

FormSync in SyteLine is the Mongoose capability that keeps multiple open forms synchronized on the same record context, so moving to a different order or item in one form automatically refreshes the related forms to show data for that same record.

Infor SyteLine

IDO Collection

An IDO collection in SyteLine is the in-memory set of rows a LoadCollection call returns from an IDO. It holds the property values plus each row's state - unchanged, modified, inserted, or deleted - so the framework knows exactly what to persist on update.

Infor LN & Baan

Baan IV

Baan IV is the mid-1990s release of Baan Company's ERP suite, built on the bshell runtime with 3GL and 4GL development tools, package-based modules such as td, ti and tf, and multi-company table sharing.

Infor LN & Baan

BaanERP (Baan V)

BaanERP, also called Baan V, is the late-1990s successor to Baan IV. It introduced a separate Warehousing package, Enterprise Planning, and a component-based Windows architecture, and it is the direct code ancestor of SSA ERP LN and today's Infor LN.

Infor LN & Baan

Infor LN

Infor LN is Infor's enterprise ERP suite for complex discrete manufacturers, descended from Baan IV and Baan V. It is deployed as on-premises LN 10.x or as an Infor CloudSuite edition, and covers manufacturing, warehousing, projects, service and multi-company finance.

Infor LN & Baan

DEM (Dynamic Enterprise Modeling)

DEM (Dynamic Enterprise Modeling) is Baan and Infor LN's business process modeling layer. It captures an organization as enterprise structure, business function, business process and organization models, then generates role-based menus so each user sees only the sessions their process steps require.

Infor LN & Baan

Baan Session

A Baan session is the atomic executable unit of Baan and Infor LN - one screen, report or batch program - identified by a coded name such as tdsls4101m000, where the letters give the package and module and the digits give the session number.

Infor LN & Baan

Table Sharing

Table sharing is the Baan and Infor LN mechanism that lets several companies read and write the same physical database table, so master data such as items, business partners and currencies is maintained once and used by every company that shares it.

Infor LN & Baan

Logistic Company

A logistic company in Baan and Infor LN is a company number that holds operational data such as items, warehouses, and sales, purchase, production and warehouse orders. Each logistic company is linked to exactly one financial company that carries its ledger.

Infor LN & Baan

Financial Company

A financial company in Baan and Infor LN is a company number that owns the general ledger: the chart of accounts, dimensions, fiscal calendars, currency setup and all financial transactions for one or more logistic companies linked to it.

Infor LN & Baan

Baan 3GL and 4GL

Baan 3GL and 4GL are the two programming languages of Baan and Infor LN. 4GL is the event-driven script attached to a session and its form fields, while 3GL is the C-like language used for libraries, DAL objects and standalone programs.

Infor LN & Baan

Porting Set

A porting set is the platform-specific layer of Baan and Infor LN that adapts the application to a given operating system and database. It ships the bshell executable, the database drivers and system utilities, and is versioned and patched independently of the application software.

Infor LN & Baan

Bshell

The bshell is the Baan and Infor LN virtual machine. It loads compiled pseudo-code for sessions, scripts and reports, executes it, and translates the application's database calls into native SQL through porting set drivers - one bshell process typically serves one user session.

Infor LN & Baan

Exchange Scheme

An exchange scheme is the reusable definition in the Baan and Infor LN Exchange module that maps external flat-file records to LN tables field by field, in either direction, so data can be imported or exported in batches without writing custom programs.

Infor LN & Baan

Infor LN Studio

Infor LN Studio is the development environment for Infor LN. It gives developers an Eclipse-based workspace to write and debug 3GL and 4GL code, design sessions, forms and reports, and manage components and VRC software layers against a live LN development environment.

Infor LN & Baan

LN Extensions

LN Extensions are Infor LN's supported customization framework: extension scripts and objects attached to standard sessions, tables and business logic through defined extension points, stored outside standard code so vendor upgrades and cloud updates do not overwrite them.

Infor LN & Baan

Freeze and Unfreeze (LN Planning)

Freeze and unfreeze are Infor LN Enterprise Planning actions that lock planned orders or planned quantities so a planning run cannot change or delete them. Freezing protects near-term commitments inside the planning horizon; unfreezing returns the records to normal regeneration.

Infor LN & Baan

Warehouse Order

A warehouse order is the execution document in Infor LN Warehousing that carries out an inbound, outbound or transfer movement. It is generated from an originating order such as a sales, purchase, production or transfer order, and its order type defines the procedure steps.

Infor LN & Baan

Handling Unit

A handling unit in Infor LN is an identified physical package - a pallet, carton or container - tracked with its own ID. Handling units nest hierarchically, carry item, lot and serial data, and move through warehousing and shipping as a single scannable entity.

Infor LN & Baan

Cost Component

A cost component in Infor LN is a bucket that classifies part of an item's cost - material, operation, subcontracting, surcharge or general cost - so standard and actual cost prices can be broken down, rolled up through the bill of material and compared.

Infor LN & Baan

Project Pegging

Project pegging in Infor LN links demand, supply, inventory and cost to a specific project, element or activity. Pegged material is reserved for that project only and its costs flow to the project budget, which is how contract-driven manufacturers segregate program inventory.

Infor LN & Baan

Integration Transaction

An integration transaction is the record Infor LN creates when a logistic event such as a receipt, issue, shipment or production completion must post to finance. Mapping schemes translate each transaction into the correct debit and credit ledger accounts and dimensions.

Infor LN & Baan

Multisite (Infor LN)

Multisite in Infor LN is the configuration pattern for running several plants, warehouses or legal entities in one environment using sites, enterprise units and multiple companies, with intercompany trade rules that price and post transfers between them automatically.

Infor LN & Baan

Enterprise Planning

Enterprise Planning is Infor LN's supply planning engine. It runs master planning and order planning across plan items, sites and companies, generating planned orders and reschedule signals from demand, inventory, lead times and the supply strategies defined per plan item.

Infor LN & Baan

LN Item Code

An LN item code is the unique key identifying a part, product, service or cost item in Infor LN. It is carried on every order line, bill of material and inventory record, and its length and segmentation are fixed in the data dictionary before go-live.

Infor LN & Baan

LN Bill of Material

An LN bill of material is the parent-child list defining which components and quantities go into a manufactured item in Infor LN. Each line carries a position number, net quantity, scrap factor, effective dates and, optionally, the routing operation that consumes it.

Infor LN & Baan

Baan/LN Tools Package

The Tools package is Baan and Infor LN's own development and administration layer. It holds the data dictionary, session and form definitions, the 3GL and 4GL compilers, VRC software layering, authorization management, job scheduling, company setup and the Exchange module.

ERP Core Concepts

ERP (Enterprise Resource Planning)

ERP (Enterprise Resource Planning) is integrated business software that runs finance, manufacturing, supply chain, sales, and HR processes on a single shared database, so a transaction entered once updates every dependent function in real time.

ERP Core Concepts

MRP (Material Requirements Planning)

MRP (Material Requirements Planning) is the ERP calculation that explodes demand through bills of materials, nets it against on-hand and on-order inventory, and offsets by lead time to produce timed purchase and work order suggestions.

ERP Core Concepts

MRP II

MRP II (Manufacturing Resource Planning) is the closed-loop extension of MRP that adds capacity planning, shop floor control, purchasing execution, and financial simulation, so the material plan is validated against labor, machine, and money constraints before release.

ERP Core Concepts

BOM (Bill of Materials)

A BOM (bill of materials) is the structured, hierarchical list of components, sub-assemblies, and quantities required to produce one unit of a parent item, and it is the primary input MRP explodes to calculate material requirements.

ERP Core Concepts

Routing

A routing is the ERP record that defines the ordered sequence of operations required to manufacture an item, specifying for each step the work center or machine, setup time, run time per unit, move and queue time, and any outside processing.

ERP Core Concepts

Lead Time

Lead time is the elapsed time between initiating an order and having the material available for use. In ERP it is the offset MRP applies when back-scheduling a planned order from its due date to its release date.

ERP Core Concepts

Safety Stock

Safety stock is buffer inventory carried above expected demand to protect against variability in demand and supply during replenishment lead time. In ERP it is netted out of available inventory so MRP plans replenishment before the buffer is consumed.

ERP Core Concepts

Reorder Point

A reorder point is the inventory level at which a replenishment order is triggered. It equals expected demand during the replenishment lead time plus safety stock, so stock on hand covers usage until the new supply arrives.

ERP Core Concepts

Lot Sizing

Lot sizing is the ERP rule that determines how much to order or produce in each replenishment, converting net requirements into order quantities by balancing setup or ordering cost against inventory carrying cost.

ERP Core Concepts

ATP (Available to Promise)

ATP (available to promise) is the portion of inventory and scheduled production not yet committed to existing customer orders, calculated by date so order entry can quote a delivery a customer can rely on.

ERP Core Concepts

CTP (Capable to Promise)

CTP (capable to promise) is an order promising method that goes beyond available inventory by simulating whether required material and work center capacity can be obtained in time, returning a delivery date the plant can actually achieve.

ERP Core Concepts

MPS (Master Production Schedule)

An MPS (master production schedule) is the time-phased statement of which end items a plant will produce, in what quantities, and in which periods. It is the anchor demand input MRP explodes into component requirements.

ERP Core Concepts

Demand Planning

Demand planning is the process of forecasting future customer demand by combining statistical analysis of shipment history with sales, marketing, and customer intelligence, producing the consensus demand signal that drives master scheduling and MRP.

ERP Core Concepts

Capacity Planning

Capacity planning is the process of comparing the labor and machine hours a production plan requires against the hours each work center actually has available, identifying overloads early enough to reschedule, add shifts, or outsource.

ERP Core Concepts

Work Order

A work order is the ERP document that authorizes production of a defined quantity of an item by a due date. It snapshots the bill of materials and routing, drives material issues and labor reporting, and collects actual cost for the job.

ERP Core Concepts

Purchase Requisition

A purchase requisition is an internal request to procure goods or services, routed for approval against budget and authority limits before a buyer sources it and converts it into a purchase order sent to a supplier.

ERP Core Concepts

Three-Way Match

A three-way match is the accounts payable control that compares the purchase order, the goods receipt, and the supplier invoice on quantity and price before releasing payment, ensuring a company pays only for what it ordered and actually received.

ERP Core Concepts

General Ledger

The general ledger is the central accounting record of a business, where every financial transaction posts as balanced debits and credits against accounts in the chart of accounts, producing the trial balance behind all financial statements.

ERP Core Concepts

Chart of Accounts

A chart of accounts is the organized list of accounts a business uses to classify every financial transaction, typically structured with segments for natural account plus dimensions such as site, department, product line, or project.

ERP Core Concepts

Period Close

Period close is the recurring process of finalizing a fiscal period: completing transactions, reconciling subledgers to the general ledger, posting accruals and adjustments, running costing, then locking the period so reported results cannot change.

ERP Core Concepts

Standard Cost

Standard cost is a predetermined per-unit cost for material, labor, and overhead, set in advance and used to value inventory and transactions. Differences between standard and actual are captured as variances that isolate where cost deviated.

ERP Core Concepts

Actual Cost

Actual cost is the real cost incurred to produce or acquire an item, accumulated from the specific material prices paid, labor hours reported, and overhead applied, rather than from a predetermined standard.

ERP Core Concepts

Landed Cost

Landed cost is the total cost of bringing purchased goods into inventory, combining the supplier price with freight, insurance, customs duty, tariffs, brokerage, and handling, so inventory is valued at what the material truly cost to acquire.

ERP Core Concepts

Cycle Counting

Cycle counting is the practice of continuously verifying inventory accuracy by counting a small subset of items each day or week, rather than shutting down for one annual wall-to-wall physical inventory.

ERP Core Concepts

ABC Analysis

ABC analysis is an inventory classification method that ranks items by annual consumption value and groups them into A, B, and C classes, so planning attention, counting frequency, and control effort are concentrated where the money is.

Manufacturing & Operations

OEE (Overall Equipment Effectiveness)

OEE (Overall Equipment Effectiveness) is a manufacturing metric that multiplies Availability, Performance, and Quality into a single percentage showing how much of scheduled production time an asset spends making good parts at its rated speed.

Manufacturing & Operations

Takt Time

Takt time is the rate at which a process must complete one unit to meet customer demand, calculated as available production time divided by demand for that period. It sets the pace a line must sustain, not the speed it can achieve.

Manufacturing & Operations

Cycle Time

Cycle time is the actual elapsed time a process takes to complete one unit, measured from the start of work on that unit to the start of work on the next. It describes what the process does, not what the customer requires.

Manufacturing & Operations

Throughput

Throughput is the rate at which a process, line, or plant produces good, sellable units over a defined period, expressed as units per hour, shift, or week. Only conforming output that can be sold counts toward throughput.

Manufacturing & Operations

WIP (Work in Process)

WIP (Work in Process) is inventory that has been released to production and has begun conversion but is not yet a finished good. It includes partially machined parts, subassemblies, and material sitting in queue between operations.

Manufacturing & Operations

Kanban

Kanban is a visual pull system in which a card, bin, or electronic signal authorizes production or replenishment of a specific part only after downstream consumption, capping inventory at the number of signals in circulation.

Manufacturing & Operations

Lean Manufacturing

Lean manufacturing is a production philosophy derived from the Toyota Production System that maximizes customer value by systematically identifying and eliminating waste, defined as any activity that consumes resources without adding value the customer would pay for.

Manufacturing & Operations

Six Sigma

Six Sigma is a data-driven improvement methodology that reduces process variation and defects using statistical analysis, targeting a capability level of 3.4 defects per million opportunities, typically executed through the five-phase DMAIC cycle.

Manufacturing & Operations

SPC (Statistical Process Control)

SPC (Statistical Process Control) is the use of control charts and statistical rules to monitor a process in real time, distinguishing normal common-cause variation from special-cause variation that signals a genuine change requiring action.

Manufacturing & Operations

Cp and Cpk

Cp and Cpk are process capability indices. Cp compares the specification width to the process spread, assuming perfect centering, while Cpk also accounts for how far the process mean has drifted from the center of the tolerance.

Manufacturing & Operations

FMEA (Failure Mode and Effects Analysis)

FMEA (Failure Mode and Effects Analysis) is a structured risk assessment that identifies how a product or process can fail, the effects and causes of each failure, and existing controls, then prioritizes actions to prevent the highest-risk failures.

Manufacturing & Operations

PPAP (Production Part Approval Process)

PPAP (Production Part Approval Process) is a standardized supplier approval package, originating in the automotive industry, that demonstrates a supplier's production process can consistently produce conforming parts at the quoted rate before shipments are authorized.

Manufacturing & Operations

First Pass Yield

First pass yield is the percentage of units that complete a process correctly the first time, without rework, repair, or retest. Units that eventually pass after rework are excluded, so the metric captures hidden factory loss.

Manufacturing & Operations

Scrap Rate

Scrap rate is the percentage of production output discarded as unusable because it cannot be economically reworked to specification, calculated as scrapped units divided by total units produced over a defined period.

Manufacturing & Operations

Rework

Rework is action taken on nonconforming product to bring it into full conformance with the original specification. It differs from repair, which makes product acceptable for use without meeting the original requirement.

Manufacturing & Operations

SMED (Single-Minute Exchange of Die)

SMED (Single-Minute Exchange of Die) is a structured method developed by Shigeo Shingo for reducing equipment changeover time to under ten minutes, primarily by converting setup steps that require the machine stopped into steps performed while it still runs.

Manufacturing & Operations

TPM (Total Productive Maintenance)

TPM (Total Productive Maintenance) is a company-wide approach to equipment reliability that shifts routine care to operators, builds structured planned maintenance, and targets the elimination of breakdowns, speed loss, and defects caused by equipment condition.

Manufacturing & Operations

MTBF (Mean Time Between Failures)

MTBF (Mean Time Between Failures) is the average operating time a repairable asset runs between consecutive failures, calculated as total operating time divided by the number of failures in that period. It measures reliability, not expected lifetime.

Manufacturing & Operations

MTTR (Mean Time To Repair)

MTTR (Mean Time To Repair) is the average time required to restore a failed asset to full operating condition, calculated as total repair time divided by the number of repairs. It measures maintainability rather than reliability.

Manufacturing & Operations

Andon

Andon is a visual and audible signaling system that lets any operator immediately alert supervisors and support staff to an abnormal condition, and in its full form authorizes stopping the line until the problem is addressed.

Manufacturing & Operations

Poka-Yoke

Poka-yoke is a mistake-proofing technique that designs a process or fixture so a human error either cannot occur at all or is detected immediately, before it becomes a defect passed to the next operation.

Manufacturing & Operations

Value Stream Mapping

Value stream mapping is a lean method that diagrams every material and information flow required to deliver a product family from order to delivery, quantifying processing time, waiting time, and inventory at each step to expose where lead time is consumed.

Manufacturing & Operations

Theory of Constraints

Theory of Constraints is a management methodology, developed by Eliyahu Goldratt, holding that every system is limited by a small number of constraints, and that overall performance improves only when effort is focused on the constraint.

Manufacturing & Operations

Traceability

Traceability is the ability to follow the history, location, and application of a part or material through the supply chain and production process, both backward to its sources and forward to every product and customer it reached.

Manufacturing & Operations

Bottleneck

A bottleneck is the resource in a process whose available capacity is less than the demand placed on it, making it the step that limits total system output. Every other resource can only run as fast as the bottleneck permits.

On-Prem AI

On-prem AI is the practice of running artificial intelligence models on hardware you own and control, inside your own data center or plant, so prompts, documents, and model weights never leave your network.

On-Prem AI

LLM (Large Language Model)

An LLM (Large Language Model) is a neural network with billions of parameters trained on huge volumes of text to predict the next token in a sequence, which gives it broad language understanding, reasoning, and generation ability.

On-Prem AI

SLM (Small Language Model)

An SLM (Small Language Model) is a compact language model, typically between about 1 and 15 billion parameters, designed to run on a single modest GPU or even a CPU while still handling focused enterprise tasks well.

On-Prem AI

RAG (Retrieval-Augmented Generation)

RAG (Retrieval-Augmented Generation) is an architecture that searches your own document store for passages relevant to a question, inserts them into the LLM prompt, and instructs the model to answer only from that retrieved evidence.

On-Prem AI

Vector Database

A vector database is a system that stores high-dimensional embedding vectors and retrieves the ones most similar to a query vector, enabling semantic search where results match meaning rather than exact keywords.

On-Prem AI

Embedding

An embedding is a list of numbers, typically several hundred to a few thousand dimensions, that represents the meaning of a piece of text, image, or code so that semantically similar items produce vectors that sit close together.

On-Prem AI

Fine-Tuning

Fine-tuning is the process of continuing to train an already pretrained model on a smaller set of your own examples, adjusting its weights so it reliably produces the formats, vocabulary, and behavior your task requires.

On-Prem AI

LoRA (Low-Rank Adaptation)

LoRA (Low-Rank Adaptation) is a fine-tuning method that freezes a model's original weights and trains small low-rank matrices injected alongside them, achieving most of the benefit of full fine-tuning at a small fraction of the memory cost.

On-Prem AI

Quantization

Quantization is the technique of storing a model's weights at lower numeric precision, for example 8-bit or 4-bit integers instead of 16-bit floats, which sharply reduces GPU memory use and often increases speed at a modest accuracy cost.

On-Prem AI

Inference

Inference is the phase in which a trained model is run to produce output from new input, as opposed to training, where the model's weights are being learned. In production, inference is where essentially all AI compute cost is spent.

On-Prem AI

Token

A token is the basic unit of text a language model processes, usually a word fragment rather than a whole word. In English, one token averages roughly four characters, so about 100 tokens covers 75 words.

On-Prem AI

Context Window

A context window is the maximum number of tokens a language model can hold in view at one time, covering the system instructions, conversation history, retrieved documents, and the generated response combined.

On-Prem AI

GPU VRAM

GPU VRAM is the high-bandwidth memory physically attached to a graphics processor. For AI serving it must simultaneously hold the model weights, the key-value cache for every active request, and activation buffers, which makes it the binding constraint on most deployments.

On-Prem AI

Tensor Parallelism

Tensor parallelism is a technique that splits the weight matrices inside each model layer across several GPUs, so every card holds a slice and computes part of each operation, letting a model run that no single GPU could hold.

On-Prem AI

vLLM

vLLM is an open-source inference server for large language models that uses PagedAttention memory management and continuous batching to serve many concurrent users at high throughput, exposing an OpenAI-compatible HTTP API.

On-Prem AI

Air-Gapped AI

Air-gapped AI is the deployment of AI models on a network that has no physical or logical connection to the internet or to any less-trusted network, so all model files, dependencies, and data must be transferred through a controlled process.

On-Prem AI

Open-Weight Model

An open-weight model is one whose trained parameter files are published for download, so anyone can run it on their own hardware, inspect it, quantize it, and fine-tune it, subject to whatever license the publisher attaches.

On-Prem AI

Model Weights

Model weights are the numeric parameters inside a neural network, learned during training, that determine how input is transformed into output. They are the model: distributed as files, they encode everything the system has learned.

On-Prem AI

Hallucination

A hallucination is output from an AI model that is fluent and confident but factually wrong or entirely fabricated, produced because language models generate statistically plausible text rather than retrieving verified facts.

On-Prem AI

AI Guardrails

AI guardrails are the technical and policy controls placed around an AI system that constrain what it may receive as input, what it may produce as output, and what actions it is permitted to take, independent of what the model itself decides.

On-Prem AI

Prompt Injection

Prompt injection is an attack in which instructions are hidden inside content an AI system reads, causing the model to follow the attacker's directions instead of the operator's, because a language model cannot reliably separate instructions from data.

On-Prem AI

AI Agent

An AI agent is a system that wraps a language model with tools, memory, and a goal so it can plan a sequence of steps, call external systems, observe the results, and iterate until the objective is met or a limit is reached.

On-Prem AI

MCP (Model Context Protocol)

MCP (Model Context Protocol) is an open standard that defines how AI applications connect to external tools and data sources, so any compliant client can use any compliant server without custom integration code for each pair.

On-Prem AI

Sovereign AI

Sovereign AI is the principle that a nation or organization should retain legal and physical control over the AI models, data, and computing infrastructure it depends on, rather than relying on systems governed by another jurisdiction.

On-Prem AI

Model Distillation

Model distillation is a training technique in which a small student model learns to reproduce the behavior of a larger teacher model, retaining most of the teacher's quality on a target task at a fraction of the size and serving cost.

Compliance & Standards

CMMC (Cybersecurity Maturity Model Certification)

CMMC (Cybersecurity Maturity Model Certification) is the US Department of Defense program that verifies defense contractors implement required cybersecurity controls before contract award, using three levels of assessment tied to the sensitivity of the information they handle.

Compliance & Standards

ITAR

ITAR (International Traffic in Arms Regulations) is the US State Department regulation, at 22 CFR Parts 120-130, that controls the export of defense articles, defense services, and related technical data listed on the United States Munitions List.

Compliance & Standards

EAR (Export Administration Regulations)

EAR (Export Administration Regulations) is the US Commerce Department regulation, at 15 CFR Parts 730-774, that controls the export, reexport, and transfer of dual-use commercial items, software, and technology through the Commerce Control List.

Compliance & Standards

DFARS 252.204-7012

DFARS 252.204-7012 is the Defense Department contract clause requiring contractors to protect covered defense information by implementing NIST SP 800-171, to report cyber incidents to DoD within 72 hours, and to flow the same obligations down to subcontractors.

Compliance & Standards

NIST SP 800-171

NIST SP 800-171 is the National Institute of Standards and Technology publication that defines the security requirements for protecting Controlled Unclassified Information in nonfederal systems, and it is the technical baseline behind DFARS 252.204-7012 and CMMC Level 2.

Compliance & Standards

CUI (Controlled Unclassified Information)

CUI (Controlled Unclassified Information) is unclassified information created or possessed by the US government, or by a contractor on its behalf, that laws, regulations, or government-wide policy require to be safeguarded or restricted in dissemination.

Compliance & Standards

FedRAMP

FedRAMP (Federal Risk and Authorization Management Program) is the US government program that standardizes security assessment, authorization, and continuous monitoring for cloud services, allowing one rigorous authorization to be reused across federal agencies.

Compliance & Standards

AS9100

AS9100 is the quality management system standard for the aviation, space, and defense industry, published by SAE and the International Aerospace Quality Group, which incorporates all of ISO 9001 and adds sector-specific requirements for safety, configuration, and risk.

Compliance & Standards

AS9102 First Article Inspection

AS9102 First Article Inspection is the aerospace requirement to verify that a production process reliably produces conforming parts by fully inspecting and documenting a representative first article against every drawing and specification characteristic.

Compliance & Standards

AS5553 (Counterfeit Parts)

AS5553 is the SAE standard defining requirements for avoiding, detecting, mitigating, and disposing of counterfeit electrical, electronic, and electromechanical parts, used by aerospace and defense organizations to build a documented counterfeit parts control plan.

Compliance & Standards

ISO 9001

ISO 9001 is the international standard specifying requirements for a quality management system, enabling an organization to consistently provide products and services that meet customer and regulatory requirements and to improve through a process-based, risk-aware approach.

Compliance & Standards

IATF 16949

IATF 16949 is the automotive industry quality management system standard, published by the International Automotive Task Force, that supplements ISO 9001 with automotive-specific requirements and must be implemented together with ISO 9001 rather than on its own.

Compliance & Standards

ISO 13485

ISO 13485 is the international quality management system standard for organizations involved in the design, production, installation, or servicing of medical devices, emphasizing regulatory compliance, risk management, and documented traceability throughout the product lifecycle.

Compliance & Standards

FDA 21 CFR Part 11

FDA 21 CFR Part 11 is the US regulation that defines the criteria under which the FDA accepts electronic records and electronic signatures as equivalent to paper records and handwritten signatures, requiring validation, audit trails, access controls, and signature integrity.

Compliance & Standards

FSMA Rule 204

FSMA Rule 204 is the FDA Food Traceability Rule requiring companies that handle foods on the Food Traceability List to capture defined key data elements at critical tracking events and produce those records to FDA in a sortable electronic format within 24 hours.

Compliance & Standards

SOX (Sarbanes-Oxley) Compliance

SOX compliance is a US public company's obligation under the Sarbanes-Oxley Act of 2002 to establish, document, test, and certify internal control over financial reporting, including the IT general controls and segregation of duties in the systems that produce financial data.

Compliance & Standards

EDI (Electronic Data Interchange)

EDI (Electronic Data Interchange) is the computer-to-computer exchange of business documents such as purchase orders, invoices, and shipping notices between trading partners in a standardized structured format, without human intervention.

Compliance & Standards

EDI 850 Purchase Order

The EDI 850 is the ANSI X12 transaction set used to transmit a purchase order electronically from a buyer to a supplier, carrying header terms, ship-to and bill-to parties, and line item details such as quantity, price, and product identifiers.

Compliance & Standards

API (Application Programming Interface)

An API (Application Programming Interface) is a defined contract that lets one software system request data or trigger functions in another, specifying the available operations, the structure of requests and responses, and the rules for authentication and error handling.

Compliance & Standards

REST API

A REST API is a web interface that exposes application data as addressable resources over HTTP, using standard verbs such as GET, POST, PUT, PATCH, and DELETE, stateless requests, and conventional status codes, typically exchanging JSON.

Compliance & Standards

SOAP

SOAP is an XML-based messaging protocol for exchanging structured information between systems, in which each message is an envelope containing an optional header and a body, and the available operations are formally described by a WSDL contract.

Compliance & Standards

Webhook

A webhook is an HTTP callback in which a source system automatically sends an HTTP request to a subscriber-provided URL when a defined event occurs, pushing notification to the consumer instead of requiring the consumer to poll for changes.

Compliance & Standards

Middleware

Middleware is software that sits between applications and handles the communication, transformation, routing, and reliability concerns of moving data between them, so that the applications themselves do not need to know about each other directly.

Compliance & Standards

iPaaS

iPaaS (Integration Platform as a Service) is a cloud-hosted platform that provides prebuilt connectors, data mapping, orchestration, and monitoring so organizations can build and operate integrations between applications without running their own integration infrastructure.

Compliance & Standards

Data Mapping

Data mapping is the process of defining how each field in a source system corresponds to a field in a target system, including the transformation rules, value translations, default values, and conditional logic needed to move data correctly between them.

Definitions are the easy part

Configuring these systems correctly is the hard part. That is what our AI agents and engineers do - for SyteLine, LN, M3, and on-prem AI stacks.

Talk to an Expert Try Our Free Tools