Available with Network Analyst license.
This document introduces vocabulary that is essential to understanding the ArcGIS Network Analyst extension help. They are listed in the order you are likely to encounter them. Some definitions are related to documents with more thorough descriptions.
Transportation networks—like street, pedestrian, and railroad networks—can allow travel on edges in both directions. The agent on the network—for instance, a truck driver traveling on roads—is generally free to decide the direction of traversal as well as the destination.
Source features are the line and point features used to create a network dataset. They can be thought of as the physical network, which doesn't have topology embedded within the features. The network dataset can be thought of as the logical network, which does embed the topological relationships needed to perform network analyses.
The network dataset is a collection of interconnected network elements (edges, junctions, and turns) that model undirected flow. Its most common application is in modeling road networks.
Any network analysis with the ArcGIS Network Analyst extension requires a network dataset, which is a logical network. You might wonder why a network dataset is necessary if you already have a line feature class that looks like a network. The reason is that line features don't inherently know what they are connected to, but network elements do. When a network analysis is performed, the solver needs to quickly scan a multitude of network elements to figure out which paths it can take to reach a destination. If it had to reference simple line features, it would need to perform time-consuming spatial operations for every line it inspected; the process would be too slow. Furthermore, setting up complex connectivity schemes would be difficult—if not impossible—to define with line features alone. Alternatively, when the solver scans a network dataset, the elements can provide accurate information without slowing the solver down as much.
The terms network and network dataset are often used interchangeably in the help documents.
Network datasets are made up of edges, junctions, and turns; these components are generically referred to as network elements.
Network attributes hold information about the network dataset. There are four kinds:
Network cost or impedance
Network cost and impedance refer to the same concept. Whenever an agent traverses a network element, it is charged some amount, which is the network cost. For example, a path from one city to another might have a network "cost" of 45 miles.
Network cost can be anything you choose but typically is distance or travel time. To be used in a network analysis, a network dataset must have at least one cost attribute because network analyses always optimize some cost. For instance, a route analysis finds the least-cost path between two or more points. In some cases, even more than one cost attribute is required.
After you add a network attribute to a network dataset, the attribute values need to be calculated. Evaluators serve this function.
There are many evaluators for each network attribute. A network attribute has a unique evaluator for each network element (junctions, edges, and turns) and each of their source feature classes. Moreover, there are two evaluators for each edge source feature class: one for the from-to side of edges and another for the to-from side. For example, if a cost attribute is added to a network dataset made from a streets feature class and a turns feature class, there can be an evaluator that calculates the cost attribute values for each of the following:
If more source features are added to the network, more evaluators would be necessary to calculate costs for their corresponding network elements.
There are various kinds of evaluators, such as constant evaluators that assign a single value to all elements in the group and field evaluators that can pull values from source features and assign them to their corresponding network elements.
When a network dataset is represented in ArcMap as a layer, it is called a network dataset layer or, more simply, a network layer.
Network analysis layer
Network analysis layers, or analysis layers for short, can be thought of as a framework for setting up and solving a network problem.
For instance, a route analysis layer is created when you choose to solve a route (least-cost path) problem in ArcMap. It is automatically associated with the active network dataset. Also, the analysis layer has properties that let you further define the problem, such as a property that lets you specify whether the stops should be optimally sequenced or visited in the order they are listed.
Network analysis layers also contain a set of network analysis classes that are predefined for the network problem type: route analysis layers contain network analysis classes for stops, routes, and the various barriers; origin-destination analysis layers contain origins, destinations, lines, and barriers. The analysis classes in turn hold the input data you provide and the output data provided by the solve operation.
Network analysis class
Network analysis classes are feature classes and tables. The features and records they contain serve as input and output data for network analysis layers; for example, in a route analysis layer, the stops and barriers you provide and the resulting route features are stored in network analysis classes.
Network analysis feature classes are displayed in the ArcMap table of contents as sublayers. They are not maintained on disk; rather, they are stored in memory and saved in the map document.
Network analysis objects
This is a generic term for the records and features stored in a network analysis class. If the data is a feature and is located on a network, the more precise term, network location, can be used instead.
A network location can refer to a network analysis object that is positioned on a network, or it can simply refer to a point on a network.