What is a schematic feature?
Schematic features are implemented by a schematic feature class.
These schematic features display in schematic diagrams that are generated/opened in ArcMap, previewed in ArcCatalog, or created/visualized via any application developed with the ArcGIS Schematics API.
The schematic features contained in a schematic diagram are regrouped into feature layers. There is a feature layer for each schematic feature class that is associated with the diagram template which implements the diagram. Each feature layer has specific layer properties that control how to draw the related schematic features (in one symbol, showing categories, and so forth) or whether to label them, in the exact same way the display is controlled for standard geographic features. The default layer properties are set within Schematic Dataset Editor and upheld by the schematic diagram template.
Schematic features are always stored in a feature class table in the schematic dataset.
ArcGIS Schematics implements four types of schematic features: node, node on links, link, and sublink. These schematic features are defined below.
Node and node on link
A schematic node is a junction in a network graph. It is usually represented by a point feature in schematic diagrams.
Nodes can also be represented by a polygon or polyline feature. This mainly applies to schematic features that need to behave like containers and relate a set of schematic features.
The symbol used to display a node can be a composed symbol with symbol layers tagged as schematic ports (using the specific SchematicPort string), which can then be used to manage schematic links connection to ports.
Learn about how to represent schematic nodes with ports
Nodes can have geographic, semigeographic, or graph coordinates. After an edit session is started on the schematic diagram in which they are displayed, nodes can be manually moved using many Schematics editing tools. Several nodes can be aligned along the same horizontal or vertical axis. Two nodes can have a binding-logical and hierarchical relationship, which means that certain rules apply when these nodes are manipulated—for example, moving a parent node causes the associated child node to be moved accordingly.
A schematic node on link is a node positioned on a link path route. Its location on the schematic link is either absolute or relative. A node on link is completely dependent on the link to which it is related. The 
Edit/Move Node On Link tool allows you to move a node on link along its reference link during an edit session. Selecting and moving a link automatically selects and moves its associated nodes on links. The destruction of the link causes the automatic destruction of its nodes on links. Links that connect nodes on links are sublinks.
All the attributes that can be used to symbolize schematic nodes/nodes on links must be specified within Schematic Dataset Editor on the schematic feature class that implements those nodes/nodes on links.
Link and sublink
A schematic link is an edge in a network graph. It is represented by a line feature in schematic diagrams. A link can go through a number of link points, or vertices, that modify its path. It is represented by either a direct single-line segment or several line segments passing through one or several link points. Schematics tools for links editing are available after an edit session is started on the schematic diagram that contains the schematic links.
A schematic sublink is a link that exists as part of another link. It connects one node on link to one of its related link's extremity nodes or connects two nodes on links. Sublinks are completely dependent on the schematic links to which they are related and on the nodes on links that they connect. Selecting and moving a link automatically selects and moves its sublinks. The destruction of the link causes the automatic destruction of its sublinks.In the same way, deleting a node on link causes the deletion of its connected sublinks.
All the attributes that can be used to symbolize schematic links/sublinks must be specified within Schematic Dataset Editor on the schematic feature class that implements those links/sublinks.
Defining nodes on links and sublinks for your links is useful when you have to graphically display a line that is composed of several different parts. For example, in the electrical domain, electrical lines combine an overhead part and an underground part along their path route. In this case, it could be useful to graphically distinguish the overhead from the underground using sublinks and nodes on links.