- Collocation constraint link
The collocation constraint link specifies that the units on each end of the link must be hosted on the same hosting stack. To validate this constraint, the topology editor traverses the units' hosting stacks, following hosting links and realization links to ensure that eventually the same unit is present in both stacks. In the following example, two application components are connected by a collocation constraint link. This constraint link is valid because both units are hosted on the same operating system, even though the path to the operating system goes through other units.
- Anti-collocation constraint link
The anti-collocation constraint link is the opposite of the collocation constraint link; it is violated if the units are hosted on the same hosting stack.
As with the collocation constraint link, you can provide more detail about the anti-collocation constraint link by specifying a capability type. The anti-collocation constraint is violated only if a unit with the specified capability is present in both units' hosting stacks. For example, you can specify that two units can be on the same piece of hardware but not on the same operating system. The following illustration shows two Web application components that are connected by an anti-collocation constraint link; the constraint specifies the j2ee.ServletContainer capability, allowing the units to be hosted on the same piece of hardware and operating system but not on the same Web server.
- Deferred hosting constraint link
- Like the collocation constraint link, the deferred hosting constraint
link ensures that two units are in the same hosting stack. However,
the collocation constraint link is used between peers, such as two
applications; the deferred hosting constraint link is used between
a host and a hosted unit. In this way, the deferred hosting constraint
link specifies that one unit is hosted on another unit, though there
may be other units in between them. For example, a Web application
may have a deferred hosting constraint link to an operating system
unit, indicating that the Web application runs on the operating system,
but that there may be other units between the two, such as a Web server.
This constraint is violated if the units are not in the same hosting
stack, as in this example of a Web server with a deferred hosting
constraint link to an operating system that does not host the application's
Web server:
- Application communication constraint link
- The application communication constraint link indicates that two
application units must be able to communicate. Therefore, the hosting
stacks of those units must have a network connection or other means
of communication. This constraint is most often used in logical models
to indicate how computer systems in different locations must be connected.
For example, in the following illustration, two components are connected
by an application communication constraint link, so the locations
on which the components are hosted must have a path of communication,
such as a network communication constraint link:
The application communication constraint link can have child constraints that describe the communication between the units. For example, the application protocol constraint specifies the protocol that is used for the communication, and the communication bandwidth constraint specifies the amount of bandwidth that is used.
- Network communication constraint link
- This type of constraint link represents a path of communication
between two pieces of network infrastructure. The topology editor
does not validate network communication constraint links and thus
this type of constraint link is never violated. Instead, this type
of constraint link plays a part in the validation of application communication
constraint links, as described above. Network communication constraint
links can also serve as metadata to specify how the parts of computer
networks communicate. For example, you can correct the violation in
the application communication constraint link example by adding a
network communication constraint link between the locations:
Like the application communication constraint link, the network communication constraint link can have child constraints that describe the communication between the units. For example, the port configuration constraint specifies the port through which the units communicate, and the transmission delay constraint specifies the longest acceptable delay for communication between the units.
- Attribute propagation constraint link
- The attribute propagation constraint link ensures that two attributes
on two separate topology elements are set to the same value. For example,
if you want to ensure that two servers are running the same version
of Java, you can create an attribute
propagation constraint link between the units, specifying the two Java version attributes:
