Architectural Category Tags and Risk Metric#
The Architectural Category Tags can be created with the menu .
This will create a tag named ArchitecturalCategories with 4 sub tags: core, peripheral, utility and control.
The tags are calculated by using the technique desribed in Daniel J. Sturtevant’s paper “System Design and the Cost of Architectural Complexity” [1]. The categories are determined by using two metrics called “Visibility Fan In” (VFI) and “Visibility Fan Out” (VFO).
- Existing metrics in Lattix are used to represent the VFI and VFO metrics as follows:
VFI = Average Impact +1VFO = Average Cumulative Dependency
In Sturtevant’s paper, a “cliff” in the VFI and VFO graphs is referenced, in which there is a sudden increase in the values of VFI and VFO. An example is represented in the following images:
In order to calculate if the VFO and VFI are low or high we determine the “cliff” value for each and name them cliffVFI and cliffVFO
A value of VFI < cliffVFI is lowA value of VFI >= cliffVFI is highA value of VFO < cliffVFO is lowA value of VFO >= cliffVFO is high
VFI |
VFO |
Tag |
Description |
|---|---|---|---|
low |
low |
peripheral |
Does not influence much of the system |
high |
low |
utility |
Relied on by much of the system |
low |
high |
control |
Coordinates the behavior of the system |
high |
high |
core |
Has large cycles and co-dependent code |
Sometimes there are multiple cliffs in the graph (see below). We choose the cliff based on several heuristics, for instance: the biggest cliff or the cliff closest to the middle. If the cliff is too close to the end of the graph (as in the the above VFO graph) we will use a cliff value closer to the middle of the graph for better results.
Risk Metric#
Since the subsystems tagged as core have both high Fan In and Fan Out connections, they are considered higher risk. So we have created a metric named Risk that is calculated as follows:
The Risk metric will be calculated as 1.0 if VFI >= cliffVFI and VFO >= cliffVFOThe Risk metric will be calculated as 0.3 if VFI >= (cliffVFI * 70%) and VFO >= (cliffVFO * 50%)The Risk metric will be calculated as 0.3 if VFI >= (cliffVFI * 50%) and VFO >= (cliffVFO * 70%)
The Risk Metric can be used in the heatmap.
