Systems Architects and Engineers specialising in the effective use of Systems Architecture, Methods and Tools since 1982

Commutative Clustering

We refer to the method of clustering used by Strategy SAVIŽ as "Commutative" clustering. This is the most common means of grouping Process and Information to yield areas of responsibility in which groups of functions and the information which they create are defined. The rules under which this operates are as follows where two related spheres are referred to as Sphere-A and Sphere-B:

1) If Item AN in Sphere-A has a clusterable relation with Item BP in Sphere-B and also has a clusterable relation with Item BQ in Sphere-B then

Item AN,

Item BP and

Item BQ

are allocated to the same cluster;

2) If Item BX in Sphere B has a clusterable relation with Item AY in Sphere A and also has a clusterable relation with Item AZ in Sphere A, then

Item BX,

Item AY and

Item AZ

are allocated to the same cluster.

Clusterable relations are those selected as requiring action in the clustering process. For example, it is conventional in clustering a CRUD matrix (Process related to Information using the values Create, Read, Update, Delete) to cluster items based on the Create relation value.

A high level example of a matrix clustered in this way follows. It is shown as an Excel worksheet although other similar formats may be used. See Export the Matrix.

You can examine the interactions between one cluster and those to which it relates through common row or column items by using the cluster review facility.