ArchiMate 3.0 Technology Layer MetaModel and valid relationships

1,974 views
Skip to first unread message

Steven Mileham

unread,
Apr 25, 2017, 4:44:17 AM4/25/17
to ArchiMate

Hello ArchiMate people, after a few more models, I've got a couple of questions, I'm hoping someone can point me in the right direction!


TLDR;

Question 1: Do System Software, Device and Technology Collaboration inherit relationships as they are “Specialisations” of Node?

Question 2: How can Artifacts be “Assigned” to a Node in Archimate 3.0?

 

Expanded Questions;

I have been working with a colleague to model a legacy Identity Management platform in ArchiMate 3.0.  We are currently using Archi as our modelling tool, but I recently picked up a trial of Abacus to understand their repository features.  Unlike Archi, Abacus highlights “derived” relationships separately from “core” relationships, and this has caused me to question my usage of a couple of elements.  Here is our baseline model for the Atlas Identity Management platform;



Relationships in Black: As far as I am aware, these are “core” relationships.

Relationships in Purple: These are derived relationships (shortcutting Functions and Interfaces, etc)

Relationships in Red: The following appears in the ArchiMate 3.0 metamodel;

  My interpretation (which is probably where I’ve gone wrong) is that System Software, Device and Technology Collaboration all “inherit” from Node, thus all relationships that are valid for a Node are also valid for the “Specialised” elements.  Based on this assumption, it should be valid to aggregate System Software under a Technology Collaboration as a “core” relationship and I should be able to assign Technology Functions to System Software, is this correct?


Relationships in Orange: Looking back at ArchiMate 2.1, an Artifact can be assigned to a Node; 


In order to show this in ArchiMate 3.0, do we need to model the System Software representing File Storage?  I guess technically in this example, it could be represented by the Perl scripts being assigned to “Perl” System Software running on the Node?



Mastering ArchiMate

unread,
Apr 25, 2017, 6:49:02 AM4/25/17
to Steven Mileham, ArchiMate
On 25 Apr 2017, at 10:44, Steven Mileham <steven....@gmail.com> wrote:

Hello ArchiMate people, after a few more models, I've got a couple of questions, I'm hoping someone can point me in the right direction!


TLDR;

Question 1: Do System Software, Device and Technology Collaboration inherit relationships as they are “Specialisations” of Node?

Only in the metamodel, not in the model. Hence, if you model a Business Object that is a model-specialisation of Contract (this is possible, see: What is wrong with this picture?) the Business Object does not (according to ArchiMate) inherit any relation to or from the Contract element.

However, a Contract as a type inherits all the possible relations of Business Object. This allows you to have Access relations from Business Function to Contract, for instance. The ‘core set of relations’ and the ‘derived relations’ (including those derived) are all thrown together in the table of ‘allowed relations’.

Question 2: How can Artifacts be “Assigned” to a Node in Archimate 3.0?

Quite simple: Node - Assigned-To - Artifact is allowed.

 Expanded Questions;

I have been working with a colleague to model a legacy Identity Management platform in ArchiMate 3.0.  We are currently using Archi as our modelling tool, but I recently picked up a trial of Abacus to understand their repository features.  Unlike Archi, Abacus highlights “derived” relationships separately from “core” relationships, and this has caused me to question my usage of a couple of elements.  Here is our baseline model for the Atlas Identity Management platform;



Relationships in Black: As far as I am aware, these are “core” relationships.

Relationships in Purple: These are derived relationships (shortcutting Functions and Interfaces, etc)


Relationships in Red: The following appears in the ArchiMate 3.0 metamodel;

  My interpretation (which is probably where I’ve gone wrong) is that System Software, Device and Technology Collaboration all “inherit” from Node, thus all relationships that are valid for a Node are also valid for the “Specialised” elements.  Based on this assumption, it should be valid to aggregate System Software under a Technology Collaboration as a “core” relationship and I should be able to assign Technology Functions to System Software, is this correct?

Yes.

And in ArchiMate 3, starting from that metamodel excerpt and inheritance you can do Node - Composition - System Software - Assignment - Artifact, which turns derived into Node Assignment - Artifact.


Relationships in Orange: Looking back at ArchiMate 2.1, an Artifact can be assigned to a Node; 


In order to show this in ArchiMate 3.0, do we need to model the System Software representing File Storage?  I guess technically in this example, it could be represented by the Perl scripts being assigned to “Perl” System Software running on the Node?

You do not need to model the System Software that is responsible for the derived relation Node - Assignment - Artifact. You are free in modelling what stuff you put in or leave out.

The whole derivation thing is something that you should not take too seriously…

G





--
You received this message because you are subscribed to the Google Groups "ArchiMate" group.
To unsubscribe from this group and stop receiving emails from it, send an email to open-archimate-f...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/open-archimate-forum/f95cba4d-e704-4780-940c-0b5bbe8bbbc4%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Steven Mileham

unread,
Apr 25, 2017, 8:50:16 AM4/25/17
to ArchiMate
That's great Gerben,

Thanks for taking the time to respond, I thought I was reading it right, I guess it's just an issue with the Abacus implementation of the ArchiMate 3.0 relationship table!

On your final point of "The whole derivation thing is something that you should not take too seriously…" I've been burnt a few times with incorrect derived relationship, so I've been trying to build in as many "core" relationships as possible/is useful.

Thanks again!

Steve

Steven Mileham

unread,
Apr 26, 2017, 3:50:14 AM4/26/17
to ArchiMate
Funnily enough, I just stumbled across the "Corrigendum" document to be discussed by the OpenGroup for ArchiMate 3.0.1 and it looks like there is a more explicit description of the "inheritance" and they are adding the Node to Artifact assignment relationship back in as "core".



On Tuesday, 25 April 2017 09:44:17 UTC+1, Steven Mileham wrote:
Reply all
Reply to author
Forward
0 new messages