[aadl-modeling]: Latency Analysis

[Philip Alldredge]

Hello,
Typically for latency analysis one would connect components using ports and
port connections. Then define an end to end flow between the systems. The
end to end flows would be defined in the top level system and reference
that systems connections and the flow specifications defined in the sub
systems.



Philip Alldredge
Research Scientist
Rotorcraft Systems Engineering and Simulation Center
University of Alabama in Huntsville
VBRH A-4
256-824-4837
Philip.Alldredge at uah.edu

On Fri, Jul 20, 2018 at 4:19 AM, David K <fux1235 at googlemail.com> wrote:

> Hello,
>
> I want to do a latency analysis and have a question about ports and
> comonent-asccess again :/
> I understand that port-connections are purely logical and
> component-accesses are physical.
> If I use a system-component to assemble my software-components and
> subsystems how do I need to model the connections and especially the flows
> (physically and logical)? I want to model the subcomponent communication
> with port-connections and the communications between the systems with bus
> accesses. My question is how do I get the data from the subcomponents of a
> system e.g. "device1" to the subcomponents of the system "device3" with the
> architecture of the image below.
>
> Until now my models only contain port connections. So the subcomponents of
> the systems are connected via ports. Then systems themselves are connected
> via ports too because I don't know how to implement the data flows across
> several systems with component-accesses.
>
> As far as I tried I can't allocate multiple datatypes to one port
> interface and I can't use the flow mechanism to connections. The examples I
> use for my research doesn't contain such cases. They all use just one
> system with it's subcomponents. Neither of them handles the communication
> between two systems and their subcomponents via buses. For the FTA it was
> sufficient to change the "bus accesses" from the image below into ports and
> then allocate them to the bus. This works fine but I suppose this wasn't
> the correct way to do it.
>
>
> Thanks in advance
> David
>
> [image: image.png]
>
-------------- next part --------------
HTML attachment scrubbed and removed
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image.png
Type: image/png
Size: 25457 bytes
Desc: not available
URL: <http://lists.sei.cmu.edu/pipermail/aadl-modeling/attachments/20180720/4d5d2f32/attachment.png>