GSoSD Review Discussion Point 1: Should Arrowhead systems be "lazy" (passive) or "active" (proactive)?

[PerOlofsson-Sinetiq]

We need to make sure that the architecture in the GSoSD are capable of adressing a flora of systems, ranging from systems that basically just starts up and register themselves in appropriate registries to advanced systems that handle for example orchestration by themself.

Should we create use-cases that describes this situation? Are there other ways to incorporate these requirements?

[PerOlofsson-Sinetiq]

I suggest the following approach to the issue:

1. We define the terms "active systems" and "passive systems".
2. We come up with a suggestion of the Arrowhead 5.0 principle of interaction with active/passive systems.

[borditamas]

AITIA's view on the subject:
We think that a service oprertion could be passive or active (not a system itself) depending on the purpose of operation. Could you please explain why should we define such a restriction for the Arrowhead Systems? What made this question coming up?

[PerOlofsson-Sinetiq]

AITIA's view on the subject: We think that a service oprertion could be passive or active (not a system itself) depending on the purpose of operation. Could you please explain why should we define such a restriction for the Arrowhead Systems? What made this question coming up?

No, i do not associate the operations with the term "active"/"passive".

The question adresses whether systems can make its own decisions and act within the Arrowhead cloud (an "active system") or if they are commanded to act on behalf of some other system or an operator (a "passive system").

I think this is the reason for the definition point of this issue. I´ll adress that in the following post.

[PerOlofsson-Sinetiq]

I would like to propose the following definition of "active" and "passive" systems:
- Active systems
Active systems are systems that perform actions on existing internal and external data. These systems can initiate orchestration by themself in order to carry out tasks. These systems are acting in various levels of autonomy.
- Passive systems
Passive systems are systems that wait for external stimuli before performing any execution. These systems probably have less computational capabilities than the active counterpart. These systems rely on the Arrowhead Core services to a large extent to get necessary data.

The two types of systems are putting different requirements on the Arrowhead Core Services. The more "active" a system is, the more of the core capabilities are present within the application system itself. More "active" systems will require more tolerant core services. On the contrary, the more "passive" a system is, the more responsibility is put on the core services to provide data to the application system.

[PerOlofsson-Sinetiq]

Another aspect, related to the legacy integration. Lots of legacy systems tend to lean at the "active" system paradigm, with complex and holistic approach to the stated requirements. In the future, where we might se more microservice / microsystem solutions, a more "passive" approach might be present. However, the Arrowhead Framework will always have to deal with legacy systems that tries to lock users into their solution.

[borditamas]

AITIA view

Considering the proposed definitions we think that both kind of systems have to be feasable within an Arrowhead Local Cloud.

[jerkerdelsing]

Seams that the meeting July 28 agrees.