Hi Julian,
I think the scraper is a good fit for PLC4X where it is right now. What could
be a good option (I guess Tim is working on refactoring the scraper right now),
would be to really introduce different "trigger" implementations.
The Timer to be the one achieving what it currently does, and provide an
integration-module in Edgent that uses the PLC4X Scraper and provides a custom
trigger implementation to integrate into the "pull" or "push" of Edgent.
You think that would be possible too?
I think the scraper is a great tool on its own and I would like to not enforce
people to have to dig into Edgent too just in order to have the standalone
PLC4X scraper running.
Besides that, I +1 one the other suggestions.
However I also agree (Not mentioned in your mail, but previous ones) that we
should move the Edgent connector from the PLC4X project to Edgent project.
Chris
Am 04.03.19, 09:34 schrieb "Julian Feinauer" <j.feina...@pragmaticminds.de>:
Hi community,
as I finished some private projects, I feel like I have some time and new
energy to spend more efforts to edgent.
I have several ideas on how we can kick off the project again and would
like to share / discuss them with you guys.
* First, I suggest that we add adapters (similar to the IotpDevice for
IBM Watson) for other relevant Cloud plattforms, like Azure, Google, AWS, … .
* Second, we open sourced our Framework CRUNCH for industrial analytics
late last year [1]. I would like to integrate it in edgent to allow more
refined edge analytics.
* Third, I think a good place for something like the PLC Scraper (from
plc4x [2]) (perhaps with some configuration stuff as standalone) is Edgent. So
we could build complete pipelines there from connecting to the plc over
analytics with crunch to some kind of storage (local, cloud)
* Finally, the iotdb project [3] which recently joined the incubator is
doing really great stuff and it would be good to also integrate it in edgent as
a sink (and source). I see main purposes here, one as a buffer (think of kafka)
and second as archive store.
I’m cross posting this mail to the lists for edgent, plc4x and iotdb to let
all communities know but please reply to this mail only on the EDGENT list, as
I think its best to keep the discussion here.
What do you think of these ideas?
And perhaps we can also win some contributors of the other projects
(crunch, plc4x, iotdb) to support these efforts.
Thanks!
Julian
[1] https://github.com/pragmaticminds/crunch
[2] https://plc4x.apache.org/
[3] https://iotdb.apache.org/
