Hi All,
first of all, a very warm welcome Veronika :-)
Really happy I managed to spark her interest. I know I tried working on this
multiple times, but I always noticed my brain isn't wired the right way for
that ;-)
Data-Scientists are much better at such problems ...
@Julian Feinauer ... nope ... no jira yet, but we should create one ... I'll
take care of that.
And to all others, please let me explain a little more what Veronika is going
to try to do:
Almost all protocols support asking for multiple items in one request. However
the size of a request is usually limited.
Now there are multiple possible optimizations possible:
- Aggregate items that are close to each other (Instead of reading multiple
bytes which are located in close proximity, load one byte array instead)
- Rearange the items so request and response are filled to the maximum
- Split the request up into multiple requests
For the S7 driver I had implemented the last part, but not the others.
Instead of implementing this for one protocol (as I did for S7) she's going to
try to implement this in a generic fashion so we can add this as a layer in all
Java drivers (can port things to other languages later). Currently all
protocols except s7 read each item in a single request, which is highly
imperformant.
Really looking forward to this feature :-)
Chris
Am 06.05.19, 15:03 schrieb "Julian Feinauer" <j.feina...@pragmaticminds.de>:
Hi Veronika,
happy to hear from you (Chris announced since day one that you will come
one day to solve this problem).
So I guess his prophecy just came true!
Really looking forward to your code as there are also some other
optimizations we would like to do (like collapsing identical fields in requests
and so).
And your work will be an excellent first step!
Julian
PS: @Chris: do we have a Jira for that? If not, I can create one
Am 06.05.19, 15:01 schrieb "Veronika Schindler"
<veron...@pisquaredover6.de>:
Hey everyone!
Chris won me over to PLC4X :)
I will set out to optimize the requests to a plc, so that request and
response size are optimally used.
Veronika
