err, to clarify myself. I wrote:
I know there is ser2net and socat, but latencies introduced by bridging
might not be as effective as you would expect.

where I meant:
I know there is ser2net and socat, but latencies introduced by bridging
might make testing less effective than you would expect.

Cheers,
Lukasz

On 05.12.2018 13:11, Łukasz Dywicki wrote:
> Hey Christofer,
> I'm fine with VPN access. I had a chance to talk with people who
> deployed "my" openhab bacnet integration with several big swegon units
> (9000 m3/h each) and gave it a try. I can agree with them to collect
> PCAPs, I could also ask other users who contacted me to get sample traffic.
> This approach will definitely work with bacnet/ip, but with VPN access
> we can't reliably test serial interfaces. I know there is ser2net and
> socat, but latencies introduced by bridging might not be as effective as
> you would expect.
> More over VPN have one important side effect. I can't show "real" thing
> being controlled with software while I meet people. Small devices with
> serial interfaces have its advantage - you can take them with you at
> fare, show that they actually work with your program and your software
> can communicate with them.
> Just think of impression which could be made with bidirectional
> communication between two and more protocols with OPC-UA on top of
> everything.
> I think such things would speak nicely to automation guys at trade shows.
> 
> Ps. You can call me Lukasz and Luke, all these variants are fine!
> 
> Cheers,
> Lukasz
> --
> Apache Karaf committer & PMC member
> Founder of http://connectorio.com
> 
> On 05.12.2018 10:46, Christofer Dutz wrote:
>> Hi Lukasz (Hope you forgive me for not using that L with the strike through 
>> ... I just don't know where to find it) 
>>
>> Happy you're willing to help out with this :-)
>>
>> Regarding hardware ... I think we might have a different option here. I know 
>> a big company near Frankfurt asked about BacNet support ... I think we might 
>> be able to convince them to provide us with a reasonably big device (so we 
>> can test all different sorts of things). If they do, we would add the device 
>> to our IoT Lab network here in the FFM codecentric office. This way everyone 
>> interested could access it through our IoT Lab VPN ... would that be an 
>> option for you? As I have a meeting with them this afternoon, I'll 
>> definitely ask them. 
>>
>> Regarding the drivers ... I think the way we currently implement most of 
>> them, should make it easy to implement a BacNET driver or a CAN driver 
>> without having to stick to one hardware platform (for the CAN bus driver). 
>> The Protocol should be the same for all hardware implementations, so I 
>> assume we could implement a CAN Protocol layer and provide multiple 
>> implementations of driver modules using the CAN protocol layer. I would be 
>> more than happy to help you get started with this (Might even write that 
>> down as I think this would be a good thing to read for others interested)
>>
>> Chris
>>
>>
>>
>> Am 05.12.18, 01:31 schrieb "Łukasz Dywicki" <[email protected]>:
>>
>>     Hey Guys,
>>     I come over plc4x already few times, had chat with Christofer as I peek
>>     into "industry" side of software from time to time and I am interested
>>     in evolution of project, both as contributor (if I have time & budget)
>>     and possible user.
>>     
>>     A while ago I authored bacnet integration for openhab (actually decoded
>>     bacnet4j magic) which is used in several places. I would love to move
>>     that part forward as bacnet is widely used in bigger buildings.
>>     Obviously because bacnet4j license is not compatible with ASLv2 nor EPL
>>     I can't contribute my work back to Eclipse Smarthome nor OpenHab project
>>     which makes maintenance of bacnet integration a real burden.
>>     I see commercial spaces as another corner of "industrial" stuff since
>>     most of them have lots of equipment which is in some cases also used in
>>     manufacturing.
>>     
>>     Coming back to the point, after going over specs and learning how
>>     advanced bacnet could be, finding that most of unit tests for bacnet4j
>>     is left for historical, but not practical reasons, I come to conclusion
>>     - hey maybe it would be good idea to create a naive and simplistic
>>     bacnet api, just to have a rough read/write support. Then we could see
>>     if it will "catch" and have any traction. So far I completed easiest
>>     part - creating interfaces without actually touching a serial/udp codec. 
>> :)
>>     I know we I can get pcaps, I can even use my home ventilation unit to
>>     collect some, but I lack a bacnet mstp interface big time. I can't test
>>     bacnet4j nor even check if my code works with it.
>>     Since there is a bunch of people on this mailing list who most likely do
>>     or did some work with industrial stuff maybe, by some chance, you are
>>     able to point some cheap equipment for bac ms/tp testing. Here I mean
>>     something basic even with one property, a switch or actuator which
>>     doesn't cost 200€. The computer side I should be able to cover with
>>     standard RS485/USB adapter.
>>     I remember that in one of materials related to plc4x there was
>>     information about some firm (or someone) who contributed test equipment
>>     to project. Maybe it would be good idea to drop an information about
>>     cheap lab equipment on the webpage, which will let people to experiment.
>>     As long as costs are close to 50€/$60 (lets call it raspberry barrier)
>>     there is a chance that people will play with things. We all know that
>>     its not a industry grade hardware, but that's not what we require for
>>     most of time to test communication.
>>     
>>     Since I already mentioned openhab, I would like also to ask about one
>>     more standard/protocol which I was asked about in context of building
>>     automation - which is CANbus. I found it today on one of pictures on
>>     plc4j related to S7 - are there any plans getting CAN supported? For
>>     quite long time I presumed that CAN is car specific thing, but
>>     apparently I was wrong and it is used "in the field" as well.
>>     Having a cheap device which speaks CAN and pluggability to computer is
>>     usually first step to get things done. It might be also last one (as we
>>     are just humans and we often are short with time for fun projects), but
>>     having it is better than not having it. :-) Hope you get the point.
>>     
>>     From my own side I might be able to contribute later this year/early
>>     next year a sample lab setup for wmbus/mbus. These two are quite popular
>>     in Europe for media consumption and have plenty of vendors who use it
>>     together following DLMS/OMS spec.
>>     
>>     Kind regards,
>>     Łukasz Dywicki
>>     --
>>     Apache Karaf committer & PMC member
>>     Founder of http://connectorio.com
>>     
>>

Reply via email to