84 comments

[ 15.1 ms ] story [ 281 ms ] thread
First I was excited, then I saw that it's based on DDS.
why the hate on DDS?

seems like a perfect fit.

DDS is the least bad of the stuff that has come out of OMG, but it's not a nice middleware still. There is no underlying algebra of that composes nicely when compared to Zenoh for example, and the "vendor independence" and compatibility is usually a joke. There's a reason ROS is moving away from it again.
There's a standard that has OMG as the acronym? OMG. I'm not even going to bother to search for that!
Object Management Group I think. Better known as the inventors of UML.
UML was phase two of OMG. It originally developed CORBA.
"The horror, the horror."
Forgive the ignorant question, what made CORBA terribile?

I only used it during university and it seemed overengineered but it mostly worked, and I vaguely recall IDL to be nice.

I'm guessing interop was actually atrocious?

Synchronisation based on spooky action at a distance of mutable objects.
CORBA had the concept of distributed objects which is a terrible idea in practice.
Also BPMN, CORBA and UPOS/JavaPOS.

It's funny that you can legitimately just add " - OMG!" in the slang acronym meaning after any of its creations' names.

They also merged the concepts of Working Group and Task Force into an Open Management Group Working Task Force, or OMG/WTF.
I cannot find boomstick anywhere in here
Am I just too ignorant of Miltech, or is this that in large parts recreating Standard messaging primitives that other protocols and systems have been providing for decades already (so invented elsewhere?)
DDS sort of took over from CORBA in the military domain, and the statically-typed IDL was still quite similar to CORBA last time I looked at it, so that takes you back to around 1998 when CORBA was being used a lot in the military in my experience.

What other protocols and systems were you thinking of that can provide (depending on broker) hard-realtime delivery that can be used in safety-critical systems?

Stupid question, but why didn’t the DOD mandate IP compatibility for literally all comms like 20+ years ago? Is it pure corruption/incompetence or is there a good reason?
The DOD effectively can't "just mandate" things and it's because of politics (which is a superset of corruption and incompetence).

There's a very small number of major contractors and if you pick one of their APIs as the new standard, then you disadvantage all the others.

If you pick a greenfield project, then it will take a long time, and once you are done everybody will claim that they can't use it (see e.g. Ada; quick version they started working on a mandate for a programming language in 1975, mandated Ada in 1991 and by 1996 removed the Ada mandate).

A lot of military comms are point-to-point or broadcast between radios where TCP/IP doesn’t really make sense. On top of that, you need to understand how much legacy military networking stuff is out there. This stuff long predates the internet. We had tactical data links and target tracking where a ground station could receive a target info from a remote radar array, and automatically direct fighters on autopilot on an intercept course with enemy contacts in like 1960 (see SAGE as an example). These systems have been around for decades and been continually improved on. There’s a lot of inertia to maintain compatibility. Because you really don’t want someone to die because your ally tried to warn you but their tactical data link isn’t compatible with yours.
actual or perceived realtime / determinism requirements
Reminds me a lot of the Tactical Microgrid Standard (aka TMS aka MIL-STD-3071) [1], probably just because TMS uses DDS as well. I would really like to know if there is a protocol that functions like DDS but caters to real-time guarantees and prioritizes (at a "simple protocol" level) usability on embedded systems with no dynamic memory allocation. It would also need to be just-as-functional with non-real-time systems. One problem with DDS is that it is too heavy-handed to implement well on an embedded system.

[1] https://battery.army.mil/system-integrator-hub/tms/

Pretty sure there are DDS implementations that provide most of that, maybe with memory pools instead of no-allocations.

Other than that, PX4 uses an in-memory only pub/sub for internal data bus, somewhat inspired by DDS

I should have also mentioned that my appetite also includes not being locked into one vendor's software stack. For instance, there are nice offerings from Eprosima [1] and Twin Oaks [2] that do this, but they seem to be "outstanding in their field" and would be difficult to replace if a need to do so developed. In the case of Eprosima's offerings, they also make some architectural assumptions that are difficult to maintain for some applications (such as having a paired Linux device with each embedded device acting as a translator).

But yes - there absolutely are solutions already. I guess I was just lamenting that they weren't really functionally accessible for all applications.

[1] https://micro-xrce-dds.docs.eprosima.com/en/latest/

[2] https://www.twinoakscomputing.com/

There's a range of DDS implementations, definitely a number that are compatible with embedded and spaceflight applications that don't require dynamic memory allocations. You're going to take a hit on some DDS features, but then again you aren't going to need to handle a large queue of large dynamically-sized messages on some small embedded platform.

I think it's rarely the right choice - most of the hardware I've seen that uses embedded DDS probably should have used plain old UDP instead and then whatever client was running on beefier hardware then handled the translation into DDS - but there are shipping products that use embedded DDS.

Zenoh (https://zenoh.io/) is a good alternative to DDS in many cases - it scales down to microcontrollers with most features still intact, but scales up to a routed hybrid mesh if needed.

It does move the schema enforcement out of the middleware layer, so it’s a bit different. But it’s now a first-class alternative middleware implementation for ROS2 as an alternative to DDS.

Wow - thanks for pointing this out! I will be looking into it.
TIL there are people using DDS for non-realtime
This looks like ROS2 for missiles.
This is so ROS2 that ROS2 can't even tell that its not ROS2
Nah, they haven't open sourced anything: https://github.com/rheinmetall. They just published the documentation. Really weird, what's the point?
They've got the IDL files as well for the APIs, but you can't use them without their libs:

> The data model is defined using a custom format .rmodel based on the ddkit > framework (rmodel-api).

> Please note that the .rmodel data model is for documentation purposes only and > does not reflect the actual interface used on the communication layer. Network > communication is only possible through the provided libraries.

I think that’s what the “protcol” (sic) in the title was supposed to indicate. I agree it’s misleading.
Theres probably some tender that says it must be a "open platform" "open ecosystem" "modular vendor independent" whatever. So you upload some slop to GitHub and tick the box.
Open the C++ example up and the first thing I see:

```cpp std::cout << "Received ReportCoordinateFrameSettings" << std::endl; ```

Generally, I've heard `std::endl` instead of \n has bad performance because it forces the statement to print immediately. And because cout is guaranteed to flush anyways when the program ends, you're adding a delay for no reason.

I wonder if there's something I'm not seeing that led to them adopting this rule. e.g. "abnormal program termination" caused by the weapon blowing itself up circumventing a buffer flush?

For a long time, std::endl demonstrated that you were a proper C++ programmer, not some neckbeard C apologist. It wasn’t until C++ had „won“ that it became acceptable to use \n again.
Cool I did not know that.

Your comment is a bit salty, but I learned something from it, so I don’t think you deserve the downvotes

> And because cout is guaranteed to flush anyways when the program ends, you're adding a delay for no reason.

I mean... If your command line utility doesn't print anything til it's done, that is going to be, _at best_, an annoying user experience.

> cout is guaranteed to flush anyways when the program ends

This is not always true in the case the program crashes. I've had case where `std::endl` was necessary.

“Hi Codex, please find attached website about the specification of my Battlesuit and use the documented API to create a plugin for Home Assistant that shows groups of sensor data.

You are only allowed to use commands that retrieve data, you are not allowed to perform actions on the suit.

Treat the suit like its API is not robust and in worse shape than a cheap toy: perform read calls slowly and sequentially with at least one minute in between.

If possible please group data in meaningful categories, just like the tank and radar dome.

My Home Assistant is the default on the local network, credentials are in credentials.txt.

Please send me a text when you’re done. Kthnxbye”

"The Kampfanzugskode isn't working well." - "Why can't we use KI to improve it? I've heard from Hans that this Klaus Kode is good at Programmierung." - "We can only use Mistral because of Regulierungen." - "Let's open source it then, so others can use Klaus Kode." - "Ja! This is indeed a legit Schlupfloch and a fantastische Idee!"
(comment deleted)
I'm gonna steal "Klaus Kode", that one really got me.
I sense you'd enjoy the signup process for Klaus Programmieren: https://klausprogrammieren.com/
Having completed our company’s ISO 27001 recertification process recently, this one really hits home with me.
Haha, I had no idea this existed.
oh my god, that's gotta be the best European innovation I've ever seen in a long time

this deserves an EU grant

System is unavailable every 2nd Tuesday?

Was ist das Betriebssystem des Servers? Windows?!?

Haha love it. Long live the fax machine!
"But Wolfgang, could these Regulierungen be umgangen by hosting TiefSuch Blitz on our very own Rechneresysteme?"
> like its API is not robust and in worse shape than a cheap toy

This guy military techs :P

It is a DDS API, so if ROS 2 is anything to go by then it probably doesn't work at all at least 40% of the time.

“Fire your goddamned missiles!.json”

“Think russian^H^H^H xml”

Are they trying to recreate DIS (IEEE1278) and HLA (IEEE1516) — like they‘re replicating the FOM architecture from distributed simulations for their weapons to integrate.
I have not heard about HLA for a looong time !
Corba never died!!! OMG!!
I'm glad it died in GNOME, though. That's where it matters.

Who cares if german company wants to use this mess for their weapons.

of course the germans would use a fucking open management group academic mastrubation, bureucratic shitshow like dds.
They could Just have used ROS...
So what do they expect us to do? Write killer-plugins? Or get excited for next upgrade?

The situation gets out of hand!

(comment deleted)
Finally, Front Mission in real life.
Good. There is no security by obscurity anymore.

The only way to fight those who can just spend more tokens is to have more humans look at it, who also spend tokens.

Oh boy, just what the world needs right now. Open source weapons APIs!
MAVLink has been open source for over a dozen years and with better tooling and apis.
MCP is 100% already in the works.

> Claude, please execute the insurgents. Make no mistakes.

> The user did not specify is family and friends of insurgents are to be killed, too. > Reasoning: Since death of a loved one is a radicalizing event, they will turn into insurgents after the fact with a high likelyhood, about as likely as the insurgent identification was. So... > must kill all humans. Where is Sarah Connor?
I wonder if someone just got super annoyed at having to ask legal to share their documentation with a cooperation partner for export control clearance, that they decided getting it open sourced would be less of a hassle.

Now they can freely link anyone to the document.

This is 100% a valid workflow in DoD/ITAR fields. Suffer once through review, benefit eternally. It's especially important to get one good reviewed spec / presentation out, then all later reviews can be staged w.r.t. that one and are much easier to baseline and pass.

That'll be $1000 consulting fee on ITAR compliance.

Rheinmetal is a nice company, hope it just didn't benefit from war/fear of war.
Complete horrible wrappers. Its almost like they didnt have any real programmers for those languages.

If you cant follow naming standards which are extremely simple what chance has the rest got.