NMEA OneNet

[laika]

For those that missed the memo last year, it's the NMEA's new protocol which runs over ethernet and is supposedly due out next year:
http://www.nmea.org/Assets/nmea%20introduces%20onenet.pdf

The press release is light on detail but the IANA assigned ports list has UDP/10111 assigned to "NMEA OneNet multicast messaging" and a quick query to the NMEA revealed that it runs over IPv4

Closed, proprietary and currently unfinished protocol so you could say it's too early to tell, but what are the expectations of the forum? Exciting and forward thinking, unnecessary cynical marketing ploy, or looking dated before it is released likely to look like 2 decade old technology before it gains widespread acceptance?

 

[AngusMcDoon]

YAPP Seatalk to OneNet converter hitting the forum shelves any day now...

 

[prv]

Well, most people (including some manufacturers) haven't caught up up with NMEA 2000 yet...

Still, it's good to use standard technology. If it runs on IP (the fact that it's not IPv6 is unlikely to matter on an instrument network) then presumably one could use standard switches, connect directly to computers with no additional hardware, and plug in things like raspberry pis and arduinos with no low-level YAPPery. Just open a socket in your high-level language of choice and away you go.

Pete

 

[ianj99]

In the meantime, if you want to transmit NMEA by ethernet or wifi, you can use the free download of Navmonpc running as a server on the laptop connected to the nmea datastream and other devices as clients - Navmonpc or there's probably no reason why tablets and the like could not use the data over wifi.

 

[laika]

Pete: Agree with everything you say apart from the IPv6 part: IPv6 is pretty much ideal for a network of instruments like this. Stateless address autoconfiguration, greater opportunity for resource reservation if (as the blurb suggests) real time transmission is a requirement (without running a second N2K backbone). Never mind the question of what the IPv6 vs IPv4 landscape will look like by the time OneNet gains widespread market penetration. Given control of the complete stack, I'd be interested to know why the NMEA think they can't do reliable transmission over ethernet (Lossless Ethernet? higher level protocols?). A lot of similar problems have been addressed by multicast market data systems. I'm sure the OneNet committee have their reasons, but I wonder if the manufacturers are still thinking in terms of electronics engineering rather than computer science.

Inaj99: the current de-facto (non-) "standards" for NMEA-0183 payloads over IP (using udp broadcast or unicast or tcp) are handy but not, err, "standardised". I believe there *is* a standard for NMEA-0183 data over udp: IEC 61162 part 4, but as it costs CHF280 for a butcher's I am unaware of anything more than the (publicly available) contents pages which imply that it involves multicast UDP over IPv4. Nothing uses this AFAIK, but it might well be a basis of OneNet.

This closed commercial protocol malarkey is a rum business.

 

[laika]

For the handful of people here who are interested, the NMEA have changed their minds and onenet is now IPv6 based according to comments on the end of this panbo article:
http://www.panbo.com/archives/2012/08/onenet_nmea_finally_creates_a_marine_ethernet_standard.html

Now they just need to adequately address security and I can stop ranting about sending electronic engineers to do a computer scientist's job and focus on the evils of closed "standards". Like I said, it's a rum business this closed commercial protocol malarkey.

Oh and if anyone from Simrad, Raymarine etc. is reading...author of the only marine data multiplexer I know of to support IPv6 available for hire. Unreasonable rates but speaks good English.

 

[Boreades]

This closed commercial protocol malarkey is a rum business.

Agreed. Obviously the Marine Electronics industry is trying to protect its own interests. But it looks increasingly futile. We've already got OpenCPN and other Open Source software initiatives. Next, the hardware.

Maybe we should start an Open Marine group, to develop and promote more DIY/cheap/free hardware and software products?

 

[laika]

Agreed. Obviously the Marine Electronics industry is trying to protect its own interests. But it looks increasingly futile. We've already got OpenCPN and other Open Source software initiatives. Next, the hardware.

Maybe we should start an Open Marine group, to develop and promote more DIY/cheap/free hardware and software products?

Actually "Open Marine Group" (preferably followed by an "!", or for greater impact, possibly "Open Marine Freeware Group" (!)) was exactly the name I had been thinking of.

I could bore the pants off people with this subject, but in a word, "yes".

("Open Marine Group Protocol for Operation of Navigation Information Electronic Systems" if we want to encourage diversity in the industry)

 

[AndyJ76]

("Open Marine Group Protocol for Operation of Navigation Information Electronic Systems" if we want to encourage diversity in the industry)

OMG PONIES! Love it