Someone brought up the UDR56K-4 radio announced at Dayton Hamvention® a couple of weeks ago. So let me list some parameters and answer a couple of questions:

• The model announced is specified as a half duplex radio, running 25W on 70cm band.
• It has an integrated computer running a standard Linux server (Debian Squeeze) with additional drivers specific to the radio.
• The software platform is open source and developers are free to create their own protocols and applications
• Additional hardware and options can be added via USB or 10/100 Ethernet.
• There will be a daughter card option to hold a vocoder (AMBE family card will be available from NW Digital Radio)
• AX.25 and D-STAR network stacks are under development and test, supporting applications that use those network stacks.
• Modulation choices include GMSK, FSK, 4FSK, and possibly one or two additional choices. The modulation is performed in a high integration RF IC which talks through a bus directly at digital logic levels (no analog/base-band layer)
• The RF IC can accept up to 100 Kbps with bandwidth profiles based on modulation choice. The US regulations limit operation to 56 Kbaud and 100 KHz bandwidth.  Higher bit rates that meet those requirements can be supported up to the 100 Kbps rate. The developer/operator is responsible for insuring data rate / modulation choice meet the regulations where the radio is deployed. 

Having addressed the basics relative to the radio.  The idea is to have a platform to support multiple applications including AMPRnet.  There is an interest in other common air interface (CAI) protocols for passing various data including Net-44 packets, we have at least one 3rd party developer who wants to add a low-overhead IP transport to the radio.

Yahoo! Forum for discussions relative to the UDR are at http://groups.yahoo.com/group/UniversalDigitalRadio

One thing to remember is that for a given power one must choose between distance, reliability, and speed.  For example, using 4FSK it may be possible to push the raw bit rate above 80 Kbps within 100 KHz, however it might be worth using a forward error correcting code reducing the effective rate back to 56 Kbps while reducing retries on a given path due to bit errors.

I think that Net44 as a replacement for broadband service is not a goal that we should be seeking.  In the wireless space there are plenty of consumer devices (Part 15 in the US) that can provide high-speed/local service, e.g. WiFi, or even point-to-point links on some paths.

There are a lot of applications where repurposing consumer devices makes sense, e.g. HSMM on amateur bands, or very local mesh networks in urban areas.  Conversely, there are many applications where longer haul, higher power, modest data rate service is a better choice and this is where Amateur Radio can fill a need. 

I agree we should work to multi-home Net-44 to the rest of the Internet, but I don't believe we should do this on a LAN level.  I think at most it should be the 256 /16 subnets and probably many fewer than that.   Local LANs can tunnel using newer tunneling protocols that can use a dynamic IP on at least one end of the tunnel to these BGP'ed routers (with fail-over between them for tunnels from LANs).  There is no reason to use roll-your-own routers or protocols -- there are very capable routers with tunneling in the US$60 range that can get the connection over the Internet tunnel to the local LAN, then use our various RF interfaces (WiFi/HSMM, AX.25, DSTAR, ...) out to individual nodes.  A did a presentation on this subject earlier this spring -- http://www.microhams.com/digitalconf2012/K7VE_N7IPB_RebootNET44.pdf 

---