Exclusive Frequencies:  144.930 & 223.660 MHz
Other Common Freqs:  146.745 & 439.025 MHz

TCP/IP is the main interhost communications protocol group of the Internet.  The IP network layer is a level 3 network protocol which can run on top of the AX.25 protocol (among others).  All amateur radio TCP/IP hosts are assigned an IP address of the form "44.X.Y.Z", of which the "X" portion represents the geographic area the host is located in (see below).

Southern California is divided into several IP address spaces ("X" is):

•    6 - Santa Barbara & Ventura Counties
•    8 - San Diego & Imperial Counties
•  10 - Orange County
•  16 - Metropolitan Los Angeles County
•  17 - Antelope Valley (Kern & Northern Los Angeles Counties)
•  18 - Riverside & San Bernardino Counties
•  Any area not listed above

The "Y" and "Z" portions of  any address may be determined by each geographic region.  Each region is free to divide up its assigned address space into further and smaller sub-nets, and in fact, some have and some have not.

How does one obtain an IP address?

There exists an IP address coordinator for each region.  One obtains an IP address by requesting one or more from the respective coordinator.  Some coordinators simply issue the addresses linearly, while others may need more specific data (operating frequency, station location, etc.) in order to issue a properly subnetted address.  Furthermore, for those who will operate more than one host computer system (at one or more stations) will also need to submit a list of  hostnames so that the Internet Domain Name Service can distinguish individual machines.

The following flowchart shows the entire process (borrowed from the Australian master site):

What is my Internet Host Name?

All stations that are part of the amateur radio TCP/IP network will belong to (or have an alias name that belongs to) the domain "ampr.org."  Traditionally, for those amateur radio operators who have only a single on-the-air station, one's callsign is used as the host name.  Those who have more than one host computer will usually use their callsign as a sub-domain and choose various host names under that domain to describe their various machines.

Some machines may have other data interfaces (ethernets, serial ports, non-amateur wireless communications, etc), and therefore have more than one host name - with these other names pertaining to other domains.  This is permissible, and gateways between the ampr.org domain and others are encouraged.  However, the amateur radio ports will follow the above naming convention.

In some regions, logical sub-domains will be named by frequency and/or region as appropriate, thus bearing no callsign in their domain names.  However, since these names do not represent host machines, this convention is not a deviation of the above.  However, any hosts that do exist in these sub-domains can elect to have their host name follow the sub-domain name instead of the otherwise mandatory "<callsign>.ampr.org" format.

What can amateur radio TCP/IP stations do?

TCP/IP amateur radio stations can do anything allowed under the existing FCC Part 97 rules (47 C.F.R), just like any AX.25 BBS station can do.  Additionally, due to the use of the TCP/IP protocol family, TCP/IP stations can also communicate directly with any other host computer that is connected to the Internet.  Such communication, where not using amateur radio hardware, is not restricted by Part 97 content rules, but one must be careful to ensure that no improper content does traverse an amateur radio wireless link.

Many TCP/IP stations exchange mail and other data with the Internet and each other.

As one word of caution, there are some AX.25 BBS stations that refuse to exchange mail or general bulletins with stations that are using TCP/IP gateways (and thus benefit from the higher speed of the Internet).  Some do so because their operators are "purists" and others do so to ensure that in case of an emergency where normal communications methods fail, the amateur wireless digital network will still be in place and viable.  Alternatively, there are others whose operators are glad to take advantage of such higher speed Internet gateways when they are operational.

How is Amateur Radio TCP/IP Different from the Internet?

In most cases, the only difference that the end user will notice is that of speed.  Although many of the host systems are comparably as fast, or in some cases faster, as compared to other Internet hosts, the problem lies with the data throughput rate.  Most amateur radio digital links are of relatively slow data rates (1200 or 9600 baud, as compared to 57.6 kilobaud+) and are half-duplex, meaning that data can flow in only one direction at a time.  The Internet uses bidirectional links (both directions not necessarily of equal throughput) of much higher speed.  Other than this, and the obvious content issue, there is no significant difference.

How about IP packet routing?

In some regions, hosts and their operators have set up dynamic routing based on the RIP-2 Internet routing protocol.  In other areas, the operators have had to enter manual routes into their software routing tables. Obviously, the dynamic routing approach is much better and accounts for the possibility of host equipment failures and changes. The two most common software packages are discussed:

To implement dynamic routing for a "JNOS" host (or any other derrivative "NET" or "NOS" program sourcing from the original by Phil Karn, KA9Q), one must first either compile the source code or obtain a copy of the pre-compiled program that supports the RIP-2 protocol.  Secondly, when configuring the software, one must generally add the following two statements per radio interface to activate such routing:

ARP ADD <subnet-broadcast-IP> AX25 QST <interface>
RIP ADD <subnet-broadcast-IP> 3600 0x16 2

Note:  Some operators feel that they should route their RIP packets only to those other hosts or gateways that they will be using.  However, this defeats the purpose behind using dynamic routing in the first place, but is no worse than maliciously designed manual routing.

To implement RIP-2 dynamic routing for a system running the Linux operating system (or any other sufficiently compatible UNIX OS), one must acquire a program called gated and configure it for RIP-2 routing over the packet radio interface(s).  Since there is extensive help at the gated website and for Linux itself, such setup will not be dealt with here, except to say that make certain that the host knows what the proper subnet broadcast address is and that it's AX.25 callsign maps to "QST".  Note:  Using "routed" (which comes as part of the standard Linux distributions) is not appropriate since it implements RIP version 1 only and will broadcast incorrect subnet masks when communicating with hosts using RIP-2.  Lastly, "gated" will want to send its routing broadcasts every 30 seconds - can this be changed?

There is also a program called RIP98, which supposedly is included with the Linux AX.25 tools.  Although not evaluated by the SCDCC (yet), it's possible that this program is better suited for wireless routing than gated.

Dynamic routing for a Net/ROM interface is meaningless since:

• NetROM itself contains a dynamic routing protocol, and
• There is no ability to "broadcast" information using NetROM.