Re: RFC 1226

From: Tomi Manninen (rkde@biota.com.au)
Date: Mon Mar 01 2004 - 17:34:48 EET

  • Next message: Leo Bistmans: "Re: RFC 1226"

    On Mon, 2004-03-01 at 03:21, Nick Schmalenberger wrote:

    > A few weeks ago, I sent this message to Brian Kantor, the author of RFC
    > 1226. I haven't gotten any response, so maybe you can answer my
    > questions. Also, in the AX.25-HOWTO, it talks about these network layer
    > protocols like NETROM and ROSE, and I can't understand what they are for.
    > Why not just use IP? In the database of Linux Ham radio applications, it
    > has a fileserver for AX.25. Why not just use the existing facilities in
    > Linux? What is so special about radio that all this stuff needs to be
    > reinvented incompatibly?

    Radio does indeed pose problems that are not present in land line
    networks. Mainly the high packet loss (due channel access, collision
    detection, interference and propagation being much more challenging)
    is a problem.

    Ham radio poses even more problems. It's next to impossible to build
    a properly co-ordinated network so any L3 protocol should have an
    automatic and quickly converging routing protocol on top of it.

    When AX.25 was devised roughly 20 years ago they didn't specify
    any L3 protocol. Possibly because of the limits posed by the
    available (to the average ham) processing power and memory.
    This resulted in the development of many rivalling upper level
    schemes which is partly a good thing, partly a bad thing.

    Anyway I think TCP/IP was a late entrant to the scene, required
    relatively big amount of computing power, had a high overhead
    and didn't have a working routing protocol. And then there
    probably were social reasons why IP over AX.25 never made a
    big breakthrough.

    And one has to remember, 15 to 20 years ago TCP/IP didn't
    have nearly the status of de-facto standard of networking like
    it has now. The Internet of today didn't really exist...

    > > I read your RFC 1226, and I wonder why would an AX.25 frame be
    > > encapsulated in an IP datagram?

    Because of the above, ie. the fact that we still largely use AX.25
    with our own upper layer protocols and even without any L3 protocol,
    there still is a need for encapsulating AX.25 in IP frames for
    tunneling over the Internet. Hence RFC1226.

    > > Why
    > > doesn't APRS use IP and UDP? Thanks.

    The APRS "protocol" is basically a collection of standard text
    messages that are sent out as broadcasts and propagated in the
    "network" in a simple flood like fashion. I don't see much
    benefit in adding the overhead of IP/UDP...

    (I hope the above makes sense to those that have been around
    longer than me. I have only about 12-13 years of experience
    in packet radio...)

    -- 
    Tomi Manninen / OH2BNS / KP20ME04
    

    - To unsubscribe from this list: send the line "unsubscribe linux-hams" in the body of a message to xtjs@evva.com More majordomo info at http://vger.kernel.org/majordomo-info.html



    This archive was generated by hypermail 2b30 : Mon Mar 01 2004 - 17:35:17 EET