ATM - TCP-IP

This is a discussion on ATM - TCP-IP ; hello I know close to nothing about ATM, but my book mentions it briefly. Anyways, among other things it also claims that ATM networks need special hardware in order to work. But somewhere else the book says that ATM FORUM ...

+ Reply to Thread
Results 1 to 5 of 5

Thread: ATM

  1. ATM

    hello


    I know close to nothing about ATM, but my book mentions it briefly.
    Anyways, among other things it also claims that ATM networks need
    special hardware in order to work. But somewhere else the book says
    that ATM FORUM recommends FDDI and OC3 SONET arhitecture for ATM
    network.

    1)
    First of all, how can you first say that ATM network needs special
    hardware and then recommend FDDI hardware for ATM network?

    2)
    When ATM forum recoomends FDDI, does it recommend just hardware side
    of FDDI, or one also needs to implement part or all of FDDI protocols
    for ATM network to work?

    3)
    What is it about FDDI ( or OC3 SONET ) hardware, that it makes it
    possible for ATM protocols to run on top of FDDI, while Ethernet
    hardware can't implement ATM protocols?


    thank you

    cheers


  2. Re: ATM

    wrote:

    > I know close to nothing about ATM, but my book mentions it briefly.
    > Anyways, among other things it also claims that ATM networks need
    > special hardware in order to work. But somewhere else the book says
    > that ATM FORUM recommends FDDI and OC3 SONET arhitecture for ATM
    > network.


    The ATM Forum is now the MFA forum.

    http://www.mfaforum.org/

    They joined up with the Frame Relay and the MPLS Forums (fora, I should
    say).

    > 1)
    > First of all, how can you first say that ATM network needs special
    > hardware and then recommend FDDI hardware for ATM network?


    The only thing FDDI is, as an option, the FDDI transceivers. So they use
    FDDI's 100 Mb/s and 4B/5B signaling scheme, but that's it. Cell
    switching is entirely ATM, based on VCI and VPI in the cell header.

    > 2)
    > When ATM forum recoomends FDDI, does it recommend just hardware side
    > of FDDI, or one also needs to implement part or all of FDDI protocols
    > for ATM network to work?


    Just the optical transceivers and signaling. I.e., layer 1 stuff.

    > 3)
    > What is it about FDDI ( or OC3 SONET ) hardware, that it makes it
    > possible for ATM protocols to run on top of FDDI, while Ethernet
    > hardware can't implement ATM protocols?


    ATM's original reason for being was to be the architecture for Broadband
    ISDN (BISDN). As such, it was primarily expected to run over the
    isochronous SONET network the telcos were then deploying. And to use
    these SONET "pipes" efficiently when carrying bursty data traffic, while
    still capable of carrying voice and other circuit-oriented
    communications, such as H.320 videoconferencing or even HDTV.

    BISDN was expected to become the global infrastructure for multimedia
    communications. The time frame was late 1980s and early 1990s.

    The QoS promises ATM made were based on having this SONET
    infrastructure. Using the FDDI transceivers, as far as I could tell, was
    just a low-cost way of building local ATM nets. At the time, Ethernet
    was still limited to 10 Mb/s, and Fast Ethernet was just starting to
    become available closer to the mid 1990s. ATM was originally meant for
    at least 155 Mb/s, although later versions of the standard had options
    for as low as 25 Mb/s, if I remember right.

    Anyway, IMO, the reason ATM never became what the original visionaries
    thought is that IP became the global protocol for multimedia
    communications. Things like RTP/RTCP, around 1993 or 1994, and faster
    versions of Ethernet, all conspired to make this BISDN idea obsolete.
    Also, the circuit orientation of ATM turned out to be problematic. Hard
    to scale.

    Here's some text that puts ATM in a good light.

    http://www.mfaforum.org/education/3.shtml

    Bert


  3. Re: ATM

    thank you

    cheers



  4. Re: ATM

    kaja_love160 wrote:

    > I know close to nothing about ATM, but my book mentions it briefly.
    > Anyways, among other things it also claims that ATM networks need
    > special hardware in order to work. But somewhere else the book says
    > that ATM FORUM recommends FDDI and OC3 SONET arhitecture for ATM
    > network.
    >
    > 1)
    > First of all, how can you first say that ATM network needs special
    > hardware and then recommend FDDI hardware for ATM network?
    >
    > 2)
    > When ATM forum recoomends FDDI, does it recommend just hardware side
    > of FDDI, or one also needs to implement part or all of FDDI protocols
    > for ATM network to work?
    >
    > 3)
    > What is it about FDDI ( or OC3 SONET ) hardware, that it makes it
    > possible for ATM protocols to run on top of FDDI, while Ethernet
    > hardware can't implement ATM protocols?


    On a somewhat related note, network-related article on Wikipedia can
    (sometimes) be informative.

    http://en.wikipedia.org/wiki/Asynchronous_Transfer_Mode

  5. Re: ATM

    On Jun 25, 2:04 pm, "Albert Manfredi"
    wrote:
    > Anyway, IMO, the reason ATM never became what the original visionaries
    > thought is that IP became the global protocol for multimedia
    > communications. Things like RTP/RTCP, around 1993 or 1994, and faster
    > versions of Ethernet, all conspired to make this BISDN idea obsolete.
    > Also, the circuit orientation of ATM turned out to be problematic. Hard
    > to scale.
    >
    > Here's some text that puts ATM in a good light.
    >
    > http://www.mfaforum.org/education/3.shtml


    I am going to nit-pick a bit here.

    You points about why ATM never became what some thought it would, are,
    of course, valid. However, I find it mildly disturbing when someone
    points out specific reasons why one technology superseded another.
    Hindsight is 20/20, and often, the reasons given have to do with
    developments that occur *after* the two technologies began competing.
    What bothers me is that giving reasons after the battle has already
    been fought vaguely implies that it could not have been foreseen which
    technology would have won at the beginning of the race.

    This latter point, is what I disagree with.

    I *DO* think it is possible for someone (or some people) to see, right
    at the beginning of the race, that ATM never had a chance compared to
    Ethernet and friends.

    Someones I think it is better to say,

    "All things considered, Ethernet was better."

    -Le Chaud Lapin-


+ Reply to Thread