What is the EIGRP multicast flow timer?

Discussion in 'Cisco' started by Wendy, Dec 1, 2004.

  1. Wendy

    Wendy Guest

    Can someone clear up this ambiguous Cisco exam question once and for
    all?

    What is the EIGRP multicast flow timer?

    A. The timer after which EIGRP stops forwarding multicast data
    traffic.
    B. The timeout timer after which EIGRP retransmits, unicasting the
    neighbor not in CR mode.
    C. The time interval between consecutive received EIGRP hello
    intervals.
    D. The timer interval between consecutive transmitted EIGRP hello
    intervals.
    E. The timeout timer after which EIGRP retransmits, unicasting the
    neighbor in CR mode.

    Camps seem to be split between B. and E. I think it should be B.

    Thanks,
    Rygel
     
    Wendy, Dec 1, 2004
    #1
    1. Advertisements

  2. This is a pretty poor question. I think they are all wrong. This is hardly
    a useful certification question, anyway.

    Researching:

    How does EIGRP work? EIGRP uses a proprietary (and poorly documented)
    Reliable Multicast Scheme for EIGRP, one of several reliable multicast
    schemes. CR (Conditional Receive) is part of Cisco Reliable Multicast.

    When routers become EIGRP neighbors, any nonperiodic state change (eg
    routing table update) will be reliably multicast to the others in the group
    with multicast messages that have sequence numbers. All members of the
    multicast group must unicast an acknowledgement for each packet flow (like
    TCP SEQ/ACK, sliding window). "Flow timer" is how long a sender waits for a
    missing acknowledgement from a group member before the sender concludes that
    the lack of acknowledgement is a problem. The router will then multicast a
    sequence TLV listing the offender.

    So it does not *unicast* the offender at all, but multicasts all future
    messages in that stream with the CR bit set, and listing the offender in a
    sequence TLV, allowing the remaining functional members of the group who are
    now in CR mode to continue consuming the message until the comatose device
    gets booted from the multicast group, or comes back from a coma and plays
    ball.

    So ALL the answers are wrong. "B" is closest if you consider "not in CR
    mode" means "the device who is faling to acknowledge messages." But ALL
    devices are "not in CR mode" until the first sequence TLV is sent, so one
    could argue "E" is correct. However, since no unicast is sent, they are ALL
    wrong. Or at least the question is very poorly phrased.

    http://puck.nether.net/lists/cisco-nsp/5129.html was my main reference.

    Bottom line: Regardless of the right answer, this question sucks and has no
    place on the CCIE exam.
     
    Phillip Remaker, Dec 2, 2004
    #2
    1. Advertisements

  3. Wendy

    chek67

    Joined:
    Jun 13, 2007
    Messages:
    1
    Likes Received:
    0
    chek67, Jun 13, 2007
    #3
    1. Advertisements

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.