ARP vs. static host route

Discussion in 'Cisco' started by Maraudius Kachingo, Nov 26, 2003.

  1. If there is an ARP entry for an IP address, is it always
    used before consulting any IP routing tables?

    Most routing explanations explain that an IP address is
    mapped to an interface or next hop, and then the ARP table
    is consulted.

    Not true in implementations, correct? A match in the ARP
    table, by whatever means it got there, stops further processing.
     
    Maraudius Kachingo, Nov 26, 2003
    #1
    1. Advertisements

  2. Maraudius Kachingo

    Chris Guest

    If the routing table matches a route for a given destination then the
    packets can be put on the output interface buffer for delivery to the next
    hop. The router will then need to look at the ARP table to find the MAC
    address of the next hop IP so that the packets can be encapsulated in a
    layer 2 frame and put on the wire. If it has no cached address then it will
    make an ARP request to that IP address. This is all done when an IP routing
    decision has already been made.

    Chris.
     
    Chris, Nov 26, 2003
    #2
    1. Advertisements

  3. :If there is an ARP entry for an IP address, is it always
    :used before consulting any IP routing tables?

    :Most routing explanations explain that an IP address is
    :mapped to an interface or next hop, and then the ARP table
    :is consulted.

    :Not true in implementations, correct? A match in the ARP
    :table, by whatever means it got there, stops further processing.

    Because of Policy Based Routing (PBR), the route to a host
    might not go directly to the host even when it is in the ARP
    table directly, so the routing table -should- always be consulted.
     
    Walter Roberson, Nov 26, 2003
    #3
  4. This happens with any route. Consider a connected network
    192.168.1.0/24 on fa0 and a static route for 192.168.1.1 via
    some hop at fa1. The packet will exit at fa1 as routes demand.

    Routes and macs are two different things. First the packet is forwarded
    towards some output interface based on routing table (layer 3 info) or
    other routing information such as policy-routing.

    At the output interface a suitable layer-2 info (for example mac from the
    arp table) is looked up before forwarding at layer 2. On ethernet segment
    the packet is finally destinated to a layer-2 mac-address.

    Some switching methods like cef and older route-cache has all the previous
    info cached or stored in a quicker format. There the information is often
    readily combined for a faster lookup. The information is based on the
    route tables and arp tables originally. The forwarding information is
    actually
    made based on the cached info or the data in a quicker storage method.
     
    Harri Suomalainen, Nov 27, 2003
    #4
    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.