Inverse ARP behavior

Hi All,

I have some doubts on inverse ARP. I have read somewhere that it is not sent by default on subinterfaces (http://blog.ine.com/tag/inverse-arp/). Is this true? Can anyone explain how is the default setting for inverse ARP.

 

Thanks before

Comments

  • If an IP address and DLCI is assigned then Frame Relay Inverse ARP requests will be sent. 

    Brian McGahan, CCIE #8593 (R&S/SP/Security)

    Internetwork Expert, Inc.

    On May 18, 2011, at 10:42 PM, "prima_even" <[email protected]> wrote:

    Hi All,

    I have some doubts on inverse ARP. I have read somewhere that it is not sent by default on subinterfaces (http://blog.ine.com/tag/inverse-arp/). Is this true? Can anyone explain how is the default setting for inverse ARP.

     

    Thanks before




    INE - The Industry Leader in CCIE Preparation

    http://www.INE.com



    Subscription information may be found at:

    http://www.ieoc.com/forums/ForumSubscriptions.aspx











  • My understanding is that sub interfaces will not *initiate*
    inARP requests but will respond to an inARP request with an inARP reply.

     

    Therefore inARP will work if there is a sub interface at one end
    and a normal interface at the other but not if both ends are sub interfaces.

     

    Cheers,

     

    -Aaron.

     

    From: [email protected] [mailto:[email protected]] On
    Behalf Of
    prima_even

    Sent: Thursday, 19 May 2011 11:42 AM

    To: [email protected]

    Subject: [CCIE R&S] Inverse ARP behavior

     

    Hi All,

    I have some doubts on inverse ARP. I have read somewhere that it is not sent
    by default on subinterfaces (http://blog.ine.com/tag/inverse-arp/). Is this
    true? Can anyone explain how is the default setting for inverse ARP.

     

    Thanks before







    INE - The Industry Leader in CCIE Preparation

    http://www.INE.com



    Subscription information may be found at:

    http://www.ieoc.com/forums/ForumSubscriptions.aspx





  • Hi Aaron,

     

    I agree with you. There are some documentation that says multipoint subif will not send inverse ARP by default, but still reply. So which one is right?

     












  • Yes I do not think the multipoint interface will send inARP
    requests but they will reply.

     

    Lab this up and test with debug commands. I think ‘debug frame
    events’ will show you.

     

    Cheers,

     

    -Aaron.

     

    From: [email protected] [mailto:[email protected]] On
    Behalf Of
    prima_even

    Sent: Thursday, 19 May 2011 2:48 PM

    To: [email protected]

    Subject: Re: [CCIE R&S] RE: Inverse ARP behavior

     

    Hi Aaron,

     

    I agree with you. There are some documentation that says multipoint subif
    will not send inverse ARP by default, but still reply. So which one is right?

     







    INE - The Industry Leader in CCIE Preparation

    http://www.INE.com



    Subscription information may be found at:

    http://www.ieoc.com/forums/ForumSubscriptions.aspx





  • There is no difference in the behavior of multipoint sub interfaces vs the main interface for inarp.  P2P sub interfaces don't send inarp because they don't need it.  

    Brian McGahan, CCIE #8593 (R&S/SP/Security)
     
    Internetwork Expert, Inc.
    Toll Free: 877-224-8987 x 705
    Outside US: 775-826-4344 x 705
    Online Community: http://www.IEOC.com

    On May 19, 2011, at 7:45 AM, "ariemer" <[email protected]> wrote:

    Yes I do not think the multipoint interface will send inARP
    requests but they will reply.

     

    Lab this up and test with debug commands. I think ‘debug frame
    events’ will show you.

     

    Cheers,

     

    -Aaron.

     

    From: [email protected] [mailto:[email protected]] On
    Behalf Of
    prima_even

    Sent: Thursday, 19 May 2011 2:48 PM

    To: [email protected]

    Subject: Re: [CCIE R&S] RE: Inverse ARP behavior

     

    Hi Aaron,

     

    I agree with you. There are some documentation that says multipoint subif
    will not send inverse ARP by default, but still reply. So which one is right?

     







    INE - The Industry Leader in CCIE Preparation

    http://www.INE.com



    Subscription information may be found at:

    http://www.ieoc.com/forums/ForumSubscriptions.aspx




    INE - The Industry Leader in CCIE Preparation

    http://www.INE.com



    Subscription information may be found at:

    http://www.ieoc.com/forums/ForumSubscriptions.aspx
  • Hi Brian,

    I tried this and it seems that you are right (as always[:)]). But I wonder why there are still articles that say multipoint subinterface does not send inARP, but still sends reply (i.e on Scott articles http://blog.ine.com/2008/08/14/that-pesky-frame-relay-interface-dlci-command/).

     

    Regards,

    Prima

  • The short answer is that they’re incorrect J  You can see the packet level InARP conversation by issuing the debug frame-relay events command.

     

    Brian McGahan, CCIE #8593 (R&S/SP/Security)

    [email protected]

    Internetwork Expert, Inc.

    http://www.INE.com

    Toll Free: 877-224-8987 x 705

    Outside US: 775-826-4344 x 705

    Online Community: http://www.IEOC.com

    CCIE Blog: http://blog.INE.com

     

    From: [email protected] [mailto:[email protected]] On Behalf Of prima_even
    Sent: Thursday, May 19, 2011 3:23 PM
    To: Brian McGahan
    Subject: Re: [CCIE R&S] RE: Inverse ARP behavior

     

    Hi Brian,

    I tried this and it seems that you are right (as alwaysSmile). But I wonder why there are still articles that say multipoint subinterface does not send inARP, but still sends reply (i.e on Scott articles http://blog.ine.com/2008/08/14/that-pesky-frame-relay-interface-dlci-command/).

     

    Regards,

    Prima




    INE - The Industry Leader in CCIE Preparation
    http://www.INE.com

    Subscription information may be found at:
    http://www.ieoc.com/forums/ForumSubscriptions.aspx

  • I guess if we break it down simply to P2P Vs P2M it will avoid all confusions. P2M need resolution anyways either static or dynamic. So either you do it using static frame maps or using IARP. P2P dont need any resolution since there is only one way out so we dont need to used INARP or Static maps.

     

    HTH...

    Deepak Arora

    http://deepakarora1984.blogspot.com

     

  • Some of the confusion in this thread may be due to mixing parts of questions and answers.

    "no frame-relay inverse-arp"

    • This command will stop the sending of inverse-arp requests
    • This command will NOT stop the sending of inverse-arp replies

    main/physical interface

    • By default all DLCIs learned via LMI live here
    • DLCIs moved to a subinterface no longer live here
    • By default frame-relay inverse-arp requests will be sent on all DLCIs when:
      • an IP address is configured AND
      • encapsulation frame-relay is configured
    • Broadcast/Multicast is supported when using inverse-arp
    • Broadcast must be explicitly stated in one "frame-relay map ..." statement per protocol/DLCI to support broadcast/multicast

    multipoint subinterface

    • By default no DLCIs live here
    • By default frame-relay inverse-arp requests will be sent when:
      • an IP address is configured AND
      • one or more DLCIs are configured via "frame-relay interface-dlci <DLCI>" or "frame-relay map ip a.b.c.d <DLCI> [broadcast]"
    • Broadcast/Multicast is supported when using inverse-arp
    • Broadcast must be explicitly stated in one "frame-relay map ..." statement per protocol/DLCI to support broadcast/multicast

    point-to-point subinterface

    • By default no DLCIs live here
    • One DLCI is assigned via "frame-relay interface-dlci <DLCI>"
    • Inverse-arp is not used
    • "frame-relay map ..." statements are not permitted
    • Broadcast/Multicast is supported
  • Nicely put Darell,

    Cut and pasted to my notes. :)

    Thanks,

  • Hi all.

    I want to ask a question with InARP. When InARP packet was sent over frame relay network, What is the destination IP address of this packet.

  • According to RFC 1293: http://tools.ietf.org/html/rfc1293

     The format for an InARP request is a follows:

          ar$hrd - 0x000F the value assigned to Frame Relay
          ar$pro - protocol type for which you are searching
                      (i.e.  IP = 0x0800)
          ar$hln - 2,3, or 4 byte addressing length
          ar$pln - byte length of protocol address for which you
                      are searching (for IP = 4)
          ar$op  - 8; InARP request
          ar$sha - Q.922 address of requesting station
          ar$spa - protocol address of requesting station
          ar$tha - Q.922 addressed of newly announced virtual circuit
          ar$tpa - 0; This is what we're looking for


    The destination IP address is unknown - that is what we are trying to find out.

    The InARP request would contain the IP address of the requesting router/station, the Q.922 address (DLCI), and 0 for the destination protocol (IP) address.

    Hi all.

    I want to ask a question with InARP. When InARP packet was sent over frame relay network, What is the destination IP address of this packet.


  • Thanks Darrel for your response

Sign In or Register to comment.