VRRP - limitations on distance or latency - TCP-IP

This is a discussion on VRRP - limitations on distance or latency - TCP-IP ; Is there a limitation on VRRP with respect to latency or distance? RFC 2338 doesn't seem to address this. For example, could buildings connected by fiber host a VRRP pair? Is the limit the 9.6 microseconds interframe gap of ethernet ...

+ Reply to Thread
Results 1 to 3 of 3

Thread: VRRP - limitations on distance or latency

  1. VRRP - limitations on distance or latency

    Is there a limitation on VRRP with respect to latency or distance?

    RFC 2338 doesn't seem to address this. For example, could buildings
    connected by fiber host a VRRP pair? Is the limit the
    9.6 microseconds interframe gap of ethernet or some other factor?

    alan

  2. Re: VRRP - limitations on distance or latency

    On Aug 15, 9:04 am, pale...@sonic.net (Alan Strassberg) wrote:

    > Is there a limitation on VRRP with respect to latency or distance?


    No. Why would there be?

    > RFC 2338 doesn't seem to address this. For example, could buildings
    > connected by fiber host a VRRP pair? Is the limit the
    > 9.6 microseconds interframe gap of ethernet or some other factor?


    Why would that matter?

    DS


  3. Re: VRRP - limitations on distance or latency

    "David Schwartz" wrote in message
    news:1187218910.075478.181910@l22g2000prc.googlegr oups.com...
    > On Aug 15, 9:04 am, pale...@sonic.net (Alan Strassberg) wrote:
    >
    > > Is there a limitation on VRRP with respect to latency or

    distance?
    >
    > No. Why would there be?


    You may hit issues if there are interruption in smooth flow of traffic so
    that VRRP times out - but spanning tree can do that within a building, never
    mind over a WAN.

    the main limitation is that the protocol doesnt detect 1 way links, split
    subnets, or faults where a physical line down doesnt get signalled back to
    the router (ie if you have some L2 switches between the VRRP sender and a
    user device).

    This is really an issue with the router not seeing a port down, so still
    delivers traffic to an interface that cannot get to an end device because
    your subnet is split into 2 or more disconnected sections.

    So - it is good practice to avoid using VRRP on its own over multiple hop L2
    links or significant distance (or at least design & test to make sure these
    issues dont hurt you).

    >
    > > RFC 2338 doesn't seem to address this. For example, could

    buildings
    > > connected by fiber host a VRRP pair? Is the limit the
    > > 9.6 microseconds interframe gap of ethernet or some other

    factor?
    >
    > Why would that matter?


    The minimum hello timer in the original standard was 1 sec, so a few uSec
    are not going to be an issue.

    some implementations may use sub second timers, in a similar way to the
    changes Cisco added to HSRP to improve it a while back.
    >
    > DS
    >

    --
    Regards

    stephen_hope@xyzworld.com - replace xyz with ntl



+ Reply to Thread