Routing

 View Only
last person joined: 3 days ago 

Ask questions and share experiences about ACX Series, CTP Series, MX Series, PTX Series, SSR Series, JRR Series, and all things routing, including portfolios and protocols.
  • 1.  MX80 IPv6 and BGP considerations

    Posted 17 days ago
      |   view attached
    Hi all,

    in our network we have 2 MX80 routers that are connected to 3 ISPs and a local internet exchange(one is also connected to VIX). From ISPs we are getting full IPv4 BGP routing tables and from internet exchanges the networks from the peers connected there.

    Now we are considering requesting an IPv6 range for our AS and bringing IPv6 into our network.

    I would like to know what considerations for BGP and memory we should watch out for besides the obvious memory and CPU? I was thinking of only getting the default IPv6 route from our ISPs and then try to load balance the outgoing traffic as getting large tables just doesn't seem viable.

    For reference I'm adding the output from the router connected to 3 ISPs, VIX and a local internet exchange(show system processes extensive and show chassis routing-engine) in the attachment.

    Thanks for all the help.

    Attachment(s)

    txt
    MX output.txt   3 KB 1 version


  • 2.  RE: MX80 IPv6 and BGP considerations

    Posted 16 days ago
    Yes, MX80 are not usable for full routing those days. And if an upstream flaps, prepare to wait half an hour before convergence. If they don't crash.
    Actually the classical way to go with such gears is to receive a default only from upstreams, and peering routes from IXes, both for IPv4 and IPv6.
    The best way would be to get a pair of MX204 of course.

    ------------------------------
    Olivier Benghozi
    ------------------------------



  • 3.  RE: MX80 IPv6 and BGP considerations

    Posted 16 days ago

    The thing is we are a year or so away from getting new gear. We get full routes from our upstreams as we do some load balancing and we need specific routes.

    I was thinking if I get only def. IPv6 routes from our upstreams I could do load balancing with them for IPv6 and keep our existing IPv4 setup.

    I assume that that would be the least stressful way to implement IPv6 on our equipment.




  • 4.  RE: MX80 IPv6 and BGP considerations

    Posted 16 days ago
    But IPv4 DFZ increases every day, even without adding IPv6 DFZ. So without even doing anything, you are getting closer every day to a critical situation. Actually it's already critical (to my mind).
    You might want to get default + specific from your upstreams, accept only default and a selection of more specific (by using policies with som as-path- group), using «keep none» to avoid leaving the other routes in memory, and you would be in better shape.

    ------------------------------
    Olivier Benghozi
    ------------------------------