Do you use MPLS to transport Internet traffic?

One of the interesting questions I’ve got after announcing the Building IPv6 Service Provider core webinar was “Why would you push IPv6 into the core?

Obviously there are two mechanisms to transport Internet traffic across your (next-generation ;) Service Provider core network: forward it natively or transport it between BGP next-hops across MPLS LSP. It would be really interesting to know how many people use one or the other mechanism, so if you’re working in a Service Provider environment, please respond to the poll (use this link if you’ve disabled JavaScript).

Both transport mechanisms are described in details in my Building IPv6 Service Provider core webinar (don’t forget to register). Conversely, if you’re still not sure why you should have an IP/MPLS backbone, the Market trends in Service Provider networks webinar will definitely persuade you (or your boss).

7 comments:

  1. Hi Ivan,

    correct me if I am wrong, but there is no LDPv6, hence no IPv6 in the core.....

    ReplyDelete
  2. Don't forget about 6PE - IPv6 routes distributed via BGP with label explicit-null, with this you don't need IPv6 support in the core at all.

    ReplyDelete
  3. Ivan Pepelnjak17 March, 2010 19:16

    You're absolutely correct. MPLS is still single-protocol O:-) There is no LDPv6; it also doesn't work with MPLS TE. The only way to make use of MPLS is to use 6PE (as visir indicated).

    ReplyDelete
  4. Ivan
    I am working on the design where in pitching MPLS as core for serving IPv6 internet. This will help SP not to add any IGP for IPv6 traffic.
    I have added a HLD for the same on blog http://www.mplsvpn.info/2010/03/ipv6-internet-in-vrf-over-mpls.html

    Challenges:-
    1. The first challenge is the DNS. In that case we need to use the two interface on DNS one for forwarding and another for reverse.
    2. Proxy.

    I am also working on the solutions. Will come up with new updates :)

    ReplyDelete
  5. This seems to be an OR poll.. The previous provider I worked at did 2 of the options and the third was considered in one of the Sub-AS.

    6PE w MPLS on P Nodes & Natively w BGP on P node. 6VPE considered but not sure of outcome.

    V4 is the same which the v6 topology followed.. In fact, in my experience (And I am pure Core guy ie. no dns/nat/cpes) deploying IPv6 was as easy as another address family.

    In mobile operators I have only seen Internet in VPN and am yet to see v6 just massive NAT boxes (which imo is bottleneck)

    In the end the decision must come down to whether you already have a BGP Free Core or not.

    As a corollary,there is an argument that we buy all the super powered CRS/T1600 etc and then remove BGP from the core. In reality if they are so super (and expensive) there surely adding BGP to the core should be a walk in the park for these big boys ;-)

    ReplyDelete
  6. Hi,

    Anyone here deployed 6PE? I'm asking because I'm trying to deploy it on dynamips and I'm stuck atm with 3640 routers

    ReplyDelete
  7. I had made my topology done on Dynamips with 6PE. Please check my http://www.mplsvpn.info/search/label/IPv6


    regards
    Shivlu Jain

    ReplyDelete

You don't have to log in to post a comment, but please do provide your real name/URL. Anonymous comments might get deleted.

Ivan Pepelnjak, CCIE#1354, is the chief technology advisor for NIL Data Communications. He has been designing and implementing large-scale data communications networks as well as teaching and writing books about advanced technologies since 1990. See his full profile, contact him or follow @ioshints on Twitter.