Hi Joe Please see inline with tag <Saumya>
Thanks Saumya. On 7/2/15, 12:15 AM, "Joe Touch" <[email protected]> wrote: > > >On 7/1/2015 11:38 AM, Saumya Dikshit (sadikshi) wrote: >> Hi Joe, >> >> Vxlan is a Layer-2 tunnel and carries layer-2 frame(complete packet >> including l2 header) as payload. > >If that is correct, then it has no business interacting with ICMP. <Saumya> That¹s what we intend to solve here. For a typical data center deployment, where there is need/clamour to save an iota of resource units. Hence digging in deep and checking for L2 packet payload for an IP|ICMP or IPv6|ICMPv6 encap and convey it to end point may do some good to that. Imagine a web server application trying to reach out to its storage Node (blade server) in a cloud network over a vxlan network, in a time critical operation. > >It should be supporting the minimum required transit MTU for IPv4 or >IPv6 (respectively), and where that's not directly possible it must >support fragmentation and reassembly on its own. > > >... >> Secondly, Vxlan Gateway can potentially act as a Layer-2 gateway and >> Layer-3 gateway for client devices >> connected over same or different subnet respectively. For Layer-2 >>gateway >> case, MTU derivation may not >> make sense, as the destination is one L3-hop away. > >IP hops are measured by the number of routers, not link. I.e., >traversing one hop is one router, not one link. See RFC1812. <Saumya> From a deployment perspective or rather client perspective, two devices with same subnet are viewed (although virtually) to be directly Connected over an L3 network, although there can be L2 switches (Vxlan gateways) which can potentially bridge the packets between them, as devices will be part of same bridge domain (vni). > >Although I appreciate you're trying to optimize, ICMP PTBs are for "I >can't carry this message at all", not "I wish it were smaller". If you >try to interpret the PTBs incorrectly, you'll only create black holes. <Saumya> Blackholing will anyway happen without this solution. All we are looking for a IP|ICMP or Ipv6|ICMPv6 encap in l2 payload and working on it further, in Vxlan gateway. > >Further, the use of PMTUD is deprecated in favor of PLMTUD (RFC4821), >precisely so the path properties to not rely on ICMPs - which are often >blocked anyway. <Saumya> This solution can be viewed as an enabler for PLMTUD if required on any end-point device connected to vxlan network and not the Vxlan gateways themselves. >Again, these comments are NOT intended as issues to be "fixed" with >document edits. > >Joe _______________________________________________ nvo3 mailing list [email protected] https://www.ietf.org/mailman/listinfo/nvo3
