Robert,

Does that mean that the EVPN PE/NVE network will never get an ARP-request/reply 
with duplicate sender's MAC/IP?
If so and I understood you, the proxy-arp function will have to learn the 
IP->MAC bindings for the next-hops, and reply to requests for those next-hops 
but no the anycast loopbacks themselves. If that is the case, there won’t be 
duplicate IPs unless the NIC changes MAC (as Erik mentions), there is a human 
error or there is an ARP spoofing attack. Those cases are/will be covered in 
the draft.

As discussed and mentioned by Erik below, IPv6 anycast can be identified by the 
O flag in the NA messages. We’ll add that to the draft.

Thank you both.
Jorge

From: Robert Raszuk <[email protected]<mailto:[email protected]>>
Date: Wednesday, April 15, 2015 at 2:53 PM
To: Erik Nordmark <[email protected]<mailto:[email protected]>>
Cc: Jorge Rabadan 
<[email protected]<mailto:[email protected]>>, 
Antoni Przygienda 
<[email protected]<mailto:[email protected]>>, 
"Henderickx, Wim (Wim)" 
<[email protected]<mailto:[email protected]>>, 
"[email protected]<mailto:[email protected]>" <[email protected]<mailto:[email protected]>>
Subject: Re: [bess] ARP ND draft

Erik,

How about /32 IPv4 anycast addresses with multiple subnet per linux NIC ? It is 
typical to be able to overload host networking with same anycast loopbacks.

It does not need to be ARP resolved .. the resolution is indirect via connected 
next hops.

Cheers,
R.



On Wed, Apr 15, 2015 at 11:48 PM, Erik Nordmark 
<[email protected]<mailto:[email protected]>> wrote:
On 3/31/15 1:10 PM, Rabadan, Jorge (Jorge) wrote:
Hi Robert and Tony,

As Wim mentioned, ipv6 anycast is something that we will add to the draft in 
the next rev. There is an easy way to know if a given proxy-ND entry belongs to 
an anycast address or not and disable the duplicate IP detection for those.

The challenge for IPv4 is that I don’t see an easy way to learn _dynamically_ 
from access attachment circuits that a given ipv4 is anycast. Even for default 
gateways, if they are integrated in the EVPN PE, we are good, but if they are 
external and connected to a MAC-VRF, it is not so clear how to learn that 
(unless you learn those entries from the management interface).
Jorge,

IPv4/ARP doesn't have any support for anycast address on the same subnet. While 
IPv6/ND has such support (using the O-flag) the common anycast deployment for 
both is to have the anycast instances deployed on different subnets and, in the 
case of DNS servers, in different ISPs.

Thus for IPv4 I think you can assume that the same IP address appearing with 
different MAC addresses is either a duplicate IP address or a case of a host 
having changed the MAC address on its NIC. (I don't know if NIC bonding can be 
configured in a way where it looks like an IP->MAC change each time there is a 
failure; if so that would be a third case.)

   Erik


One of the reasons why we have lots of “SHOULDs” in the draft and not “MUST” is 
because the implementation has to be flexible enough to be configured in a 
different way depending on the use-case, which is one of the points that Tony 
mentions below. In the use-case described at the moment there is no anycast and 
duplicate IP detection is very important. We will add the DC use case in the 
next rev as suggested by Robert and others.
Thanks.
Jorge


From: Antoni Przygienda 
<[email protected]<mailto:[email protected]> 
<mailto:[email protected]<mailto:[email protected]>>>
Date: Monday, March 30, 2015 at 12:12 AM
To: Robert Raszuk <[email protected]<mailto:[email protected]> 
<mailto:[email protected]<mailto:[email protected]>>>, "Henderickx, Wim (Wim)" 
<[email protected]<mailto:[email protected]> 
<mailto:[email protected]<mailto:[email protected]>>>
Cc: Erik Nordmark <[email protected]<mailto:[email protected]> 
<mailto:[email protected]<mailto:[email protected]>>>, 
"[email protected]<mailto:[email protected]> 
<mailto:[email protected]<mailto:[email protected]>>" 
<[email protected]<mailto:[email protected]> 
<mailto:[email protected]<mailto:[email protected]>>>, Jorge Rabadan 
<[email protected]<mailto:[email protected]> 
<mailto:[email protected]<mailto:[email protected]>>>
Subject: RE: [bess] ARP ND draft

    I’m also skeptical whether IP duplicate detection would be a good
    default thing. Especially in case of what I call ‘aliased default
    gateway’ which section 10.1 specifically allows, i.e. default GW
    IP address is same but each PE may use a different MAC when
    advertising it and consequently responses for same IP with
    different ARPs may be seen in the network.  Yes, default GW
    ExtComm is there to differentiate so it can be called an exception
    but nevertheless.

    I also thought a tad about VRRP but I think the IP duplicate
    detection will not apply there, it’s all same IPx->MACx from all
    routers so if anything, it’s more of a MAC move thing.

    Generally I think someone who wants a secure, stable eVPN wants IP
    duplicate detection, someone who runs a very dynamic network with
    tons gateways, possibly anycast & floating IPs will probably not
    be too enamored with it.

    Thanks

    --- tony

    //

    /There are basically two types of people. People who accomplish
    things, and people who claim to have accomplished things. The
    first group is less crowded.
    <http://www.brainyquote.com/quotes/quotes/m/marktwain393535.html>/

    /~~~ Mark Twain
    <http://www.brainyquote.com/quotes/quotes/m/marktwain393535.html>/

    *From:*[email protected]<mailto:[email protected]> 
<mailto:[email protected]<mailto:[email protected]>>
    [mailto:[email protected]<mailto:[email protected]>] *On Behalf Of *Robert 
Raszuk
    *Sent:* Monday, March 30, 2015 1:19 AM
    *To:* Henderickx, Wim (Wim)
    *Cc:* Erik Nordmark; Antoni Przygienda; [email protected]<mailto:[email protected]>
    <mailto:[email protected]<mailto:[email protected]>>; Rabadan, Jorge (Jorge)
    *Subject:* Re: [bess] ARP ND draft

    Hi Wim,

    > There is anycast at IPv4 level for sure but I am not ware this is
    supported at arp level.

    Precisely right. It needs to be documented and addressed if anyone
    is up to proposing automated IP duplicate address detection and
    disabling.

    RFC1546 is rather too old to consider here as solution :)

    Cheers,

    R.

    On Mon, Mar 30, 2015 at 1:10 AM, Henderickx, Wim (Wim)
    <[email protected]<mailto:[email protected]>
    
<mailto:[email protected]<mailto:[email protected]>>>
 wrote:

    To be clear: RFC4861 section 7.2.7 explains the anycast behaviour
    in IPv6.

    I am not aware of such thing at IPv4/ARP level. Do you have a pointer?

    There is anycast at IPv4 level for sure but I am not ware this is
    supported at arp level.

    *From: *<Henderickx>, Wim Henderickx
    *Date: *Monday 30 March 2015 07:38
    *To: *Robert Raszuk
    *Cc: *Erik Nordmark, Antoni Przygienda, "[email protected]<mailto:[email protected]>
    <mailto:[email protected]<mailto:[email protected]>>", Jorge Rabadan


    *Subject: *Re: [bess] ARP ND draft

    At interface level you get dad in most stacks I know.

    Sent from my iPhone


    On 30 Mar 2015, at 06:45, Robert Raszuk 
<[email protected]<mailto:[email protected]>
    <mailto:[email protected]<mailto:[email protected]>>> wrote:

        Hi Wim,

        What makes you say that in IPv4 there is no anycast ? All
        anycase I have played so far is IPv4 :)

        Cheers,

        r.

        On Sun, Mar 29, 2015 at 11:18 PM, Henderickx, Wim (Wim)
        
<[email protected]<mailto:[email protected]>
        
<mailto:[email protected]<mailto:[email protected]>>>
 wrote:

        We will update the draft to highlight the IPv6 anycast
        behaviour better as pointed out by RObert. In IPv4 there is no
        anycast behaviour and as such there should be one option possible.




        On 30/03/15 04:59, "Antoni Przygienda"
        <[email protected]<mailto:[email protected]>
        
<mailto:[email protected]<mailto:[email protected]>>> 
wrote:

        >Yes, but of course I brought it up to show that 'the last one
        simply wins' as suggested by the draft is not enough IMO. A
        good architecture should probably keep track of what it served
        as answer and when the answer is invalid or a new, better one
        exists, provide a GARP.
        >
        >As well, when PE2 sends a newer MAC it may not be a good
        strategy to serve a GARP if PE1's MAC has already been
        offered. That could lead IMO to e.g. gateway chasing problems.
        >
        >--- tony
        >
        >
        >There are basically two types of people. People who
        accomplish things, and people who claim to have accomplished
        things. The first group is less crowded.
        >~~~ Mark Twain
        >
        >
        >> -----Original Message-----
        >> From: Henderickx, Wim (Wim)
        
[mailto:[email protected]<mailto:[email protected]>
        
<mailto:[email protected]<mailto:[email protected]>>]
        >> Sent: Thursday, March 26, 2015 6:01 AM
        >> To: Antoni Przygienda; Erik Nordmark; Rabadan, Jorge (Jorge)
        >> Cc: [email protected]<mailto:[email protected]> 
<mailto:[email protected]<mailto:[email protected]>>
        >> Subject: Re: [bess] ARP ND draft
        >>
        >> For this case you should sent a GARP with the new MAC/IP
        >>
        >>
        >>
        >>
        >> On 25/03/15 18:56, "Antoni Przygienda"
        <[email protected]<mailto:[email protected]>
        
<mailto:[email protected]<mailto:[email protected]>>>
        >> wrote:
        >>
        >> >> > b)It is worth explaining what is suggested behavior if
        eVPN
        >> >> > advertises the same IP with multiple MACs and what
        happens when
        >> >> > e.g. the served MAC vanishes
        >> >> >
        >> >> Doesn't the EVPN RFC already stating that the routes
        would be
        >> >> withdrawn in that case?
        >> >
        >> >The scenario I had in mind was when eVPN PE receives
        >> >
        >> >From PE2 IP1/M1  and  later
        >> >From PE3 IP1/M2
        >> >
        >> >while having answered with IP1/M1 per proxy alrady.
        Additionally, in
        >> >such situation ends up seeing
        >> >
        >> >From PE2  IP1/<no MAC>
        >> >
        >> >So the answer it gave is not valid anymore all of a sudden.
        >> >
        >> >--- tony
        _______________________________________________
        BESS mailing list
        [email protected]<mailto:[email protected]> 
<mailto:[email protected]<mailto:[email protected]>>
        https://www.ietf.org/mailman/listinfo/bess


_______________________________________________
BESS mailing list
[email protected]<mailto:[email protected]>
https://www.ietf.org/mailman/listinfo/bess

_______________________________________________
BESS mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/bess

Reply via email to