Re: BFD over VXLAN: Trapping BFD Control packet at VTEP

["Joel M. Halpern" <[e-mail]>]

What is special about the management VNI is precisely that it is NOT a 
tenant VNI.  The VxLAN administration does know how it allocates VNI to 
tenants, and which VNI it has allocated.  In contrast, it does not know 
which IP addresses or MAC adddresses teh tenant is using or may plan to use.

Yours,
Joel

On 8/2/2019 6:41 PM, Dinesh Dutt wrote:
> The assumption of an IP address within any VNI is suspect that way. 
> What's special about a single VNI, the management VNI? The VTEP IP 
> address does not belong in reality in any VNI.
> 
> Dinesh
> 
> On Fri, Aug 2, 2019 at 3:17 PM Joel M. Halpern <[e-mail] 
> <mailto:[e-mail]>> wrote:
> 
>     Your response seems to miss two points:
> 
>     First, the problem you describe is not what the document says it is
>     solving.  To the degree it discusses it at all, the document says " 
>       In
>     most cases, a single BFD session is sufficient for the given VTEP to
>     monitor the reachability of a remote VTEP, regardless of the number of
>     VNIs in common. "
> 
>     Second, you assume the existence of an IP address for a VTEP within a
>     VNI.  As with the MAC address, the VTEP does not have an IP address
>     within the VNI.  Some implementations may have created such a thing,
>     but
>     the general construct, as defined to date, does not support such.
> 
>     In short, you are requiring a behavior that violates the architectural
>     structure of overlay / underlay separation, and common usage.  And you
>     are doing so to support a use case that the working group has not
>     indicated in the document as important.
> 
>     Yours,
>     Joel
> 
>     On 8/2/2019 5:01 PM, Dinesh Dutt wrote:
>      > Joel,
>      >
>      > You understood correctly.
>      >
>      > The VNIs may not share fate due to misconfiguration. And I strongly
>      > suspect someone will want to use BFD for that because its about
>     checking
>      > path continuity as stated by the draft. As long as there's a
>     valid IP
>      > (because it's BFD) owned by the VTEP in that VNI, you can use BFD in
>      > that VNI. Thats all that you need to dictate.  That IP address
>     has a MAC
>      > address and you can use that on the inner frame. That is all normal
>      > VXLAN processing. The outer IP is always that of the VTEP.
>      >
>      > Dinesh
>      >
>      > On Fri, Aug 2, 2019 at 11:03 AM Joel M. Halpern
>     <[e-mail] <mailto:[e-mail]>
>      > <mailto:[e-mail] <mailto:[e-mail]>>> wrote:
>      >
>      >     If I am reading your various emails correctly Dinesh (and I
>     may have
>      >     missed something) you are trying to use the MAC address
>     because you
>      >     want
>      >     to be able to send these BFD packets over arbitrary VNI to
>     monitor the
>      >     VNI.  That is not a requirement identified in the document. 
>     It is not
>      >     even a problem I understand, since all the VNI between an
>     ingress and
>      >     egress VTEP share fate.
>      >
>      >     Yours,
>      >     Joel
>      >
>      >     On 8/2/2019 1:44 PM, Dinesh Dutt wrote:
>      >      > Thanks for verifying this. On Linux and hardware routers
>     that I'm
>      >     aware
>      >      > of (Cisco circa 2012 and Cumulus), the physical MAC address is
>      >     reused
>      >      > across the VNIs on the VTEP. Did you check on a non-VMW
>     device?
>      >     This is
>      >      > more for my own curiosity.
>      >      >
>      >      > To address the general case, can we not define a
>     well-known (or
>      >     reserve
>      >      > one) unicast MAC address for use with VTEP? If the MAC
>     address is
>      >      > configurable in BFD command, this can be moot.
>      >      >
>      >      > Dinesh
>      >      >
>      >      > On Fri, Aug 2, 2019 at 10:27 AM Santosh P K
>      >      > <[e-mail]
>     <mailto:[e-mail]>
>      >     <mailto:[e-mail]
>     <mailto:[e-mail]>>
>      >     <mailto:[e-mail]
>     <mailto:[e-mail]>
>      >     <mailto:[e-mail]
>     <mailto:[e-mail]>>>> wrote:
>      >      >
>      >      >     I have cross checked point raised about MAC address
>     usage. It is
>      >      >     possible that tenant could be using physical MAC
>     address and
>      >     when a
>      >      >     packet comes with valid VNI with a MAC address that is
>     being
>      >     used by
>      >      >     tenant then packet will be sent to that tenant. This rules
>      >     out the
>      >      >     fact that we could use physical MAC address as inner
>     MAC to
>      >     ensure
>      >      >     packets get terminated at VTEP itself.
>      >      >
>      >      >     Thanks
>      >      >     Santosh P K
>      >      >
>      >      >     On Wed, Jul 31, 2019 at 11:00 AM Santosh P K
>      >      >     <[e-mail]
>     <mailto:[e-mail]>
>      >     <mailto:[e-mail]
>     <mailto:[e-mail]>>
>      >     <mailto:[e-mail]
>     <mailto:[e-mail]>
>      >     <mailto:[e-mail]
>     <mailto:[e-mail]>>>>
>      >      >     wrote:
>      >      >
>      >      >         Joel,
>      >      >             Thanks for your inputs. I checked
>     implementation within
>      >      >         Vmware. Perhaps I should have been more clear
>     about MAC
>      >     address
>      >      >         space while checking internally. I will cross
>     check again for
>      >      >         the same and get back on this list.
>      >      >
>      >      >         Thanks
>      >      >         Santosh P K
>      >      >
>      >      >         On Wed, Jul 31, 2019 at 10:54 AM Joel M. Halpern
>      >      >         <[e-mail] <mailto:[e-mail]>
>     <mailto:[e-mail] <mailto:[e-mail]>>
>      >     <mailto:[e-mail] <mailto:[e-mail]>
>     <mailto:[e-mail] <mailto:[e-mail]>>>> wrote:
>      >      >
>      >      >             Sorry to ask a stupid question.  Whose
>     implementation?
>      >      >
>      >      >             The reason I ask is that as far as I can tell,
>     since the
>      >      >             tenant does not
>      >      >             have any control access to the VTEP, there is no
>      >     reason for
>      >      >             the VTEP to
>      >      >             have a MAC address in the tenant space.  Yes, the
>      >     device has
>      >      >             a physical
>      >      >             MAC address.  But the tenant could well be
>     using that MAC
>      >      >             address.  Yes,
>      >      >             they would be violating the Ethernet spec. 
>     But the whole
>      >      >             point of
>      >      >             segregation is not to care about such issues.
>      >      >
>      >      >             On the other hand, if you tell me that the VMWare
>      >      >             implementation has an
>      >      >             Ethernet address that is part of the tenant
>     space, well,
>      >      >             they made up
>      >      >             this particular game.
>      >      >
>      >      >             Yours,
>      >      >             Joel
>      >      >
>      >      >             On 7/31/2019 1:44 PM, Santosh P K wrote:
>      >      >              > I have checked with implementation in data
>     path.
>      >     When we
>      >      >             receive a
>      >      >              > packet with valid VNI then lookup for MAC will
>      >     happen and
>      >      >             it is VTEP own
>      >      >              > MAC then it will be trapped to control
>     plane for
>      >      >             processing. I think we
>      >      >              > can have following options
>      >      >              > 1. Optional managment VNI
>      >      >              > 2. Mandatory inner MAC set to VTEP mac
>      >      >              > 3. Inner IP TTL set to 1 to avoid
>     forwarding of packet
>      >      >             via inner IP
>      >      >              > address.
>      >      >              >
>      >      >              >
>      >      >              > Thoughts?
>      >      >              >
>      >      >              > Thansk
>      >      >              > Santosh P K
>      >      >              >
>      >      >              > On Wed, Jul 31, 2019 at 9:20 AM Greg Mirsky
>      >      >             <[e-mail]
>     <mailto:[e-mail]> <mailto:[e-mail]
>     <mailto:[e-mail]>>
>      >     <mailto:[e-mail] <mailto:[e-mail]>
>     <mailto:[e-mail] <mailto:[e-mail]>>>
>      >      >              > <mailto:[e-mail]
>     <mailto:[e-mail]>
>      >     <mailto:[e-mail] <mailto:[e-mail]>>
>      >      >             <mailto:[e-mail]
>     <mailto:[e-mail]>
>      >     <mailto:[e-mail]
>     <mailto:[e-mail]>>>>> wrote:
>      >      >              >
>      >      >              >     Hi Dinesh,
>      >      >              >     thank you for your consideration of the
>      >     proposal and
>      >      >             questions. What
>      >      >              >     would you see as the scope of testing the
>      >      >             connectivity for the
>      >      >              >     specific VNI? If it is
>     tenant-to-tenant, then
>      >     VTEPs
>      >      >             will treat these
>      >      >              >     packets as regular user frames. More
>     likely, these
>      >      >             could be Layer 2
>      >      >              >     OAM, e.g. CCM frames. The reason to use
>     127/8 for
>      >      >             IPv4, and
>      >      >              >     0:0:0:0:0:FFFF:7F00:0/104 for IPv6 is
>     to safeguard
>      >      >             from leaking
>      >      >              >     Ethernet frames with BFD Control packet
>     to a
>      >     tenant.
>      >      >              >     You've suggested using a MAC address to
>     trap the
>      >      >             control packet at
>      >      >              >     VTEP. What that address could be? We
>     had proposed
>      >      >             using the
>      >      >              >     dedicated MAC and VTEP's MAC and both
>     raised
>      >     concerns
>      >      >             among VXLAN
>      >      >              >     experts. The idea of using Management
>     VNI may
>      >     be more
>      >      >             acceptable
>      >      >              >     based on its similarity to the practice
>     of using
>      >      >             Management VLAN.
>      >      >              >
>      >      >              >     Regards,
>      >      >              >     Greg
>      >      >              >
>      >      >              >     On Wed, Jul 31, 2019 at 12:03 PM Dinesh
>     Dutt
>      >      >             <[e-mail] <mailto:[e-mail]>
>     <mailto:[e-mail] <mailto:[e-mail]>>
>      >     <mailto:[e-mail] <mailto:[e-mail]>
>     <mailto:[e-mail] <mailto:[e-mail]>>>
>      >      >              >     <mailto:[e-mail]
>     <mailto:[e-mail]>
>      >     <mailto:[e-mail] <mailto:[e-mail]>>
>     <mailto:[e-mail] <mailto:[e-mail]>
>      >     <mailto:[e-mail] <mailto:[e-mail]>>>>>
>      >      >             wrote:
>      >      >              >
>      >      >              >         Hi Greg,
>      >      >              >
>      >      >              >         As long as the inner MAC address is
>     such
>      >     that the
>      >      >             packet is
>      >      >              >         trapped to the CPU, it should be
>     fine for
>      >     use as
>      >      >             an inner MAC is
>      >      >              >         it not? Stating that is better than
>     trying to
>      >      >             force a management
>      >      >              >         VNI. What if someone wants to test
>      >     connectivity
>      >      >             on a specific
>      >      >              >         VNI? I would not pick a loopback IP
>      >     address for
>      >      >             this since that
>      >      >              >         address range is host/node local
>     only. Is
>      >     there a
>      >      >             reason you're
>      >      >              >         not using the VTEP IP as the inner IP
>      >     address ?
>      >      >              >
>      >      >              >         Dinesh
>      >      >              >
>      >      >              >         On Wed, Jul 31, 2019 at 5:48 AM
>     Greg Mirsky
>      >      >              >         <[e-mail]
>     <mailto:[e-mail]>
>      >     <mailto:[e-mail] <mailto:[e-mail]>>
>      >      >             <mailto:[e-mail]
>     <mailto:[e-mail]>
>      >     <mailto:[e-mail]
>     <mailto:[e-mail]>>> <mailto:[e-mail]
>     <mailto:[e-mail]>
>      >     <mailto:[e-mail] <mailto:[e-mail]>>
>      >      >             <mailto:[e-mail]
>     <mailto:[e-mail]>
>      >     <mailto:[e-mail]
>     <mailto:[e-mail]>>>>> wrote:
>      >      >              >
>      >      >              >             Dear All,
>      >      >              >             thank you for your comments,
>      >     suggestions on
>      >      >             this issue,
>      >      >              >             probably the most challenging
>     for this
>      >      >             specification. In the
>      >      >              >             course of our discussions,
>     we've agreed to
>      >      >             abandon the
>      >      >              >             request to allocate the
>     dedicated MAC
>      >     address
>      >      >             to be used as
>      >      >              >             the destination MAC address in
>     the inner
>      >      >             Ethernet frame.
>      >      >              >             Also, earlier using VNI 0 was
>     changed from
>      >      >             mandatory to one
>      >      >              >             of the options an
>     implementation may
>      >     offer to
>      >      >             an operator.
>      >      >              >             The most recent discussion was
>     whether
>      >     VTEP's
>      >      >             MAC address
>      >      >              >             might be used as the
>     destination MAC
>      >     address
>      >      >             in the inner
>      >      >              >             Ethernet frame. As I recall it, the
>      >     comments
>      >      >             from VXLAN
>      >      >              >             experts equally split with one
>     for it
>      >     and one
>      >      >             against. Hence
>      >      >              >             I would like to propose a new
>     text to
>      >     resolve
>      >      >             the issue. The
>      >      >              >             idea is to let an operator select
>      >     Management
>      >      >             VNI and use
>      >      >              >             that VNI in VXLAN encapsulation
>     of BFD
>      >      >             Control packets:
>      >      >              >             NEW TEXT:
>      >      >              >
>      >      >              >                 An operator MUST select a VNI
>      >     number to
>      >      >             be used as
>      >      >              >                 Management VNI. VXLAN
>     packet for
>      >      >             Management VNI MUST NOT
>      >      >              >                 be sent to a tenant. VNI
>     number 1 is
>      >      >             RECOMMENDED as the
>      >      >              >                 default for Management VNI.
>      >      >              >
>      >      >              >             With that new text, what can be the
>      >     value of
>      >      >             the destination
>      >      >              >             MAC in the inner Ethernet? I
>     tend to
>      >     believe
>      >      >             that it can be
>      >      >              >             anything and ignored by the
>     reciever VTEP.
>      >      >             Also, if the
>      >      >              >             trapping is based on VNI
>     number, the
>      >      >             destination IP address
>      >      >              >             of the inner IP packet can from
>     the range
>      >      >             127/8 for IPv4,
>      >      >              >             and for IPv6 from the range
>      >      >             0:0:0:0:0:FFFF:7F00:0/104. And
>      >      >              >             lastly, the TTL to be set to 1 (no
>      >     change here).
>      >      >              >
>      >      >              >             Much appreciate your comments,
>      >     questions, and
>      >      >             suggestions.
>      >      >              >
>      >      >              >             Best regards,
>      >      >              >             Greg
>      >      >              >
>      >      >
>      >
>