Re: [Lsvr] Adoption of draft-ymbk-lsvr-l3dl-(signing|ulpc)

[Randy Bush <[e-mail]>]

>>> C: It feels like more text is needed around "If a peering address has
>>> been announced as a loopback, a two or three (one or both ends could
>>> be loopbacks) hop
>>>    BGP session will be established. " -- how do I know it is announced
>>> as a loopback? How do I know what route to install to reach the
>>> loopback? etc.
>>
>> l3dl 13.2 tells you it is a loopback
> 
> Doh, I missed that...
> 
>>
>> 13.2.  Encapsulaion Flags
>>
>>    The Encapsulation Flags are a sequence of bit fields as follows:
>>
>>     0           1            2            3            4  ...       7
>>    +------------+------------+------------+------------+------------+
>>    |  Ann/With  |   Primary  | Under/Over |  Loopback  | Reserved ..|
>>    +------------+------------+------------+------------+------------+
>>
>> so how about
>>
>>    For each BGP peering on a link here MUST be one agreed encapsulation,
>>    and the addresses used MUST be in the corresponding L3DP IPv4/IPv6
>>    Announcement PDUs.  If a peering address has been announced as a
>>    loopback, i.e. MUST BE flagged as such in the L3DL Encapsulation PDU,
>>    a two or three hop BGP session will be established.  Otherwise a
>>    direct one hop session is used.
> 
> Close but no cigar -- I suspect I didn't fully explain my (second) question
> If router A says that I should peer with the loopback address
> 192.0.2.1, I need to know how to route *to* that address

you will likely want to forward, <pedantry> not route </pedantry>, to
the peer's address which was marked as Primary, iff it is in a subnet
which you also share.  if not, pick a forwarding next hop that is in a
subnet you share.  if there are multiple choices, you could have
signaled which subnet in an Upper Layer Protocol Configuration PDU
Attribute.

if you share no subnets, then you do not have a valid underlying l3dl
session see l3dl

   13.  The Encapsulations
   ....
   Further, to consider a logical link of a type to formally be
   established so that it may be pushed up to upper layer protocols, the
   addressing for the type must be compatible, e.g. on the same IP
   subnet.

yes, you have to hold a lot of gorp in your head.  sorry.

but some clarifying text might keep others from falling down this rabbit
hole.  thanks.

randy