[6lo] Link Local address and 6BBR

["Pascal Thubert (pthubert)" <[e-mail]>]

Dear all :

We are ready to publish draft 10 for the backbone router and ask for WGLC, but for one pending issue that was raised during the internal review between authors on the 6BBR draft.

The initial intention was not to cover Link Local addresses at all, since the multilink subnet is made of different links and the 6BBR is a L3 box that isolates them. This is true for a routing proxy, but not exactly right for a bridging proxy, which is really a layer-3 switch with MAC layer bridging continuity on the data path.

But doing so, we bar Link Local traffic that could have happened between nodes attached to different 6BBRs, e.g., in a Wi-Fi environment where the 6BBRs can be collocated with APs and maybe operating as Bridging Proxies. The proposal on the table is thus to proxy ND for Link Local addresses in the case of a bridging proxy. The registration and proxy operation would be the same as for a Global Address, but there's at least one caveat.

Consider the flow in https://tools.ietf.org/html/draft-ietf-6lo-backbone-router-09#section-3.1. The EDAR/EDAC exchange with the 6LBR would contain the Link Local address as opposed to a Global Address, but the EDAR/ EDAC does not give an indication of a link ID. If the 6LBR is on-link then that's not an issue but if we place it farther away, we could detect collision that would be a false positive for link local addresses used in different MLSNs served by a same 6LBR.

There are a number of possibilities to solve this:

  *   A new ND option in the DAR/DAC to indicate a Link ID (complex to specify, setup and deploy)
  *   Make the scope of uniqueness for a Link Local Address the collection of links covered by a 6LBR (easy, no change in the spec)
  *   Indicate the MLSN prefix (e.g., in the source address of the EDAR) and make that MLSN the scope of uniqueness for a Link Local Address (must force the 6BBR to source the DAR with an address in the subnet)

What do people think?

Pascal