[6lo] PASA smart home use case

[Esko Dijk <[e-mail]>]

Hi PASA authors,

As requested in the 6lo WG meeting I’m relaying my comments on the smart home use case (section 3.2 of draft-li-6lo-path-aware-semantic-addressing-00).

In Fig 2, it could be clarified better where the PASA Root sits (i.e. the root and 6lo Border Router as depicted in the architecture Fig 5). A first assumption I had was that each PLC GW contains a PASA root, because the PLC network segment beneath it is the constrained 6lowpan network technology benefitting from PASA addressing.  Then each PLC segment has to deal with constrained nodes and constraints on the network throughput (i.e. low-speed PLC). In this case, the link connecting all the PLC gateways and the Home Gateway would be another e.g. high-speed (Ethernet, Wi-Fi) technology. This link uses in typical homes its own existing addressing which may be based on SLAAC or DHCPv6 delegated prefix via CPE / Home Gateway, or ULAs, or something else. The HomeNet WG defined solutions for such addressing using a new protocol I think and SNAC WG is now aiming to provide solutions for that as well that re-use already-present protocols & products in the home. That would not use PASA addresses and there seems to be no need to do that.

But, another interpretation of this figure is that Home Gateway provides the PASA Root. In this case all underlying devices including PLC Gateways themselves will get a PASA assigned address. This then would suggest that the link connecting the PLC GWs and Home Gateway is another kind of wired link, not PLC. What would this link technology be like -- is Ethernet assumed?  In this case where would the other home devices like laptops, smartphones, wireless smart home devices etc., be connected? It could be to a separate network that is not shown and that doesn’t use PASA addressing. Or maybe these devices are not shown but are on the same Ethernet link, not using PASA addresses but another type (the DHCPv6 / SLAAC / ULA / SNAC solutions mentioned above).

In summary, it’s hard for me to understand the current use case because not fully clear which interpretation of the 2 above I should assume.

In case of interpretation 1, the diagram does not show any tree topology per PLC network so I don’t understand yet why using PASA addresses here would be useful? SLAAC would work equally well. The diagram suggests that all PLC devices in a segment can talk to all other PLC devices in the segment in a single hop – no tree. That suggests all PLC devices are leaf devices. Is that intended? I recall there are also PLC technologies that build a tree when some device can’t reach the GW in one hop.

Regards
Esko Dijk


IoTconsultancy.nl  |  Email/Teams: [e-mail]