On 1.7.2015 15.19, Juliusz Chroboczek wrote:
HNCP maximum packet size is 4000 bytes (Section 3 of -06). Assuming
a router wishes to publish 400 bytes of per-router information, and for
each interface two neighbours, two assigned prefix TLVs (one /64 and one
IPv4 /24) and two node addresses (one IPv6 and one IPv4), the total size
of the state being flooded is
400 + interfaces * (32 + 16 + 24 + 24 + 24) = 400 + interfaces * 120
This implies that without fragmentation we can handle up to 30 interfaces
per router. If Markus changes the protocol to not flood NODE-ADDRESS
globally, or if implementors choose to use SLAAC/DHCPv4 for assigning
router addresses or use unnumbered interfaces, this becomes:
400 + interfaces * (32 + 16 + 24) = 400 + interfaces * 72
which allows up to 50 interfaces.
Only one node per common link should have AP for each prefix. So your
model of a network is actually bit better in the wild. However, you do
not include e.g. SD stuff so maybe it's a wash.
That leaves the interesting question about whether or not advertising
_all_ router addresses is really useful or not; SLAAC-ish allocation
(for example) could be used here instead, and just particular addresses
used for (e.g.) SD be advertised in the service-specific TLVs.
On a related note, hncp-06 only says the following about the packet size
(Section 3):
Each node MUST be able to receive (and potentially reassemble) UDP
datagrams with a payload of at least 4000 bytes.
It should also say that a node MUST NOT send a datagram with a larger
payload, and that it MUST silently drop any larger datagrams (optionally
log to system management, etc.). If this is not done, persistent state
desynchronisation may occur.
I am not sure I want to cripple use in non-crippled networks, just
provide guaranteed base value which works everywhere and *RED ALERT*
light should light up on crippledbox if it encounters this.
(At a guess, I have 0 devices that have issues with quite a bit bigger
packets than 4k :-p)
Cheers,
-Markus
_______________________________________________
homenet mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/homenet
