Re: [6tisch] MSF adapts to traffic only for secured packets

[Mališa Vučinić <[e-mail]>]

Yatch,

This is simply a matter of not allocating resources, i.e. cells, in the network in response to the unauthenticated traffic. Even if the intermediate nodes have rate limiting on AF43, how could we ensure that the “acceptable” amount of unauthenticated traffic that actually gets forwarded does not trigger cell allocation? The cells used to transport the join traffic will be marked as used by MSF, and depending on the amount of regular traffic as well as the join rate limit, this may cause the cell allocation threshold to be surpassed. Right?

Mališa

> On 5 Dec 2019, at 18:25, Yasuyuki Tanaka <[e-mail]> wrote:
> 
> Thanks, Malisa,
> 
> Then, why cannot the IPv6 layer on an intermediate have rate limiting in order not to forward too much packets having AF43...? Forwarding decisions are made at the IPv6 layer.
> 
> Even if the intermediate node drops excess amount of forwarded join requests, the scheduling function in use still needs to do something? I may be confused...
> 
> Yatch
> 
> On 12/5/2019 6:07 PM, Mališa Vučinić wrote:
>> The “join rate” parameter takes care that any single JP at the edge of the network does not inject too much traffic. But this traffic is forwarded along multiple hops towards the root, and therefore gets aggregated with (join) traffic from other JPs in the network. The purpose of the traffic tagging mechanism in minimal-security is for such nodes, closer to the DAG root, to avoid allocating cells in response to the join traffic.
>> Mališa
>>> On 5 Dec 2019, at 17:48, Yasuyuki Tanaka <[e-mail] <mailto:[e-mail]>> wrote:
>>> 
>>> Why can't the "join rate" avoid such undesired cell allocations? If the join rate is properly configured, incoming join requests don't cause such allocations, do they?