Re: per neighbor packet buffering "issue"
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: per neighbor packet buffering "issue"
At Tue, 21 Oct 2008 16:08:09 +0200, Julien Abeillé wrote:
>
> uIPv6 holds on 11K of flash, and a minimum of 1.8K of RAM (3K with IP
> fragmentation support)
>
> - The most costly feature is "per neighbor packet buffering", mandated
> by RFC4861. On constrained platforms, there is no dynamic memory
> allocation, hence if the neighbor cache can hold four entries (a
> minimum in a sensor network), we must define four 1280 bytes buffer,
> which is huge in comparison to the initial 1.8K.
>
> My question is: do you think it makes sense to change the MUST into a
> SHOULD?
I would support SHOULD. Packet buffers are a precious resource when
you get far enough into the embedded world. Protocol specifications
have to allow for the possiblity that both packet buffers might
already be in use. :)
--------------------------------------------------------------------
IETF IPv6 working group mailing list
ipv6 at ietf.org
Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6
--------------------------------------------------------------------
Note: Messages sent to this list are the opinions of the senders and do not imply endorsement by the IETF.