Re: [tcpm] SYN/ACK Payloads, draft 01

Joe Touch wrote:
> 
> >
> > It's certainly a consideration, which is why the draft recommended
> > that 64 bytes be the maximum payload size. At this size there are
> more
> > juicy targets for amplification, like DNS servers.
> 
> FWIW, since this is payload data, the size of the data is an
> implementation detail - unless, as I've noted, you're changing the
> semantics of TCP from a byte stream to delineated messages.
> 
> (yes, you're already changing the semantics from one where connection
> info is available only after TWHS to one that allows info to be used
by
> the app before the TWHS completes, which is a different issue)
> 

If this were to move forward, the Security Considerations would need
to document the amplification attack, and probably have SHOULD language
about a default maximum packet size.

But this is probably not something that the TCP stack should enforce.
There are many situations where the system administrator would have
valid reasons for knowing that a DoS was already blocked by other means.

_______________________________________________
tcpm mailing list
tcpm@ietf.org
https://www.ietf.org/mailman/listinfo/tcpm