Re: [v6ops] draft-taylor-v6ops-fragdrop WGLC

[George Michaelson <[e-mail]>]

On Tue, Aug 20, 2013 at 12:40 AM, joel jaeggli <[e-mail]> wrote:

> On 8/18/13 9:27 PM, George Michaelson wrote:
>
> >
> > I am a little worried the document makes unattributed statements about
> > ASIC and FPGA costs. I think they need to be fleshed out, because this
> > kind of unsubstantiated and unattributed comment can lead to people
> > continuing to assume things which applied once, to some vendors, but
> > doesn't apply now (consider the whole thing about secure disk wiping
> > which was predicated on a very specific generation of HDD hardware.
> > Or, the "BGP is dying we need locator+id" debate, which arguably stems
> > from a false premise)
> >
> Actually I think we pretty carefully skirted that in fragdrop, it's
>
> http://tools.ietf.org/html/draft-wkumari-long-headers-01
>
> where there hardware is front and center.
>
> here it's fast vs slow (potentially no path) / no feasible way to do
> reassembly.
>
>
I think a citation at this point in the document would help. Really thats
all my point is: If you repeat text inline which says "...is a problem..."
then without a backchain to where you explore it fully, it becomes
unsubstantiated, and untestable against changes in the h/w horizon.

Why skirt it?

Not a big deal anyway. I like the document and think its worth publishing.

-G