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

[Lorenzo Colitti <[e-mail]>]

On Thu, Nov 1, 2012 at 9:51 AM, Joe Touch <[e-mail]> wrote:

> However, you've answered the key issue - this is not about fragments, but
> rather about two things:
>
> a) DPI
>         DPI cannot be done on packets with any options
>         so this either means operators need to ignore DPI
>         on such packets or drop them
>

I don't think we should use the term DPI, because we don't really have
consensus on what it means. I suspect some would assert that just looking
at layer 4 headers is not DPI and that it's only DPI if you look into layer
4 payloads as well.


> There are certainly places where DPI is expected (e.g., by a government),
> and so only DPI-capable packets are permitted, but this doesn't seem
> tenable anywhere except at the boundaries of customer networks (otherwise
> all encrypted traffic would be dropped all over the place, and that doesn't
> appear to be happening)
>

... and saying that looking at layer 4 headers is "DPI" will also lead to
this kind of confusion.