On Mon, Dec 8, 2014 at 7:06 PM, Joe Touch <touch@xxxxxxx> wrote:
> That is because it is a workshop on how the Internet can evolve to
> realize the architecture of the stack, not a workshop on how the
> architecture of the stack can evolve to address the way people use it.
I had thought that part of the meeting would be to address the tension
between these two issues, but I have also since learned it has become
"how do we evolve the Internet to accept whatever middleboxes want to
do" - i.e., precisely the latter of your examples.
> The only end points that can be fixed end to end on a network are
> cryptographic keys and data bound to cryptographic keys. Ports and IP
> addresses are ephemera.
Not according to the current Internet architecture, but that's back to
my point above.
An architecture is defined as much by what it is not as what it is. If
everything is on the table as changeable, then there is NO architecture
anymore.
If you believe in 'permissionless innovation' then everything is always on the table. The Internet architecture to date has been what survived a Darwinian process.
I don't believe in in ancestor worship. Grace Hopper was a visionary pioneer but I am really glad we stopped using COBOL and FORTRAN even if many of those ideas made it into later languages.
The statement I made was not about Internet architecture. It is a basic fact about a network. The only thing that you can rely on being constant end-to-end is what you protect with cryptography.