Re: [TLS] Deployment ... Re: This working group has failed

[Watson Ladd <[e-mail]>]

On Mon, Nov 18, 2013 at 7:02 AM, Salz, Rich <[e-mail]> wrote:
>> TLS 1.2 solves the same problem as TLS 1.0. It should therefore have the same API.
>
> Do you really believe this or are you trying to just be provocative?
Do you really believe that backwards compatibility at the source level
had to be sacrificed?
Stable interfaces are the norm, unstable ones are remarkable. How many
prongs does a ground-fault detecting electricity socket have?
BLAS has multiple implementations, all with the same API. MPI looks
the same on a bunch of Xboxes wired together with Ethernet as it does
on a several million dollar supercomputer. GMP doesn't gratuitously
change its interface with every performance enhancement, and to this
day "Hello World" works unchanged on my machine, some 40 years after
it first ran on a PDP-11. Everything about the environment it runs on
is different, from the word size, to the endianness, to the output
mechanism, to the interface I am using. And yet it works with only a
recompile.

TCP has undergone many changes to the implementation, yet the BSD
sockets API still works. Why exactly does an application need to care
about which ciphersuite is used? Why does it need to do more than hand
over some trusted PKI roots, and a server certificate?

The current APIs have caused lots of security bugs as people don't use
them correctly. The solution: high level APIs that won't change when
the implementation is upgraded. Is this really too lofty a goal? The
costs of the current approach are obvious: what are the costs of
making better APIs?
Sincerely,
Watson Ladd

>
>
> --
> Principal Security Engineer
> Akamai Technology
> Cambridge, MA



-- 
"Those who would give up Essential Liberty to purchase a little
Temporary Safety deserve neither  Liberty nor Safety."
-- Benjamin Franklin