It is possible to define formal semantics for data integrity in the
OPES model and implement them in a verifiable manner, overlay
network or not.
Hilarie
"Ian Cooper" <icooper(_at_)equinix(_dot_)com> 08/10/01 01:59AM >>>
At 15:42 8/9/2001 -0600, Hilarie Orman wrote:
OPES will increase the integrity of content delivery over what is
available today.
I agree to some extent, but as I've mentioned before (and Ted reiterated in
the meeting on Tuesday) there are some issues with such an "overlay
network" that might reduce the integrity.
It is of only academic interest to contrast
a pure end-to-end world against a world with OPES.
I'm not so sure. As I've stated before, I have some serious concerns in
failover and roaming-user conditions. (These are not limited to OPES,
admittedly.)
As another
purely academic exercise, contrast a pure voice switching
model against the Internet. The Internet offers many opportunities
for traffic monitoring, substitution, and malicious intervention that
the phone system did and does not. Whether or not the benefits
outweighed the risks was not evident a priori; in the case of OPES,
you cannot make such an evaluation simply by pointing out risks
that are already there without OPES and claiming that OPES somehow
makes them worse.
True, though the evaluation also acknowledges that the OPES work is likely
to be developed outside IETF (if a WG isn't formed here) anyway. I think
the point is whether IETF is seen to "sanction" such work. I don't have
any particularly strong feelings either way on that point.