►
From YouTube: RATS Architecture Design Team, 2020-06-02
Description
RATS Architecture Design Team, 2020-06-02
A
B
B
C
First,
one
goes
back
to
that
original
discussion
how's
my
audio
working
this
morning
by
the
way
alright
goes
back
to
that
original
discussion.
We
had
at
the
hackathon
where
the
editors
met
out
in
the
hallway
or
whatever,
and
there
was
this
special
that
we
had
around
the
table
about
freshness
versus
recency,
which
I
think
was
originally
Ned's
point
during
then,
and
it
was
explained
to
me
that
he
was
using
the
Chairman
I
think
it
was
Ned.
C
But
I
may
be
wrong
that
the
difference
between
the
terms
as
one
was
referring
to
how
long
ago
the
evidence
was
constructed.
Nathan
was
referring
to
how
long
ago
they
claimed
values
that
appear
on
there
were
constructed,
and
those
are
two
different
concepts.
But
this
text
is
not
talking
about
that.
This
text
is
only
talking
about
when
they,
when
the
evidence
was
constructed
and
ignores
the
points
about
the
claims
being
constructed
and
so
to
define
recency
versus
fraction
and
so
I
didn't
know.
A
No
I
think
there
is
a
misconception
about
business
unit
and
I
were
talking
about
if
developing
these
terms,
for
quite
some
time
now,
due
to
the
confusion
of
uniqueness
and
recent,
let's
both
combine
to
the
grain
freshness
and
I.
Remember
this
discussion.
The
Nets
explained
it
to
me
at
some
points
know
there
was
that
the
father
right
it.
C
A
Yes,
because
of
a
villainess
is
ready.
So
yes,
I,
remember
that
one.
So
there
are
I,
think
the
freshness
section
section,
because
I
know
we
have
conflicting
texts,
I'm,
not
sure
how
much
we
should
talk
about
the
current
freshness
text,
because
Thomas
and
I
were
what's
missing
to
the
last
minute,
literally
and
Thomas
at
the
pen
for
the
last
fine.
Oh
you
text
for
this,
so
I'm
not
sure
so.
C
Out
of
the
four
that
I've
reviewed,
this
one
is
the
one
that
is
probably
the
easiest
to
get
consensus
on,
but
none
of
them
are
actually
easy
right.
I
think
all
of
them
are
contentious
and
I
have
for
change.
Requests
on
all
of
them.
I'd
also
be
happy
accepting
none
of
them.
Well,
I.
Don't
know
that
that's
easy
to
make
progress.
I.
B
C
D
Like
everybody
confused
about
the
scope
of
this,
this
is
9t
right,
so
90.
This
is
subsumed
by
the
by
the
latest
one
we're
working
on
with
Hank.
We
should
own
umber,
but
but
this
could
be
I,
think
close,
because
the
result
of
the
gigantic
discussion
here,
who
is
you,
know,
go
out
and
make
another
PR
for
this.
B
D
B
C
D
Is
II
is
an
attempt
I
made
to
reflow
section
10
in
a
way
that
includes
most
of
Hanks
point
about
recentness
and
and
so
on.
So
I
made
a
first
commit,
and
that
was
you
know,
at
least
from
a
different
perspective,
who
was
pretty
clear,
then
Hank
added
a
couple
of
community
there
I
have
some
confusion
coming
from
you
know
conflicting
branches
or
something
because
the
the
the
view
then
becomes
a
bit
hazy.
D
A
My
fault
I
accidentally
rebased,
a
branch
locally
I,
did
not
may
realize
that
I
created
a
PR
as
I
commit
for
this
PR
from
that,
and
it
totally
screwed
everything
sorry
for
my
French.
So
all
these
these
these
edits
here
were
not
attentional
in
a
hurry.
Let's
call
it
that's
yes
and
Thomas
yeah.
That
is
the
link
where
the
text
we
are
talking
about
for
freshness,
section,
I'm,
very
sorry
about
all
these
changes
in
this
PR.
They
are
reverting
things
back
to
old
stuff.
Okay,.
B
B
C
D
B
A
And
it
is
based
on
the
Thomas
sanity
checks.
That's
really
call
him
it,
because
he
was
very
elaborate
and
explaining
when
he
thought
semantically.
This
is
basically
the
same.
This
is
just
a
finer
granularity
or
this
is
a
combination
of
the
two
above,
so
we
are
Thomas
a
sanity
check,
most
of
it.
So
if
I
was
speaking
and
you
can
copy
the
link
from
the
check,
if
you
have
not
seen
a
seen
that
the
link
is
also
in
the
chat,
you
don't
have
to
watch
it
or
get
it
on
chef's
screen
in
the
box.
B
B
D
And
then
we
have
a
paragraph
explaining
how
the
combination
of
these
two
can.
You
know,
create
creative
situations
where,
for
example,
we
we
we
do
this
beacon,
beacon
style
at
the
station.
That
Hank
was
proposing
in
the
fourth,
which
I
think
is
a
variation
on
the
nones
module
defining
what
a
nonce
is.
D
B
B
C
D
That
there's
a
lot
of
tests
that
actually
reused.
It
is
just
you
know,
I,
shuffle,
that
around
a
bit
and
and
that's
it
and
I-
have
some
precision
here
and
there
and
I
see
the
thing
that
I
did
is
substantial.
I
think
he's
removing
the
replay
attack
protection.
We
seem
to
be
the
goal
of
the
whole
thing
with
which
is
not
in
fact
for
freshness.
The
non-space
thing
is
is
only
about
locking
timelines
from
different
person,
participants,
the.
B
C
D
E
C
B
C
Easy
enough
for
you
to
just
do
them,
so
in
line
850,
where
it
says
whether
the
included
claims,
if
you
do
a
suggested
phrase,
addition
to
see
people
think
yep
right
after
claims
and
before
can't
where
it
says
claims
can
in
between
claims
and
can
I
would
say
claims
and
their
values
lose
to
you.
The
fact
that
the
claim
and
the
value
could
have
a
different
timestamp
and
that's
my
first
one.
The.
C
F
C
A
B
B
C
C
D
C
C
I
F
I
F
D
C
The
reason
that
I,
the
way
that
I
read
the
first
part
of
the
sentence
is
even
on
a
per
claim
basis.
If
wanted
actually
is
talking
about
two
cases:
either
it's
on
a
per
claim
basis
or
it's
not
and
I
believe
that's
actually
the
same
point
right
whereas
as
well
as
sounds
like
that
you're
saying
and
there's
another
orthogonal
concept
too
and
I
don't
think
they
are
I.
Think
it's
saying
the
same
thing,
two
different
ways.
F
C
C
G
B
B
A
A
A
A
G
A
C
D
G
G
Maybe
I
can
help
out
just
speaking
a
little
except
except
erroneously.
On
this
point
right
and
and
forgive
me
if
I
sound,
a
little
ignorant
I
don't
follow
every
last
change
that
goes
on
in
this
document,
but
I
think
that
so
long
as
I
think
time
is
something
that
just
has
to
be
asserted.
Like
anything
else,
and
you
know
you
you,
the
the
tester
can
assert
it
and
you
can
trust
that
value
or
not.
C
G
A
C
A
Make
sure
that
it
fits
everywhere
and
using
the
right
word,
that's
fine!
If
it's
confusing
and
it
is
literally
defined
as
the
I'm
pretty
sure
about
that.
But
if
it's
too
complicated
and
too
technobabble,
as
they're
said,
we
have
to
make
sure
that
we
mean
that
we
start
to
count
time
from
zero
here.
That's
it
I
think.
G
C
The
concept
of
having
a
zero
is
not
interesting,
only
deltas
are
interesting
and
actual
timestamps
are
interesting.
So,
for
example,
if
you
look
at
the
the
SGX
specification
as
an
example-
because
that's
not
here
right,
yes,
jack
specification-
says
that
you
can
get
relative
time,
but
the
zero
value
is
meaningless.
It's
the
same
thing
as
in
so
elegant.
You
know
like
in
counters
in
mids
right
because
they
roll
over
the
value.
Zero
is
not
meaningful.
All
that's
meaningful
is
the
Delta.
Well.
B
B
C
G
C
A
B
C
A
C
Anything
unless
you
put
the
so
like
in
the
evidence
right,
then
there
may
be
no
concept
of
the
validity
span,
as
you
put
it
in
the
evidence
right
if
the
expiration
is
pretty
on
the
receiver
saw
there's
nothing
inside
the
claims.
It
would
designate
that
it's
firmwares
like
the
station
is
open,
it's
possible
in
the
evidence,
but
it's
not
necessary
yeah.
A
I
A
For
example,
if
you,
if
you
have
a
specific
kind
of
non
system
and
the
dances
the
validity
spend,
the
nonce
is
only
viable
for
a
minutes,
and
then
you
generate
the
unique
value
here.
You
are
initiating
this
time
out
for
validity.
Everything
is
created
for
evidence
with.
This
is
considered
fresh.
Until
this,
this
unique
value
is,
is
one
minute
old.
Then
you
have
to
start
and
create
a
new
one.
So
this
new
one
creation
and
this
duration
of
a
minute
always
starts
with
an
epoch
of
zero
and
with
the
epoch
of
the
time
since
absolute.
A
A
I
Don't
know
I
forget
it,
I
mean
the
verify
side
could
could
we
could
all
hang
on
the
time
the
clock
that
the
verifier
uses,
starting
from
the
transmittal
of
the
of
the
challenge
knots,
but
there's
this
other
question
that
component
claims,
but
particularly
in
you
know,
layered
attestation
of
compositional
applications.
Those
claims
could
have
been
sitting
there
for
a
day
right
that
that's
going
to
be
bundled
up
into
ten.
You
know
claims
and
sent.
So
not
not
all
ten
will
be.
You
know,
in
the
same
time,
you
know
sphere
as
the
as
the
nonce
right.
I
A
D
A
C
A
C
C
A
C
Says
it
yes,.
A
A
B
B
B
A
D
Introduction,
we
have
first
first
paragraph
say,
says
remote
attestation
procedures,
one
peer,
your
tester
produces
believable
information
about
itself,
enable
a
remote
peer,
the
relying
party
to
decide
so
I
think
the
rest
of
the
document
these
remote.
This
is
from
the
point
of
view
of
so
what
is
remote
is
exactly
is
blank
party
another.
That's.
B
Like
just
saying
local
and
remote,
and-
and
we
also
have
the
connection
between
the
verifier
and
the
relying
party
as
being
a
local
remote
entity
anyway,
so
that's
why
I
would
rather
just
remove
the
word
remote
or
or
indicate
its
its
its
role,
which
is,
that
is
a
verifies.
The
validating
or
appraising
entity.
B
A
C
B
A
A
D
A
D
B
C
A
C
Any
rougher
than
the
precision
of
the
time
synchronization
now
time
synchronization
may
be
more
precise,
but
the
point
is:
you've
got
a
little
bit
of
wiggle
room
and
the
time
synchronization,
which
is
you
know
your
drift
and
so
on,
accounted
for
and
the
nonce
is
just
potentially
a
little
bit
longer.
That
has
to
do
with.
You
know
prop
delay
and
back,
but
it's
the
same
concept.
There's
nothing!
That's
any
more
rough
here
than
before,
rather
than
maybe
the
matter
of
degree.
B
B
B
D
C
D
C
Point
is
the
same
as
in
what
was
at
860,
whatever
it
was,
which
is
that
the
receiver
uses
the
claims
and
makes
its
own
determination
of
how
much
it
trusts
the
time
synchronization
mechanism
all
right
same
thing
is
true:
your
receiver
takes
whatever
claims
it
is
and
decides
how
much
that
trusts
your
relative
clock
after
using
relative
time
stamps,
meaning
here's
the
deltas
from
this
claim,
was
generated.
The
claim
value
was
generated.
D
C
C
C
C
A
F
A
D
Idea
was,
as
I
said,
to
have
a
very
clean
situation
where
there's
no
clock
on
the
other
side
right
and
then
and
then
whatever,
and
then
we
can
compose
norms
and
timestamp
a
separate
thing.
Okay,
but
here
we
a
light
exactly
what
are
the
base
requirements
and
what
we
can
expect
from
a
situation.
That
is,
you
know,
purple.