►
From YouTube: RATS Architecture Design Team, 2021-02-05
Description
RATS Architecture Design Team, 2021-02-05
B
Well,
it's
been
a
while
I
was
at
intel
and
so
I
had
joined
while
we
were,
but
I've
been
kind
of
on
a
hiatus,
as
I
jumped
from
intel
to
amazon.
A
B
A
A
B
A
Okay,
so
I
think
we
have
our
critical
mass
here.
I'm
sorry!
Just
writing.
I
keep
a
attendance
record
because
sometimes
that's
been
an
issue
in
past
design.
Teams.
A
All
right
anyway,
so
I
went
through
the
last
two
days
and
oh,
we
have
new
new
issues.
Okay
and
I've
been
through,
and
I've
put
up
this
tag
on
all
the
tickets,
issues
that
have
a
pull
request,
and
so
we
have
15
of
them
and
I'm
gonna
go
through
right
after
this
meeting
and
commit
the
ones
that
seem
to
have
three
or
more
approvals
without
comments.
A
A
Okay,
all
right
I
see
so
this
is
the
second
part
of
the
of
the
effort.
Do
you
want
to
do
that
first,
or
do
you
want
to
do
the
the
previous
part
relating
to
the
248?
It
was,
I
guess.
H
E
Consideration
one,
the
other
one
is
finishing
off
the
example
which
I
sort
of
handed
over
to
hank,
hank
and
eric,
because
my
my
my
task
is
finished
there
and
the
third
one
is
an
extension
to
the
freshness
section.
The
subsection
that
deals
with
handles
about
the
discussion
around
the
window
of
ebooks.
That
one
needs
to
keep
well.
I
C
248,
yes,
yeah,
I
threw
a
little
text
into
the
chat
window.
I
know
I
could
have
committed,
but
I
was
seeing
if
anybody
wanted
to
go
that
way.
I
had
a
little
tech
through
the
button
for
some
things
that
weren't
completely
handled
there.
So
if
you
just
want
to
look
at
the
text
and
people
like
it,
then
we
can
always
do
a
commit
on
it.
C
D
A
C
A
A
D
Think
it
starts
off.
Is
there
anything
above
this
in
the
green?
Is
my
question?
Oh
it's
a
blank
line.
D
D
So
is
the
illustrated
scenario
supposed
to
be
a
failure
case
or
a
success
case.
C
A
A
A
Thank
you,
okay.
So
now
I
think
I'm
going
to
just
resolve
this
conversation.
A
J
The
handle
to
evidence
mistake
resolved,
I
think,
that's.
D
G
D
A
G
A
Do
you
want
to
stop
just
go
from
the
top
to
the
bottom?
Now.
D
G
A
Yeah
I've
I've
committed
it.
I
didn't
need.
D
D
Okay,
now,
if
you
committed
it,
then
refresh
might
change
the
line.
Number.
G
D
D
Okay,
so
here
is
pointing
to
freshness,
but
the
after
we
moved
up
the
text
that
used
to
be
the
beginning
of
this
section.
There's
another
subsection
underneath
freshness,
that's
very
specific
to
handles,
and
so
first
I'd
recommend
referencing
that,
instead
of
the
less
specific
freshness
section,
so
the
first
of
two
points.
D
The
second
of
two
points
is
that
the
text
up
there
talks
about
three
ways
to
deal
with
freshness
with
handles,
or
at
least
three
example
ways.
You
can
see.
I've
quoted
that
paragraph,
so
you
have
to
scroll
up
and
it's
examples
because
it
uses
such
as,
and
so
it
would
help
if
this
one
was
gonna.
If
the
text
here
would
say
which
mechanism
it
was
going
to
use,
whether
it's
one
of
those
three
or
something
else
and
so
it'll,
be
more
understandable
in
context.
D
A
Okay,
so
the
question
is:
can
we
reference
a
subsection
of
freshness
which
is
the
implicit
time
keeping
using
epoch
handles
which
doesn't
have
a
handle?
It
doesn't
have
a
anchor
on
it.
D
And
there's
multiple
correct
pronunciations
of
that
epoch
epoch,
epic
of
all
those
different
people
say:
okay,
I
tend
to
have
the
same
one
same
pronunciation
as
epic,
but
yeah
I'll
see
what
I
can
do
to
enunciate.
A
The
suggestion,
the
suggestion
is
that,
instead
of
freshness
we're
going
to
reference,
epic
freshness,
which
is
a
handle
we'll
have
to
insert-
which
I
have
in
the
other
in
the
other
window,
to
insert
that's
what
proposal
is
dave.
D
Those
who
wrote
this
section
and
generated
the
the
updates
to
the
diagram,
which
I
understood
the
whole
diagram,
because
I
think
I
created
the
old
diagram.
If
you
want
changes,
the
diagram
which
of
those
three
mechanisms
are
we
using
here,
is
this
example
intended
to
use,
or
is
it
a
fourth
one?
D
Is
it
using
a
counter?
Is
it
including
two
handles
and
messages,
or
is
it
requiring
retries
in
the
case
of
matching
handles
or
is
it
required?
I
could
say
it's
four
or
not
three,
or
is
it
by
buffering
incoming
messages
for
some
time
or
something
else,
which
is
the
intent
in
the
example
walkthrough
we're
going
to
do
here
because
we
have.
This
is
just
an
example
right,
so
any
one
of
those
is
fine
or
even
a
fifth
one.
A
Do
any
of
those
right
now
it's
busted,
so
I'm
just
trying!
I
can
get
the
whole
the
the
diagram
on
the
screen
here.
D
Right
now
it
doesn't
do
any
of
those
right
and,
and
so
it
doesn't
actually
have
any
way
to
to
deal
with
time.
Synchronization
the
old,
the
red
text
does.
If
you
look
at
the
red
text
right
so
well,
actually
I
don't
remember.
I
can't
say
that
the
red
text
does
without
looking,
so
maybe
it
doesn't.
I
thought
it
did.
A
So
I'm
trying
to
put
the
dot.
I
think
that
dave,
we
need
to
see
the
diagram,
so
these
guys
can
tell
us
which
method
they're
using
but.
A
D
Yeah,
so
in
the
green
text
right
you
see
where
the
attestation
result
co
and
yeah.
No,
this
one
doesn't
drain.
You
can
see
in
the
green
text
here
right.
The
attestation
result
uses
h,
prime,
in
the
bottom
line,
to
the
right
and
in
the
line
right
above
that
all
of
the
entities
got
h.
Prime,
that's
not
the
question
here.
The
question
is:
what
happens
if
h,
prime?
The
relying
party
did
not
get
h,
prime,
before
that
line,
what
happens.
D
A
D
D
C
H
C
D
A
So
dave.
D
D
A
But
but
I
think
you
only,
I
think
you
only
you
only
buffer,
once
okay.
A
D
C
A
So
so
the
reason,
the
reason
why
we
need
to
have
we
need
to
know
whether
you
are
buffering
zero
results,
in
other
words
drop.
It
one
result
where
you
can
somehow
know
it's
always
the
latest
one
or
end
results
where
you
just
have
to
keep
them
forever.
Right
and
and
dave's
point
is
that
the
keep
them
forever.
There's
an
attack
on
that,
so
we
wouldn't.
We
would
prefer
not
to
specify
that,
but.
D
To
be
clear,
I'm
fine
if
we
use
that
in
this
example
right,
because
it's
just
an
example,
if
that's
the
simplest
way
to
describe
stuff,
it's
fine,
it's
fine
with
me
if
you
say,
and
we're
going
to
use
the
buffering
solution
here
in
this
example
is
in
this
example.
I
don't
have
any
objection
to
that.
So,
if
that's
the
simplest
way
to
do
the
text,
that's
fine,
I'm
just
saying
if
I
was
going
to
implement
it,
I
wouldn't
choose
that.
But
as
far
as
you
know,
this
isn't
how
you
implement
it.
K
D
D
Okay,
I
would
prefer
it
be
written
in
the
place
where
I
put.
The
comment
was
that
this
one,
I
can't
remember,
I
think.
A
D
It's
because
the
definition
provided
in
epic
freshness,
there's
four
definitions
there
in
that
paragraph,
and
so
the
definition
would
be
replaced
by
you
know,
is
fresh
as
discussed
in
freshness
where,
in
this
example
the,
if
there's
a
mish-match,
it
will
require
a
re-transmission.
You
know
maybe
there's
a
shorter
way
to
say
it,
but
something
like
that,
because
the
identification
is
meaningless
here,
we
need
to
say
which
definition.
D
And
you
see
that
paragraph
there
provides
four
of
them,
so
the
definition
is
four
of
them,
so
we
just
need
to
replace
the
definition
with
you
know
something
that
means
the
case
you
can
see
or
by
requiring
retries
in
the
case
of
matching
handles.
That's
the
case
where
the
you
have
the
drop,
and
so
that's
the
one
we're
going
to
pick
here.
D
Maybe
the
definition
provided
in
epic
freshness
where
by
there's
no
comment,
retries,
are
required
in
case
of
mismatching
handles.
If
you
want
to
use
the
same
text.
D
H
D
A
A
A
Yeah,
so
so
yeah,
so
I
have
it
in
my
buffer
and
I'm
going
to
just
add
it
to
master.
I
think
it's
the
simplest
okay
thing
added.
D
Okay,
that's
the
question
I
was
just
asking
when
we
were
looking
at
the
diagram
is
what
happens
if
that
line
didn't
make
it?
And
I'm
not
saying
we
got
to
change
the
diagram,
it
might
just
be
useful
to
add
a
parenthetical
or
something
that
says
you
know
if
the
relying
party
had
not
gotten
it
yet,
then
it
would
be
dropped.
Something
like
that
would
be
a
fine
clarifying
here
which
does
not
change
the
diagram
or
the
flow.
D
D
A
D
D
D
A
Okay,
all
right.
The
next
comment
is
here.
D
Okay,
this
is
about
diagram,
design
and
correctness
either.
This
is
true
for
every
diagram
and
the
you
know
all
the
diagrams
in
the
whole
appendix
here
either
the
only
time
that
it's
okay
to
include
a
time
in
the
actual
message,
meaning
not
a
delta,
but
the
actual
time
is
where
the
center
of
the
receiver
has
synchronized
clocks.
This
example
does
not
have
that,
so
we
should
never
ever
see
a
time
by
itself
as
opposed
to
a
delta.
D
D
Text
is
right.
Well,
old
style
is
right.
You
can
see
it's
a
different
message
being
referred
to
here.
That's
the
attestation,
which
is
the
evidence,
so
it's
actually
higher
up,
but
only
a
delta
is
interesting,
so
it
needs
to
go
back
to
I
think
whatever
was
in
the
well.
That
was
broken
there
too.
I
think
it's
supposed
to
be.
Oh
yeah.
I
think
that's
right.
What
you're
highlighting
now
is
what
it
should
say.
D
D
D
D
D
A
D
Tedious,
I
have
to
do
another
read
through
to
make
sure
it's
completely
fixed,
but
I
noticed
that
there's
no
discussion
of
the
use
of
those
times
in
the
text
under
the
table,
but
if
it's
the
same
as
the
previous
section,
it's
okay
and
I
just
go
back
and
check
now
that
I
know
what
I
don't
know,
the
drop
one
is
intended,
which
is
the
simplest
to
explain
it
might
be.
Okay,
I'd
have
to
check.
G
D
D
Up
filing
another
pull
request,
or
something
if
I
see
that
it
needs
to
be
fixed.
J
Okay,
but
they've
been,
we
have
no
state
that
capture
this.
You,
you
are
the
state.
Are
you
fine
with
that.
D
D
G
A
G
D
Sure
why
don't
you
do
that?
Why
don't
you
merge
this
leave
the
issue
open,
assign
the
issue
to
me
and
that
will
be
the
state
and
then
I
will
close
it
if
it
looks
good.
A
A
D
A
A
15
participation
night,
you
know
june
june
22
I
go
into
the
teacher,
went
like
you
know:
wtf,
like
why'd,
I
get
a
d,
I'm
she's
like
oh
you're,
a
great
student,
I'm
like
yeah,
well
where's
my
mark
and
she
said
well
richard.
I
I
couldn't
find
her
richard.
She
couldn't
find
her
richard
in
the
class,
so
you
know
anyway.
She
finally
finished
like
richard
michaelson
or
something
I
mean
that's,
who
she
thought.
I
was
the
whole
year.
Okay,
there
were.
A
Anyway,
okay,
so,
is
this
the
next
the
next
good
pr
to
deal
with
in
this
grouping.
D
Why
not
sure?
Okay,
because
I'm
thinking
we
don't
need
to
spend
time
if
things
already
have
approvals
and
no
comments,
you
can
just
deal
with
those
offloaded
merchants,
so
yeah.
This
is
fine.
So.
A
Okay,
this
looks
like
just
missing
words
to
me.
J
D
D
Use
this
example
in
text,
but
here's
what
I'm
thinking
right.
Let's
say
your
candle
distributor
is
on
earth
and
you're
a
tester
and
relying
party
and
verifier
are
all
sitting
on
jupiter
and
you're
using
you
know
a
transmission
across
space.
It's
not
lossy.
E
D
D
Is
which
means
you
don't
have
to
cache
the
subject
and
verb
of
the
goes
to
the
object?
Okay,.
D
Wait,
let
me
just
read
it
one
more
time
now
to
make
sure
the
grammar
works.
The
appraising
entity
channel
compare
the
oh
sorry
wrong
sentence.
I
gotta
read
the
first
half
of
1163,
which
is
off
the
screen.
I
think
it
works
great
now.
A
Is
indirectly.
E
D
E
A
D
So
I'm
not
sure
if
it's,
if
the
network
round
trip
time
is
negligible,
because
maybe
it's
if
you're,
okay
with
dropping
stuff
or
if
you
have
other
solutions
right,
because
if
you
have
something
in
flight,
even
if
it's
you
know
one
millisecond
or
less,
there
could
be
something
inflated
exactly
the
same
time.
D
I
yeah
I
I
like
that
idea.
Let's
just
drop
the
sentence.
H
A
D
G
D
Oh
look
he's
already
added
a
a
label
on
that,
one
that
you
had
an
accident
to
do
and
it's
different.
D
Cage
right
now
to
epic.
D
E
D
It's
only
a
denial
of
service
if
you
can
block
it
to
a
single
entity
right
like
the
receiver
right,
if
you
block
it
to
the
receiver,
then
then
then
there's
a
dial
of
service
attack.
If
you
do
it
to
just
the
sender,
it's
an
allen
service
attack.
If
you
do
it
to
both
of
them,
it's
not
a
denial
of
service
attack.
It's
just
a
stillness
attack.
D
Those
two
specific
entities-
yeah
yeah,
okay,
selective
dropping
of
handles
well
selector
to
me-
says
you
drop
one
but
not
the
other
handle
to
the
same
entity.
Okay,
that's
not
what
we're
talking
about
here,
you're
talking
about
dropping
it
to
some
entities,
not
others.
So.
A
E
I
D
B
D
That
should
say
drunk
drop
or
block
yeah
either.
One.
E
B
E
Is
true
I
mean
well,
I
I
thought.
E
D
You
could
remove
the
sentence
or
we
could
figure
out
how
to
fix
this.
E
Well,
I
I
wanted
to
have
something
there
right,
so
I
thought
this
one,
because
if
I
felt
bad
because
that
attack
is
otherwise
you
know
unaddressed
and
that's
not
good
yeah.
E
A
E
But
at
that
point
in
time
you
know
the
the
evidence
that
has
been
produced
by
the
testing
entity
is
not
is
not
fresh,
because
maybe
you
have
compromised
it
in
the
meantime
but
the
but
the
appraiser
receives
the
evidence
which
looks
good
right.
D
Yeah
so
that
basically
you're
opening
up
your
replay
attack
window
yeah
indefinitely
to
for
the
duration
of
the
delay.
A
I
I'm
happy
with
this
sentence.
You
know
I'm
happy
with
with
this
with
this
sentence.
E
G
E
Says
you
know
I
I
I
failed
to
see
how
you
can
effectively
block
something
like
this.
E
D
And
not
talk
about
how
you
can
throw
out
the
attack,
let's
say:
hey,
there's
an
attack
that
you
use
again
left
is
an
exercise
to
the
reader.
Well,.
E
I
don't
know,
I
don't
know
if
this
is
acceptable.
Okay,
I
mean
I
thought
we
would
security.
Consideration
should
at
least
state
clearly
that
this
attack
is
not
addressed.
So
I
mean.
A
D
Yeah
you're
right
because
without
that
sentence,
you're
not
actually
mitigating
the
delay,
threat,
you're
only
mitigating
the
ghetto
one
yeah,
because
it
says
to
mitigate
these
threats
to
transport.
It
should.
A
Let's
not
claim
a
mitigation
here
that
we
can't
live
up
to
and
I
think
we
should
come
back.
If
someone
comes
up
with
a
solution,
then
we
should
come
back
to
it.
E
Okay,
but
but
at
least
you
know
having
a
transport
that
is
sensible
to
reorder
and
drop
and
tells
you
you
know,
you
can
do
something,
you
can
build
a
strategy,
it's
a.
A
Different
strategy,
on
top
of
that,
but
they're
saying
if
I
use
tcp
or
tls,
which
are
yeah
reordering
that
yeah,
then
I
I
my
transport
deals
with
that
problem
with.
E
Attack
not
the
delay
attack,
but
the
the
reorder
of
the
ordering
attack.
Yes
right.
D
D
What
do
you
call
it?
The
periodic
thing
when
there's
no
data
keep
alive,
keep
a
lot
yeah.
If
you
have
keep
the
lives
turned
on,
you
can
detect
drops,
but
if
you
don't
have
keep
lives,
you.
A
A
The
ability
to
detect
reordering
and
draw
tcp
can
do
this
t
tls
definitely
does
this.
D
G
D
I
A
E
A
D
A
E
No,
it's
misleading,
because
it
seems
like
everything
before
that
can
be
mitigated
by
this
by
a
transport
that
behaves
like
that,
but
it's
not
true.
E
Okay.
The,
however,
was
for
a
purpose
which
is
you
know
there
is
a.
There
is
an
attack
that
you
cannot
so
we
should
say
explicitly.
The
attack
above
is
not
addressed
by
such
transport,
though.
G
D
Can't
be
throated
and
put
that
as
the
same
paragraph,
because
then
it's
clear
that.
A
G
A
E
G
D
H
D
We
still
have
other
pull
requests
with
comments,
and
so
in
your
criteria
of
just
merging,
some
of
them
will
fail,
which
require
discussion.
Somehow.
A
Yeah
so
yeah,
so
I
don't
know
I'll
just
pick
three
of
them
see
what
they
say.
D
A
A
Marks:
okay,
assuming
that
well
mcr,
I
want
so
I'm
I'm
one
of
the
authors.
So
that's
what
I
would
I'd
rather
have
three.
But
if
I
need
to
count
the
author,
then
that's
fine!
Okay,
because.
G
H
We
left.
D
Everybody
like
shares:
that's
you
plus
somebody
else's
thumbs
up
plus
my
plus
one,
so.
A
D
D
A
Okay,
all
right
so
we're
meeting
on
tuesday-
and
I
I
hope
we'll
have
only
the
ones
that
only
issues
that
have
comments
will
be
still
open.
So
that
would
be
five
or
six.
D
A
Let
me
let
me
go
through
the
last,
the
things
and
and
if
I
feel
that
we
can
do
it,
then
I'm
going
to
make
a
submission
today.
A
What
I
really
hope
is
that
the
shepherd
will
start
their
write
up
and
our
ad
will
start
and
that
we
may
have
one
or
two
issues
that
we
still
need
to
close,
but
I
hope
they're,
trivial
and
and
then
recognizing
that
our
shepherd
might
raise
some
issues
anyway.
So
that's
the
goal,
but
I
would
like
to
have
a
stable
document
for
them
to
read.
You
can
always
generate.
A
That's
right
exactly
that's
my
point.
We
can
always
do
another
one
on
tuesday
and
we
may
we
may
look
at.
We
may
look
at
whatever
is
remaining
on
tuesday
and
go
oh.
These
are
just
you
know,
wordsmithing,
so
we're
not
going
to
worry
about
it
and
we'll
wait
for
the
next
round
of
review.
A
Okay
cool,
so
I'll
just
mention
for
those
of
you
who
aren't
from
maybe
aren't
familiar
with
the
the
process
that
we
wind
up
with
a
shepherd
write-up.
Who
may
make
some
comments?
It
goes
to
our
area
director,
which
is
roman,
who
will
probably
do
a
fairly
large
amount
of
review,
and
then
it
goes
to
the
iesg
and
if
for
lucky,
then
there
will
be
nothing
other
than
what
roman
has
said
to
us.
E
So,
michael
there's
is
there
going
to
be
a
a
quick
working
group
plus
call
on
on
this
or
or
we
go
straight
to
the
scene.
A
So
so
we
should
be
going
straight
to
the
ad,
because
all
the
comments
that
we're
dealing
with
essentially
are
comments
that
were
working
group
last
call
comments,
even
though
that
was
seems
like
it
was
months
ago,
which
it
literally
was,
but.