►
Description
Every other week, the Retrieval Market Builders get together to share progress in their projects in a demo format. We want to sincerely thank all of our collaborators for demoing their developments and helping to improve the FIL Retrieval Markets one demo at a time.
Novem 24, 2021, Demo Day: In this video, you can find quick demos regarding the following -
• Magmo discussing scalable multi-hop payments
• KEN Labs presenting Pando, a metadata service
__________________________________________________________________________________________
If you would like more information about the Demo Days please follow the link below to the Demo Days planning document.
https://docs.google.com/document/d/1IrNbBm2_P79Xi2fXFj9YUfQ0PjWjreGdPYFjFntD1QM/edit
A
Okay,
welcome
to
this
week's
retrieval
markets
demo
day.
Today's
date
is
wednesday,
the
24th
of
november
today's
schedule.
We
have
andrew
stewart
from
magmo
introducing
the
team
and
what
they're
working
on,
and
then
we
have
whaley
yang
from
ken
labs
talking
about
pando
metadata
service,
the
new
work
that
they're
looking
at.
A
B
Okay
yeah,
so
let's
dig
in,
as
I
mentioned,
I
finished
these
slides
at
11
58.
So
please
bear
with
me,
but
today
I'm
going
to
I'm
mainly
going
to
talk
about
magmo
and
the
work
that
we've
done
in
the
past.
But
I
will
also
talk
about
some
say
some
important
things
about
the
project
that
we're
working
on
here,
which
is
implementing
scalable
scalable
multi-hop
payments
for
the
filecoin
retrieval
market.
B
So
I'll
start
with
a
brief
introduction
to
the
team.
So
magmo
is
myself
andrew
george,
mike
alex
colin
tom
and
connie,
you
can
find
out
more
information
about
us
at
megmo.com
and
in
short,
I
would
say
that
we're
a
team
of
we
try
to
be
humble
research
and
development
engineers
and
we're
also
good
friends
and
we're
willing
to
take
risks.
Germinating,
stretchable
research
ideas,
we're
always
looking
for
novel
ways
for
peers
to
securely
interact
among
themselves
leveraging
a
public
blockchain
as
a
trust
layer.
B
We
tend
to
focus
on
layer,
two
technologies,
I'm
trying
to
identify
user
pain
points
and
alleviate
them
without
sacrificing
security,
and,
more
importantly,
I
think
something
that's
really
special
about
us
is
that
we
invite
conversation
and
collaboration
with
anybody
as
long
as
it's
in
line
with
our
mission
and
our
values
and
a
little
bit
more
about
what
we
don't
do,
we
try
not
to
hold
any
egos
about
credit
care
about.
You
know
who
came
up
with
what
idea.
In
our
opinion,
our
ideas
are
your
ideas.
B
Also,
we
don't
aim
to
commercialize
our
output,
so
our
your
success
is
our
success.
To
reiterate
here,
our
goal
is
to
improve
the
security
and
usability
of
technology
used
in
decentralized
projects,
so
all
of
our
code
should
be
open
sourced.
We
might
have
forgotten
to
license
something
if
you
find
something,
please
raise
an
issue
on
github
and
we'll
get
at
it
and
I'll.
B
Also,
you
know
put
a
point
out
that
if
anybody
does
want
to
take
our
ideas
to
market
as
this
filecoin
project
demonstrates
we're
gonna
support
you
in
any
way
that
we
can-
and
you
know,
as
I
mentioned,
we
don't
really
polish
our
marketing.
So
you
know
if
somebody
wants
to
do
some
free
visual
stuff
for
us
that'd
be
great
too
now
I
I
know
that
this
is
a
retrieval
market
demo
day,
but
as
you'll
find
out.
B
I
feel
that
I
should
talk
a
little
bit
more
about
myself
and
some
things
that
have
happened
to
me.
So
to
start
with,
I'm
from
a
small
town
in
outport
newfoundland,
it's
on
the
east
coast
of
canada,
it's
a
big
island,
but
everybody
thinks
it's
small
and
somehow,
at
an
early
age
I
got
really
interested
in
mathematics.
B
I
picked
up
a
book
called
cosmos
by
carl
sagan
and
at
the
in
one
of
the
appendixes
he
proved
that
square
root
of
two
was
irrational.
I
remember
just
you
know
it
blew
my
mind
that
you
could
do
things
like
that,
but
I
didn't
know
that
you
could
be
a
mathematician
until
I
met
a
guy
who's.
Also
named
andrew,
his
name
is
andrew
kritsch
and
we
met.
A
B
A
math
camp
back
in
newfoundland.
I
also
didn't
know
that
there
were
math
camps.
So
I
have
to
thank
my
teachers
for
kind
of
putting
pushing
me
in
that
direction
and
eventually
I
earned
a
phd
in
probability
theory,
which
is
key
here,
because
a
recent
series
of
events
has
shaken
my
entire
foundational
beliefs
about
statistics,
and
I
can't
really
help
myself
from
telling
the
next
story,
which
is
how
I
became
a
probabilist
and
what
is
going
on
right
now.
B
B
B
Okay,
that's
how
andrew
kritsch
thinks-
and
he
taught
me
to
think
that
way
and
now
I
can't
stop
it.
What
else
does
andrew
kritsch
do?
Well,
I,
I
have
to
admit
we
fell
out
of
contact
for
a
couple
of
years,
but
it
shocked
me
when
yesterday
one
of
magmo's
teammates
colin
kennedy
pointed
out
to
me
that
he
just
gave
a
talk
at
the
funding.
The
common
summit
hosted
by
no
other
than
protocol
labs,
so
he
was
talking
about
this
s
process
funding.
What
is
s
process
funding?
B
What
that
means
is
that
various
network
participants
such
as
alice
bob,
charlie
and
david
here,
can
connect
to
some
hub
through
some
on-chain
interaction
and
afterwards,
after
they've
made
these
connections
they
can.
They
can
quickly
and
securely
form
off-chain
connections,
allowing
them
to
transact
directly
between
each
other
using
nitro
protocol.
B
If
the
liter
detects
that
they
asked
the
if
the
seeder
detects
that
they
asked
the
leacher
for
some
pieces
of
data
and
the
leacher
refused
to
provide
them,
this
incentivizes
leechers
to
be
to
also
be
seeders
and,
as
I
remember
I
think,
cedars
are
also
also
altruistically
upload
pieces,
which
I
think
is
a
key
part
of
making
bittorrent
function
well.
But
if
anybody
knows
better
than
me,
then
then
please,
please,
let
us
know
so
web3
torrents
kind
of
mutates.
B
A
B
B
B
B
A
B
Testability
of
their
implementations,
you
can
imagine,
for
instance,
if
we
can
compile
these
protocols
to
wasm,
then
we
can
use
them
in
the
browser.
We
don't
have
to
rewrite
them.
We
don't
have
to
worry
about
interoperability
issues,
so
I
guess
I
was
talking
too
quickly,
because
I
thought
I
would
have
to
skim
over
this
slide.
B
But
let
me
let
me
say
a
little
bit
about
what
we've
gotten,
where
we've
gotten
so
far
in
our
go
nitro
package.
The
long
story
short
is
we're
not
a
golang
team.
We
didn't
write
any
golang
before
starting
this
project,
so
the
first
thing
that
we
did
is
we
prototyped
some
of
the
work
included
in
our
grant
proposal
as
a
language
learning
opportunity.
B
I
would
say
we
have
a
substantial
amount
of
the
interfaces
defined
already
and
some
some
rudimentary
prototype
implementations
of
things
like
the
the
store
or
the
messaging.
I
think
somebody
on
my
team
might
correct
me
there,
maybe
I'm
over
promising,
but
we
do
have
so
far.
One
implementation
of
one
of
the
off-chain
protocols
and
what
we're
currently
doing
is
we're
getting
code
and
design
review
from
two
experienced
group
of
golang
developers.
B
Each
group
is
a
domain
expert
in
their
respective
half
of
the
state
channels
and
the
filecoin
equation,
so
we're
hoping
to
maybe
adjust
our
design
based
on
any
of
that
feedback
and
to
wrap
things
up
a
little
bit
here.
I'll
give
you
a
sneak
preview
into
our
next
demo
day,
during
which
one
of
my
teammates
will
give
a
more
detailed
overview
of
where
the
project
is
where
we
plan
to
take
it
and
maybe
even
a
short
demo.
B
I
don't
want
to
make
public
promises,
but
you
know
it
could
happen
and
that
oh
right
there's
one
more
thing.
Well,
we
found
out
yesterday
that
nitro
was
already
taken
by
some
other
high
impact
project,
so
we've
had
to
suffer
a
little
bit
of
a
rebranding
here
with
that,
it's
my
pleasure
to
announce
nitro's.
B
Which
is
the
stateful
asset
transfer
protocol
boring,
but
you
know
whatever
so
this.
This
paper
that
you
see
in
the
slide
here
is
the
end
result
of
a
several
month.
Long
collaboration
with
polycrypt,
who
are
the
maintainers
of
perun,
which
is
another
state
channel
protocol
happen
which
happens
to
be
implemented
in
golang.
B
Together
with
these
folks,
we're
continuously
discovering
protocol
improvements
which
massively
reduce
the
latency
and
overhead
required
to
construct
these
virtual
channels,
especially
across
more
than
one
hub.
In
addition,
we're
increasing
the
the
security
of
these
protocols
and
reducing
the
cost
to
do
the
the
on-chain
operations
necessary
to
recover
your
funds.
B
So,
as
you
can
see
on
the
date
here,
we've
I
we
finished,
I
don't
know
why
it
says
november
24th
and
we
finished
this
paper
at
11
p.m.
Last
night
to
make
a
deadline,
and
we
hope
to
review
the
paper
next
week,
make
some
final
edits
and
we'll
ultimately
release
our
joint
paper
with
polycrypt
and
with
some
more
details,
and
with
that
I
will
thank
everybody
for
the
time.
A
C
C
C
We
provide
interface
to
to
provider
and
consumers
in
general,
the
the
provider
will
publish
their
ipod
data
update
to
pando,
with
all
the
requests
received
and
pando
will
pull
all
the
data
from
providers
over
graph
sync
there.
You
can
see
a
in
a
metadata
service,
a
mod
box,
so
there's
a
rate
limiter.
This
is
to
to
limit
the
traffic
between
providers
and
the
bundle
to
ensure
the
service
quality.
C
C
C
We
have
three
scenarios.
First
of
all,
we
need
to
register
a
provider,
for
example
a
deadbolt.
They
can
use
this
pando
command
line
tool
to
register
itself,
and
then
it
can
send
the
update
to
pando.
So
in
this
demo,
I'm
not
going
to
use
a
billboard.
Instead,
we
build
some
mock
provider
called
back.
It
will
publish
a
a
certain
amount
of
dark
data
pieces
to
the
panel
instance,
the
panda
service,
and
then
we
can
use
a
metadata
query
interface
to
carry
the
state
and
the
snapshot.
Information.
C
This
is
also
what
I
provided
here.
So
this
is
the
idea
provided.
The
idea
is
that
I
want
to
tell
the
pando
service.
I
want
to
register
this
peer
to
the
service
and
I'm
going
to
publish
data
change
to
the
service,
and
this
is
a
parameter
to
tell
where
the
panel
is,
and
this
is
a
provider
address
provider
is
this-
is
the
most
provider,
the
multi-address
of
the
mob
provider
optionally?
You
can
also
provide
a
minor
account.
C
C
C
C
C
C
C
C
C
So
this
is
the
mock
provider
and
let's
go
back
to
the
output
here,
the
more
provider
output.
There
are
five
cid
use
studies,
so
if
we
search,
we
should
see
all
of
them
here
in
the
list.
So
five,
I'm
not
going
to
search
all
of
them,
so
you
can
see
5c
id.
It
forms
its
wrap.
They
are
wrapped
in
in
this
snapshot
and
in
future
we
will
also
enhance
the
snapshot
mechanism
to
support
more
features.
You
don't
have
to
generate
every
five
seconds,
but
for
now
for
demonstration
purpose
we
implement
in
this
way.
C
C
C
C
Okay
and
next
step
is
the
state
query
and
the
snapshot
query.
I
just
show
you
how
to
list
a
snapshot
and
carry
the
detailed
information
of
each
one,
and
this
is
the
graphql
interface
that
we
can
carry
more
information
so
for
state
you
can
understand,
or
you
can
learn
the
the
latest
up
to
status
of
the
all
the
providers.
C
C
C
C
C
C
C
Considering
the
times,
I'm
not
going
to
go
through
all
the
details,
so
in
general,
it's
based
on
token
bucket
algorithm,
and
this
is
the
the
data
flow
so
for
each
request
it
will
check
it
will
try
to
get
a
token
so
until
it
gets
a
token
it
won't
fire
off
to
to
the
the
provider
and
for
the
data.
So
it
will
wait
for
with
every
one
second
and
try
again
until
a
token
is
retrieved,
and
this
is
the
gate
limiter
and
we
also
have
peer
limiter.
C
C
C
C
C
C
C
C
The
last
thing
will
be
the
performance
testing.
We
might
not
be
able
to
finish
all
of
them
in
next
two
weeks,
but
hopefully
we
can
make
three
of
them
the
first
three
on
them,
the
next
two
weeks
so
xiaomi
and
then
then
are
the
main
contributors
to
all
these
feature.
Development
and
the
bug
fixing
really
thanks
to
them.