►
From YouTube: 🙌 IPFS Weekly Call 📡 💫 2020-01-13
Description
Today introducing IPFS Collaborative Clusters
For more information on IPFS
- visit the project website: https://ipfs.io
- or follow IPFS on Twitter: https://twitter.com/IPFS
Sign up to get IPFS news, including releases, ecosystem updates, and community announcements in your inbox, each Tuesday: http://eepurl.com/gL2Pi5
A
A
A
A
A
Let's
start
by
saying
that
we
had
a
new
cluster
is
a
new
idea.
First,
cluster
release
by
the
end
of
December,
and
this
release
adds
a
new
binary.
A
new
application
called
IPS
cluster.
Follow
this
application
is
made
to
make
it
very
simple
for
anyone
to
join
what
we
call
a
collaborative
cluster.
A
collaborative
cluster
is
an
IDF
s
cluster,
which
allows
random
people
to
join
that
cluster
and
participate
in
to
Amin
what
that
cluster
is
supposed
to
pin.
A
A
The
interesting
thing
there
is
that
all
these
clustered
beers
that
run
side
by
side
with
the
IPF
SP,
you
will
join
a
network,
usually
a
private
network
but
as
I
say
with
collaborative
clusters,
a
parallel
network
that
potentially
anyone
can
join,
and
they
will
use
that
network
to
agree
on
a
common
state
and
that
common
state
is
a
shirt
against
it.
It's
a
shared
list
of
items
that
they
are
supposed
to
be
pinning
each
on
on
their
own
idea
of
estimate.
A
The
result
is
that
those
items
are
not
only
the
CID
of
the
thing
that
they're
supposed
to
pin,
but
they
may
also
contain
additional
information.
For
example,
faster
pins
have
an
names
attached.
They
can
have
something
like
expiration
dates
in
which
they
will
be
automatically
unpinned
and
they
have
replication
factors.
A
Replication
factors
may
not
have
too
much
sense
in
collaborative
clusters,
because
you,
you
cannot
trust
whether
people
are
effectively
being
content
or
not.
For
that
we
have
five
own,
which
runs
great
lengths
into
ensuring
that
people
are
storing
what
they're
supposed
to
pin.
But
replication
factors
have
a
lot
of
sense
when
you're
running
a
cluster
of
peers
that
you
actually
control-
and
you
want
with
this
review-
how
many
of
your
peers
should
actually
pin
certain
content.
B
A
A
They
have
two
trusted
peers,
so
two
beers,
which
are
able
to
modify
the
pin
set
in
that
cluster,
two
clues
that
we
control
and
for
each
of
those
clusters
we
have
added
an
archive
which
is
not
too
big
and
which
we
consider
interesting.
For
example,
the
file
going
burns
which
are
necessary
to
launch
the
piping,
the
Lotus
implementation
on
a
test
meant
Network.
A
We
added
some
books
in
Spanish
from
the
Gutenberg
Project,
which
has
three
mats
very
good
selection
of
Fremont's.
Every
book,
which
is
free,
has
no
copyright
or
anything
in
the
world.
But,
of
course,
that's
a
very
sizable
archive,
so
I
started
adding
the
Spanish
section
of
it,
and
also
we
have
another
cluster
for
ipfs
websites,
which
means
I
BFS
I,
elderly
p2p
website,
the
cluster
website,
the
docs
website
and
so
on
and,
as
you
can
see,
the
yummy
necessarily
seem
to
join.
A
This
clusters
is
to
download
the
new
ipfs
custard
follow
application
which
is
available
in
this
ipfs
io,
along
with
go
idea,
face
and
the
other
applications
and
run
in
a
single
command,
which
is
a
TFS
faster,
follow
five
going
run
and
in
and
the
image
URL
I'm
going
to
blame.
What
this
that's
particularly
time.
I
can
do
that,
for
example,
by
showing
you
life
on
screen,
which
is
essentially
more
interesting,
so
I'm
gonna
stop
this
share
and
I'm
gonna
share
a
terminal.
A
A
A
The
latest
fashion,
and
now
I
can
run
it
if
I
run
in
without
any
arguments,
you
will
see
that
it
says
I
have
confirmation
for
two
for
our
viewers.
This
is
one
nice
in
the
IDF
asked
us
to
follow
is
that
it
allows
you
to
actually
run
several
custard
years,
several
follower
tiers
in
parallel,
so
that
you
can
be
tracking
several
archives
at
the
same
time,
and
I
can,
for
example,
remove
the.
A
If
I
run
a
cluster
follow
now,
it
just
tells
me:
I
have
one,
and
the
only
thing
I
need
to
do
now
is
to
go
to
co-op
ipfs
classes
that
I
own
and
copy
paste
the
command,
but
this
is
going
to
do
is
to
initialize
my
cluster
beer,
give
it
an
identity,
and
after
that
it
is
going
to
fetch
a
configuration
from
this
cluster
here
from
Gutenberg.
Yes,
collab
ipfs
cluster
that
IO,
how
does
it
fetch
it?
It
is
not
actually
fetching
through
HTTP
HTTP.
A
This
is
configured
as
a
DMS
link,
so
it's
going
to
resolve
to
an
IP
office
address
and
it's
going
to
use
ipfs
to
actually
fetch
that
configuration
that
will
make
my
beer
run
with
the
optimal
parameters
well
optimal
with
the
necessary
parameters
to
connect
to
this
cluster,
in
particular.
In
order
to
connect
to
this
custom
in
particular,
we
need
a
Buster
secret
and
I
will
need
the
list
of
trusted
peers
for,
for
the
tester
and
I
will
need
the
addresses
of
those
peers
alike
so
that
I
can
connect
and
I
can
bootstrap
to
them.
C
A
A
C
A
Good,
it
seems
that
was
more
likely
than
as
we
see
and
the
block
messages
we
initialized
appear
and
it
just
launched
a
class
appear
be
connected
to
to
the
truss
type.
You
said
it
is
right
now
sinking
all
the
pins
in
this
archive.
A
There
should
be
around
400,
I,
believe,
400
and
something
you
can
do
this
any
time.
So
peers
can
come
and
go
anytime
if
I,
if
I
stop
the
beer
now
I
could
too
restarted
and
I
would
get
up
to
the
latest
version
of
the
archive
and,
if
I
send,
that
to
the
background
and
I
run
I
get
those
clustered
follow.
A
Now
you
see
that
I
had
ideas
classic
follow
good
number
guess,
which
is
again
reconfigured
and
if
I,
if
I
call
this
command
with
that
tracer
name,
it
will
I
will
be
able
to
perform
actions
directly
on
that
cluster.
So,
for
example,
I
can
list
all
the
pins
which
are
listed
on
that
Chester.
The
screen
is
a
bit
narrow,
but
do
you
see
that
it
says
the
state
pinned
the
CID
and
then
it
would
list
the
pin
name
which
corresponds
in
this
case
to
the
name
of
the
book
that
we
are
finan?
A
If
I
do
an
info
command,
I'm
able
to
see
information
for
this
cursor
here,
specifically
like
where
the
confirmation
folder
is
where
the
configuration
source
URL
is
and
whether
I
am
online.
The
funny
thing
other
good
thing
here
is
that
I
can
actually
stop
the
beer.
All
the
arrows
that
you
see
is
because
it
didn't
finished
pinning
on
ipfs,
didn't
finish
going
through
that
queue
of
400
items,
but
I
can
run
info.
Also,
when
my
follower
Pierre
is
not
running,
and
in
that
case
it
will
tell
me
that
the
cluster
beer
is
not
online.
A
The
ipfs
beer
is
online,
though,
and
I
cannot
also
list
the
items
in
the
archive
while
I'm
on
of
offline,
so
I
don't
need
to
run
my
gear
to
know
what
items
that
cluster
beers
in
it
and
those
are
interesting
things
to
do,
because
you
yeah
you
can
basically
inspect,
but
the
archive
is
about
without
having
to
relaunch
your
beer.
So
I
say
you
can
stop,
and
you
can
start
start
your
gear
again
with
with
the
wrong
command.
A
You
will
be
able
to
connect
to
them.
You,
the
the
ipfs
forms,
will
probably
be
connected
directly,
but
those
peers
don't
have
access
to
you
to
making
operations
in
the
cluster
or
accessing
information
from
your
peer
directly.
For
example,
after
appear
that
you
do
not
trust,
it's
not
able
to
know
whether
you
are
pinning
something
at
a
certain
moment
or
three
here,
any
operation
in
you
or
obtain
any
information
other
than
your
cluster
peer,
ID
and
so
on.
So
the
the
trusted
peers.
However,
they
can
just
like
they
can
been
using
your
passive
view.
A
They
can
also
potentially
check
whether
you
are
pinning
something
or
you
have
an
error
when
pinning
something
or
are
you
finished
at
this
point,
I
would
like
to
open
the
room
for
for
questions.
I'm
gonna,
stop
the
this
green
shirt
and
I'm.
Sorry
for
the
little
problems
that
we
had.
I'm
gonna
have
to
verify
that.
C
A
It
is,
it
is
a
list
of
peers
in
the
configuration,
but
the
thing
here
is
that,
since
we
use
these
remote
confirmations
that
you
fetch,
you
fetch
that
confirmation
every
time
you
start
you
don't
fetch
it
only
when
you
initialize,
therefore,
the
stores
can
update
the
compression,
basically
budget
right
you.
We
can
basically
add
new
beers
to
that
configuration
and
whenever
the
followers
are
restarted,
they
will
pick
up
the
changes
and
I.
A
There
is
we
will
use
in
DNS
link
for
that
yeah.
So
we
we
update
the
DNS,
but
if
we
don't
update
the
DNS,
then
you
will
essentially
reserve
to
us
yet
I
see
any
that
you
already
had
locally
in
your
ipfs
demon.
So
you
don't
have
to
you.
Don't
even
pay
the
price
of
of
going
to
some
remote
server
to
fetch
your
confirmation,
you
you're,
basically
loading
local.
C
A
A
But
we
get
from
that
is
that
you
can
sync
the
CR
DT
from
any.
You
can
sync
the
state
from
any
of
the
participating
creation
by
just
fetching
back
the
bad
thing.
There
is
that
here
essentially
building
a
block
chain.
Therefore
it
will
tend
to
grow
the
more
things
you've
been
in
the
cluster,
the
no
longer
this
chain
will
be
and
there's
where
you
need
like.
Potentially,
you
will
need
like
more
optimizations.
D
A
Fact
is
that
American
mecha
docks
are
essentially
a
structures;
they
only
grow.
So
just
like
rock
change,
you
always
append
at
the
root.
Therefore,
they
are
very
similar
to
grow.
Only
sets
and
grow
only
sets
they're.
Essentially,
our
only
set.
This
is
that
you
interpret
it
and
traverse
it
in
certain
ways.
Grow
only
sets
are
well-known.
Civt,
the
characteristic
of
all
CRE
disease
is
that
they
are
they
inflate
the
the
states,
they're
usually
don't
get
smaller.
They
inflate
at
least
four
state-based
realities
for
operation
based
realities.
A
You
need
certain
characteristics
in
your
network,
balancer,
operation,
delivery
and
so
on
for
a
state-based
realities.
You
have
things
which
basically
inflate-
and
this
is
similar
to
a
state
by
see
oddity
except
it
is
synced
more
efficiently,
taking
advantage
of
with
the
ipfs
and
and
the
fact
that
you
only
need
to
to
ship
the
new
blocks
around,
not
the
whole
thing.
Every
time.
D
A
It's
fat
underneath-
and
this
is
a
stack
which
is
its
idea
first
light.
So
it's
bit
surplus,
DHD
plus
pops
up
on
some
bucks,
a
piece
used
to
to
propagate
the
new
when
you
pin
something
there
will
be
a
broadcast
and
the
new
cid
is
broadcasted
and
then
every
peer
in
the
cluster
will
use
beats.
Fact.
If
you
copy
the
new
blocks,
it's
essentially
ipfs.
It's
essentially
an
application
on
top
of
all
the
ipfs
things.
B
A
B
A
A
A
What
that
will
do
is
to
choose
the
peers
in
your
cluster,
which
have
most
free
space,
as
reported
by
ipfs,
so
there's
an
a
storage
max
confirmation
value
in
ipfs.
So
knowing
how
many?
How
much
of
the
repo
storage
we
have
used
and
what
that
compression
value
is,
can
calculate
how
much
ipfs
free
space
you
have
and
it
will
rank
them
according
to
free
space
and
I.
Will
it
will
assign
based
on
that?
A
So
if
you
say
your
replication
factor,
3
will
pick
up
3
3
beer
sweets
with
the
most
free
space
and
send
the
pins
to
those
specifically.
But
as
I
say,
if
you're
running
in
a
untrusted
open
cluster,
it
will
be
very
easy
for
someone
to
simulate
other
opinion,
something
and
they're
actually
not
or
to
simulate
that
they
have
infinite
free
space.
So
we're
just
saying
now
everyone
pins
everything,
and
we
don't
do
that.
Selection.