►
From YouTube: IPFS Cluster Update - Hector Sanjuan
Description
Originally recorded during the Berlin Developers Meetings from July 9-13, 2018.
A
Hi
everyone,
my
name,
is
hector
seung-wan
and
I'm.
The
lead
for
ipfs
cluster
I'm
gonna,
give
you
a
quick
update
of
what
happens
in
the
last
quarter,
and
I
would
like
to
expand
on
Cluster
in
in
one
conference
session
a
bit
longer
to
talk
about
the
internals
and
all
the
pieces
in
the
ipfs
world
that
we
use
and
all
the
cool
things
that
we
do.
But
let's
start
with
this
I'm
gonna
tell
you
what
cluster
does
and
what
it
wants
to
do.
A
A
Pushing
content
to
different
ipfs
demons
in
your
network
and
replicating
them
to
serve
our
replication
factors.
We
choose,
which
peers
are
the
most
the
better
suited
for
the
content
based
on
the
free
disk
space
we
provide
api's.
We
provide
authentication
on
the
on
those
api's.
We
provide
we
exposed
to
them
through
the
p2p.
We
can
attach
HTTPS,
so
quite
versatile
and
quite
easy
to
integrate
with.
A
We
provide
what
we
call
composition,
which
is
a
proxy
to
the
ipfs
daemon
underline
the
idea
of
caste
appear,
which
means
that
your
cluster
can
behave
like
an
IPSS
demon
or
expose
ni,
BFS
compatible
api.
We
have
awesome
documentation.
We
worked
really
hard.
Last
four
in
setting
up
a
website
and
updating
documentation.
We
have
guides
all
everything
is
documented
in
much
data
and
what's
something
that
we
do
as
a
team
is
also
running
the
storage
cluster
and
the
bean
pods
for
the
community.
A
A
We
use
pops-up,
we
use
ipfs,
we
usually
p2p,
we
use
IP
le
now,
so
these
are
all
things
which
we
are
testing
on
top
and
this
allows
you
to,
and
it
has
happened
before,
uncover
things
that
we
need.
It
will
be
good
to
have
in
these
libraries
below
and
things
which
are
not
working
and
that
we
can
report
and
fix
for
everyone
in
the
community.
A
Managed
to
retake
the
dag
shard
in
branch.
This
is
a
feature
of
cluster
that
will
allow
you
to.
If
you
have
something
that
is
too
big
to
store
in
a
single
ipfs
daemon,
it
will
be
divided
across
your
cluster
ipfs
demons.
It's
really
cool.
We
need
a
couple
things
from
God
BFS
and
we
submitted
those
so
we're
close
to
having
no
blockers
to
actually
finish
this.
A
As
iwe
finish,
the
eye,
PFF
pastor
website
looks
wrinkled.
Thank
you.
I
gotta
for
the
design
and
I
had
ruined,
joins
a
cluster
team,
so
that
meant
I'm
bored
in
making
him
a
productive
member
of
the
whole
thing
and
welcome
it's
great
to
have
you
on
board
in
the
short
term
and
I
don't
want
to
make
many
promises.
That's
why
I
reduce
this
to
the
minimum,
kill
three
will
be
dark.
A
Shardene
I
will
be
spearheading
the
monitoring,
exposing
metrics
and
doing
things
about
how
cluster
beers
monitor
each
other
and
distribute
the
internal
metrics
and
so
on.
Adrian
will
be
working
on.
The
Michelle
will
be
doing
user
research
and
outreach
figuring
out
where
our
users
want
reaching
out
to
users
trying
to
have
people
run.
Cluster
and
see
how
we
can
support
them,
I
will
skip
that.
I
will
skip
that.
I
would
get
that
yeah,
because
those
are
hidden
you
see
they
shouldn't
show,
even
in
presentment,
in
the
mint
midterm.
We
want
better
composite
clusters.
A
We
want
collaborative,
pin
sets
which
is
basically
clusters
that
can
scale
to
the
thousands
and
we
won
since
we
will
have
more
metrics
on
monitoring.
We
want
to
start
measuring
how
long
it
takes
to
ingest
data
and
to
distribute
it
using
cluster
and
using
IP
FS,
more
automation
and
deployment
for
people
that
run
cluster
an
idea
for
some
production.
How?
A
How
can
they
automate
deployment
of
these
tools
in
their
data
centers
in
the
very
long
term,
forward,
error,
correction
and
other
storing
methods
that
will
allow
you
to
efficiently
store
your
data
and
replicate
it
and
being
able
to
basically
backup
very
large
amount
of
data
and
that
cluster
scales
the
way
it
needs
to
scale
I,
think
that
was
it.
Thank
you
very
much
I'm
happy
to
answer
any
questions.