►
From YouTube: The libp2p overview talk
Description
Discover libp2p, a modular and extensible networking stack which solves many challenges of peer-to-peer applications.
For more information on libp2p
- visit the project website: https://libp2p.io/
- or check out the project GitHub: https://github.com/libp2p
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
Lol,
oh
I'm,
the
beep
and
I'm
the
Jessica
TFS,
and
justly
peer
to
peer
tech,
lead
and
I.
Welcome
you
to
the
first
sleep
you
peer
overview
talk.
It
is
actually
a
preview,
but
one
that
should
bring
a
lot
of
clarity
to
all
the
loop.
Your
peer
contributors,
users
and
stakeholders,
I
welcome
all
kinds
of
feedback.
My
goal
is
to
have
a
talk
that
we
would
be
proud
to
deliver
at
a
tech
conference.
A
So,
let's
get
to
it.
The
agenda
is
quite
straightforward.
I'll
start
by
explaining
the
motivation,
and
why
do
we
need
to
think
about
networking
in
2018
after
that,
I'll
jump
right
into
a
liquid
appear,
what
it
is
and
what
it
is
followed
by
a
progress
report
and
last,
but
not
the
least
I'll
talk
about
what
it
is
in
the
horizon
for
the
project.
A
If
you
have
been
following
a
TFS
for
a
while,
you
might
have
heard
us
describe
all
the
web,
the
platform
that
we
love
and
learn
to
rely
on
it's
pretty
fragile
and
that
it
has
several
design
problems,
most
of
which
come
from
location.
Addressing
and
now
the
solution
is
to
move
to
a
Content
addressed
model.
However,
what
doesn't
get
exclusively
SEM
is
that,
in
order
to
give
content
addressing
properties
to
the
web,
you
have
to
go
deeper
into
the
networking
stack
and
overcome
many
of
the
challenges
that
the
current
Internet
infrastructure
asks.
A
Things
like
firewalls
that
apply
to
restrictive
filters,
mats
that
bind
peers
from
each
other
I
already
see,
networks
that
make
the
experience
horribly,
slow
networks
that
are
just
not
reliable
and
keep
failing
the
lack
of
mobile
addressing
across
different
types
of
networks
and
protocols.
This
is
also
known
as
roaming.
The
fact
that
there
are
government's
that
completely
that
can
completely
shut
down
the
internet
for
entire
countries
and
lots
of
devices
with
very
different
connectivity
properties,
forcing
developers
to
work
really
hard
around
their
limitations
to
get
connected
with
each
other
and
a
lot
of
times.
A
They
are
just
simply
not
able
to
also
really
important
innovation
in
networks.
Land
using
Carlo
takes
years
for
a
new
protocol
to
go
down
the
OSI
stack
and
to
be
deployed
in
a
large
scale,
even
Oryx,
with
huge
amounts
of
funding
might
still
take
years
to
decades
to
deploy
new
fantastic
ideas
to
the
Internet's
fabric.
In
order
to
build
ipfs
the
peer-to-peer
file
system,
we
had
to
come
up
with
ways
to
work
around
these
limitations,
some
of
which
required
to
redesign
pieces
from
the
ground
up.
A
The
fun
fact
is
that
ipfs
is
not
the
first
peer-to-peer
protocol.
Actually,
there
are
multiple
projects
and
some
have
been
around
for
decades.
If
you
have
been
on
the
internet
piece
for
a
while,
you
might
remember
when
Skype
was
originally
launched,
it
was
a
peer-to-peer
voice
service
projects
like
Nutella
as
a
Amul,
BitTorrent
anotherís
enabled
file
sharing
in
a
peer-to-peer
fashion.
A
There
was
a
lot
of
ketamine
research
at
Kapaa,
which
some
of
which
got
implemented,
others
didn't,
and
even
though
these
projects
existed
before
it
was
epically
hard
to
really
leverage
the
work
done
in
the
past.
Most
of
these
projects
had
the
classic
software
visibility
problems.
Things
like
lack
of
widow,
mutation
or
not
at
all
licensing
was
too
restrictive
for
nowhere
to
be
found.
The
projects
were
old,
some
of
them
really
old,
with
a
last
update
more
than
a
decade
away,
and
they
simply
didn't
work
in
modern,
OSS
or
modern
language.
A
Runtimes
gnome
intended
to
be
seen
the
source
was
closed
or
the
source
didn't
exist
anymore.
This
is
very
common
in
academia
where
the
source
is
not
seen
as
something
valuable
and
typically
gets
thrown
away.
After
the
paper
gets
published,
the
specification
was
missing
and
the
the
implementations
didn't
expose
a
really
friendly
API
that
was
easy
to
use
and
play
with.
Most
of
these
projects
were
tightly
coupled
with
a
specific
use,
making
it
very
hard
to
that,
and
this
is
something
that
we
still
see
today.
A
One
example
is
where
to
see
the
peer-to-peer
transport
of
the
web
platform,
whoever
this
is
so
incredibly
hard
piece
together
that
more
than
five
teams
of
open-source
hackers
and
or
companies
after
ID,
and
continue
trying
to
extract
the
source
code
from
the
Chrome
browser
and
we
package
it
to
other
runtimes
still
today,
most
of
those
implementations
are
quite
buggy.
They
are
not
well
maintained
and
simply
not
reliable.
A
Another
common
pattern
across
these
past
projects
is
that
there
weren't
a
lot
of
assumptions
made
probably
for
the
sake
of
a
product
a
bit
at
a
time,
but
now
it
makes
them
really
hard
to
that.
So
there
had
to
be
a
better
way
and
I
know
what
you're
thinking
the
famous
XT
CD
97
on
the
standard
story,
but
no
think
about
it.
A
What
if,
instead
of
like
creating
a
new
standard
and
new
projects
that
it
is
competing
with
the
old
ones,
we
actually
created
a
project
that
is
a
frame,
a
toolkits
full
of
options
to
integrate
all
the
previous
ones
and
that
wants
to
exist
in
the
future.
Well
enter
a
peer
to
peer.
We
peer
to
peer
is
a
module
module.
A
Teams
and
stakeholders,
a
simple
way
to
think
about
with
your
peer,
is
if
a
PFS
brought
content
addressing
to
the
world
where
peer
to
peer
brings
process
addressing
a
way
to
find
connect
and
authenticate
processes,
so
that
we
know
we
can
trust
the
connection.
End-To-End
we
prepare
is
a
collection
of
building
blocks
that
expose
very
well-defined,
documented
and
tested
interfaces
so
that
they
are
peace,
composable
swappable
and
with
that
upgradeable.
A
Another
way
that
I
like
to
describe
it
is
by
saying
it.
We
grab
the
OSI
vertical
layer
system
and
made
it
horizontal
by
the
way,
just
to
page
in
everyone's
mind,
the
OSI
layer
system
out
there
being
the
one
still
taught
in
many
networking
classes,
doesn't
actually
represent
where
all
the
internet
works,
tcp/ip
got
deployed
worldwide,
and
so
it's
model
it's
a
better
representation
either
way.
Both
of
them
are
just
conceptual
and
in
practice
they
do
have
shortcomings
such
as
multiple
action
that
I
repeated
through
many
of
the
layers
duplicating
work
and
wasting
resources.
A
Information
gets
even
missing
significant
opportunities
for
optimizations.
If
you
are
new
to
this
idea,
I
strongly
recommend
you
going
through
the
RFC
1958
and
IRC
34:39
on
internet
architecture,
principles
and
guidelines
to
improve
their
understanding.
The
summary
is
that
what
happens
today
that
the
Internet
Protocol
spread
over
multiple
areas,
which
made
things
very
tightly
coupled
internally
and
hard
to
change
and
to
innovate
with
peer-to-peer
changes
this
and
just
gives
you
a
frame
in
which
all
these
protocols
can
co-exist
and
cooperate.
A
You
pick
your
own
liquid,
appear
recipe
and,
if
I'll
at
a
time
you
want
to
change,
you
just
have
to
change
the
recipe
by
the
top-level
API.
The
link
to
the
PAPR
remains
the
same,
so
your
app
doesn't
need
to
change.
A
good
example
of
this
in
practice
is
our
PMS
runs
in
node.js
and
in
the
browser
here
you
can
see
each
recipe.
A
The
differences
are
that
both
runtimes
at
different
capabilities,
different
transports,
and
so
they
energy
as
we
get
to
use
TCP
and
WebSockets.
Well
in
the
browser
we
use,
WebSockets
and
I
know
we're
were
to
see
to
learn
more
about
each
of
these
pieces
and
how
they
work
and
how
they
pull
together.
You
can
check
the
very
comprehensive
tutorial
suggest
we
put
the
peer
repo,
these
back
explanations,
with
code
samples
that
you
can
try
on
your
own.
So
this
is
the
peer
to
peer.
What
about
progress?
How
far
have
we
gotten
so
far?
A
We
have
three
language
of
limitations,
one
in
go,
one
a
rest
and
one
in
JavaScript
and
one
slowly
in
the
works
in
c-sharp
by
a
community
member
and
many
pieces
built
in
multiple
other
languages,
thanks
to
booty
formats,
the
self-described
about
data
types
that
are
extensively
used
in
the
peer-to-peer.
The
community
is
growing
fast,
with
over
70
contributors
dedicated
to
peer
to
peer
and
over
a
thousand
people
that
have
participated
in
the
project.
A
Somehow
he
threads
on
the
browser,
and
this
was
kind
of
the
test
of
fire
for
a
leap
year
and
in
Division
it
also
runs
on
mobile
browsers,
which
super
exciting.
It
has
enabled
apps
like
collaborative
docket,
leading
thanks
to
the
p2p
publish/subscribe
API
for
real-time
communication.
We
demoed
early
peer-to-peer
connecting
going
through
your
notes
to
a
TV
movie,
ms
running
on
the
browser
using
the
liquid
appear
to
transport
the
blocks
from
the
atrium
chain.
A
There
are
also
new
products
that
do
peer-to-peer
live
streaming
using
the
peer-to-peer
and
our
friend
at
Mathematica
doing
tons
of
research
on
Alto
ever
trip.
It
appeared
the
Libre
blocks
faster
to
their
giant
user
network
and,
of
course,
any
project
that
is
using
IP
FS
is
using
a
peer-to-peer
by
design
and
in
the
future.
A
So
let's
talk
about
the
timeline
and
disclaimer,
the
timeline
is
highly
speculative.
It
is
a
mix
of
what
is
possible
and
what
are
our
wildest
hopes
and
dreams.
It
only
serves
the
purpose
of
giving
some
notion
of
direction
so
use
it
with
Fatima
somewhere
in
the
next
three
months.
The
priority
is
to
support
the
three
language
teams
and
help
them
achieve
feature
priority
with
robust
integration
and
interoperability
testing.
That
should
be
projected
all
over
up
to
date.
Specification
of
will
appear
also
to
improve
the
improbability,
debugging
ability
and
overall
excitement.
A
We
want
to
have
a
a
way
to
visualize
different
Network
topologies.
This
will
also
essential
to
foster
innovation.
Six
months
in
is
where
we
see
that
like
well,
adoption
will
kick
in.
We
we
want
to
see
new
language
of
limitations
or
even
better
will
appear,
appear
getting
package
on
the
language
or
at
times.
So
it's
just
there
for
every
single
developer.
We
also
want
to
have
a
packet
switching
API.
A
Today,
all
the
ideas
are
stream
based
which
we're
well
for
the
use
cases
appear
today,
but
for
Iowa
to
see
in
low
bandwidth
connections
and
really
long
reliable
connections.
Packet-Switching
can
bring
significant
improvements
to
the
experience
of
the
users
that
usually
peer-to-peer,
and
yes,
we
also
want
to
get
some
new
transports,
such
as
booted
more
in
the
future.
We
want
to
make
we
prepare
OS
level
service
and
a
kernel
module.
We
want
to
promote
tons
of
innovation.
A
Networking
go
to
university
networking
courses,
researchers
and
labs
and
help
them
explore
it
in
depth
and
build
on
top
of
it.
We
also
going
to
want
to
go
into
radio
and
use
all
the
latest
and
greatest
software
defined
radio
software
to
create
radio
modules
for
a
lip-reader
peer.
The
long
term
goal
is
really
to
make
people
appear
that
go
to
peer
to
peer
and
networking
stack
and
convince
the
world
that
there
is
no
need
for
the
client
and
server
model
and
all
of
its
limitations.
A
Infrastructure
is
more
reliable
when
it
is
built
with
protocols
that
are
fully
distributed
and
can
adapt
to
many
networking
scenarios
still
lots
of
work
to
be
done,
but
something
that
we
can
do
together.
Thank
you.
So
much
for
your
time.
I
hope
this
presentation
was
valuable.
Please
send
your
feedback
to
David
at
protocol
AI
thanks.
So
much.