►
From YouTube: View from the Trenches: Practical Issues In Enterprise Blockchain Implementations and... Baohua Yang
Description
View from the Trenches: Practical Issues In Enterprise Blockchain Implementations and Potential Solutions - Baohua Yang, Oracle
This session will discuss some critical problems and requirements when implementing the Hyperledger Fabric blockchain in real-world implementations, including node recovery, blockchain checkpointing, and byzantine fault-tolerant consensus.Although the original public blockchains rely on a self-sovereign management style with complete decentralization, in the enterprise-based consortium blockchains the participating enterprises are often concerned about maintaining their own nodes with efficiency and resiliency, while at the same time, working as part of a cross-company blockchain network.This requires a secure and flexible governance model and on-chain collaboration mechanisms to address the many operational issues at different layers of the blockchain network – from interoperable connections, to storage management, membership governance, chaincode distribution, etc. We will share our experience and approaches to addressing these considerations and invite further technical discussions in the community towards practical solutions.
----
Hyperledger – Open Source Blockchain Technologies
Hyperledger is an open source community focused on developing a suite of stable frameworks, tools and libraries for enterprise-grade blockchain deployments.
It serves as a neutral home for various distributed ledger frameworks including Hyperledger Besu, Fabric, Sawtooth, Indy, as well as tools like Hyperledger Avalon, Cactus and libraries like Hyperledger Aries, Ursa. Learn more about Hyperledger projects: https://www.hyperledger.org/use
Case Studies: https://www.hyperledger.org/learn/case-studies
Training & Certification: https://www.hyperledger.org/learn/training
Tutorials: https://www.hyperledger.org/use/tutorials
Webinars: https://www.hyperledger.org/learn/webinars
Events: https://www.hyperledger.org/events
Vendor Directory: https://www.hyperledger.org/use/vendor-directory
Subscribe to the Hyperledger Newsletter: https://www.hyperledger.org/newsletter
Follow-us on Twitter @Hyperledger
Learn about Hyperledger Membership: https://www.hyperledger.org/about/join
#Hyperledger #HyperledgerGlobalForum2020 #Blockchain
A
Having
helped
will
close
the
door
I'd
behind.
Thank
you
so
I'm
baba
young
from
Oracle
and
today,
I
want
to
share
with
you
some
appearance
when
we
designed
our
record
ocean
platform
and
also
some
of
the
issues
that
come
from
our
customers
feedback,
and
we
also
will
discuss
the
potential
solutions.
Relatedly
issues.
A
So
this
is
the
agenda
for
today
it
looks
so
many
idioms
right.
However,
the
first
part
I
want
to
introduce
the
Oracle
latest
blow
chin
platform
and
all
the
following
10
ATMs,
agree
to
the
issue
I
want
to
cover
today
included
CQ,
supporting
small
contract
and
also
the
data
recover
data,
backups
recovered
and
also
the
later
checkpoint
and
the
pruning
and
the
PFT
governance
performance,
privacy,
protection,
internet
internet
work,
work
protocol
and
also
the
pluggable
critical
implementations
and
the
auditing
capabilities.
I
believe.
A
A
So
for
Oracle
auction
platform,
we
actually
have
two
editions:
the
clouds
when
is
released
into
2018,
and
also
we
have
the
Promised
One
yeah,
some
of
the
customers.
They
want
to
own
everything,
all
the
data
in
their
own
data
centers,
and
we
recently
named
as
wangled
leading
plush
and
technology
providers
and
quoted
from
when
go
our
customers.
We
are
when
we
are
when
go
the
best
platform
and
we
have
customers
in
several
areas.
A
A
This
is
the
one
architecture
out
our
cloud
auditioning
of
the
blockchain
platform
and
you
can
see
ticket
taking
advantage
of
Oracle
cloud
technologies,
including
our
identity,
management
and
event,
and
all
the
other
data
protection
at
the
DB
technologies.
We
build
on
how
about
that
and
we
adopted
open
source
have
a
later
fabric
kernel
and
suddenly
we
have
some
enhancements
on
these
domains
of
you
know.
For
example,
we
enhance
the
smoke
on
child
support.
We
provided
fine-grained
access
control
and
we
also
provided
some
assistant
to
also
related
to
the
Consensus
and
also
the
governance.
A
I
want
to
point
it
at
our
auction
platform.
It's
totally
compatible
with
the
open
source
standards
like
hybridized
fabric.
So
if
you
have
a
open
source
need
to
have
little
favorite
deployment,
you
can
feel
free
to
connect
with
each
other,
and
you
can,
if
you
have
a
promise
version,
you
can
connect
with
the
crowded
Edison
is
totally
compatible,
so
this
one
is
the
Enterprise
Edition
or
the
on-premise
version.
So
compared
with
the
cloud
released
with
these
on-premise
deployments,
you
can
control
everything
like
and
it
brings
all
the
clouds.
A
A
On
top
of
the
open
source
tablet
or
fabric,
we
certainly
add
some
new
features
for
the
blockchain
developers
and
also
the
notion
of
users.
It
includes
dudes
like
the
recipe
policy
it
provided
some
RESTful
API
to
allow
you
to
operate
and
send
the
translation,
the
tulip
logic
Network,
and
also
we
have
the
sequence
of
hold
into
a
small
contract,
and
it's
enabled
rich
history
query
and
is
easy
to
doing
some
analysis
and
how
about
ledger
works
later
data
and
in
the
past
two
years.
A
A
It
might
be
the
next
very
hot
area
to
adopt
approach,
and
you
know
due
to
the
coronavirus
this
time,
mates,
there's
a
increasing
the
amounts
to
distribute
and
also
to
provide
the
health
related
disciplines.
But
it's
very
low
efficient
is
not
efficient.
You
using
the
traditional
centralized
system
to
record
all
all
the
data,
so
Plaza
might
be
promising
in
this
area.
A
So
for
the
Oracle
side,
actually
we
focus
on
the
blockchain
platform
layer,
so
we
provided
technologies
and
we
work
together
with
our
application
team
and
our
SAS
team
and
at
the
bi
SI,
and
also
we
have
those
partners
who
work
with
us.
Together
we
provided
an
entire
from
the
top
to
the
bottom
so
loose
to
the
customers
and
the
ecosystem
actually
is
very
open.
A
So
it's
welcome
everyone
to
join
if
you
feel
interesting,
okay,
so
based
on
all
these
developments
and
also
the
feedback,
so
the
Yogi's
customers,
we
have
summarized
some
of
the
issues
that
we
need
to
address
when
you
use
blow
changing
REO
scenario.
So
the
first
one
is
thank
you
lots
of
our
customers
talking
about
this,
the
how
development
applications
using
Seco
for
the
legacies
reasons
and
currently
they
need
to
develop
to
the
smart
contract.
So
why
not?
We
support
that.
A
There's
some
support
in
years,
so
seeker
is
already
very
popular
and
you
can
see
developer
efforts
if
you're,
smart
control
logic
sees,
you
know
very
complete.
You
can
just
use
CQ
to
let
it
keep
the
transient
consistency.
Also,
there
are
some
against
opinions.
For
example,
security
is
not
good
in
performance.
Come
here
with
the
pure
key
value,
storage
and
another
Avenue
is
that
blockchain
contract
itself,
which
should
be
kept
a
very
simple
Euston
used
to
not
using
the
to
it
using
C
code
to
implement
those
very
complete
logics
inside
of
some
a
smart
country.
A
A
One
two
three
four:
five:
six,
seven,
okay,
fair
enough-
and
our
opinion
is
that
we
think,
is
a
good
option.
It's
a
it's
optional
to
our
to
your
customers,
because
we
have
some
other
reason
that
the
no
CQ
events
and
no
single
activities
actually
recently
is
becoming
is
adding
more
and
more
features
from
Security's
become
very
similar.
As
the
CQ
and
our
platform,
we
supposed
to
CQ
data
structure
actually
from
the
very
beginning,
and
we
used
the
open
source
Berkeley
database.
A
That's
the
to
let
the
customer
that
he
can
follow
those
open
source
and
open
standards.
They
do
not
need
to
bind
it
to
to
the
proprietary
product
like
the
Oracle
database
right-
and
we
are
our
experience
and
also
the
evaluation
we
founded.
The
performance
is
very
good
to
support
more
than
2,000.
Tps
is
good
enough
for
most
use
case
and
our
enterprise
customers
do
have
a
very
positive
feedback.
With
this
feature,
you
can
just
use
the
tradition,
CQ
grammar
inside
the
smog,
contract
and
everything
will
be
handled
by
the
secure
database
in
yourself.
A
The
customers
think
it's
useful
if
we
can
provide
it.
The
node
backup
and
recovery
capability
to
end
to
keep
the
know
the
resilience
from
those
data
crashing
and
especially
when
you
want
to
migrate
note
from
one
data
center
to
another
one.
Then
we
think
these
features
were
useful
and
also
there
are
some
against
the
pioneers
like
cloud
change.
The
self
is
already
distributed
is
a
distributed
season.
You
do
not
need
to
keep
every
pot
single
node
to
be
resilient
enough
right
and
also
the
production.
Node
is
very
cheap.
To
create
it's
easy
to
great.
A
A
So
in
theory,
we
totally
agree
that
single
no.
This
should
not
be
kept
as
a
resonance
yes
from
entire
network
system
point
of
view.
But
in
fact,
if
you
operate
the
bloody
network
by
yourselves
in
real
case,
you
will
found
that
the
time
of
sync
up
the
data
is
a
disaster.
So
it
depends
on
your
network
disk
CPU
performance
at
any
time
to
rebuild
all
the
TDP
stability
to
the
worst
it
and
also
the
history
database,
some
real
number
from
the
from
the
experiments
that
the
ledger
with
more
than
one
linear
transition.
A
A
So
if
we
want
to
back
up
what
in
turn
should
be
back
up,
we
have
an
L
will
have
analysis
or
different
know
the
rules
like
the
fabrics,
a
node.
You
need
to
take
care
of
the
credentials
database
and
also
the
configurations
and
for
the
other
node
you
need
to
take
care
of
the
certificates,
the
credentials
of
the
config
and
the
Rafah
data
and
also
those
blocks
you
can
you
make
hands
you
Missy
the
the
red
blocks.
All
these
red
blocks
need
to
be
handled
manually.
A
The
only
thing
you
do
not
need
to
handle
that
is
those
in
the
green
box.
That's
the
index
database
and
index
tips
and
were
stated
in
the
vault
appears.
However,
in
your
case,
it
takes
huge
our
time
to
rebuild
these
index
database
and
also
the
the
worst
state
database,
so
actually
from
protocol
purpose.
You
need
to
actually
handle
all
of
the
data
by
yourself.
So
that's
why
we
think
we
need
to
do
the
data
backup
and
recover
functionality.
A
Another
question
is
related
to
the
later
checkpoint,
so
letter
check
condom
means
you
can
let
all
the
members
of
some
channel
T.
However,
some
agreement
on
the
snapchat
snapshot,
then
we
some
will
join
the
new
channel
or
the
just
restarted.
He
can
from
the
steps
name:
checkpoints
not
from
the
very
beginning
so
data
current.
This
can
reduce
the
amount
of
time
and
also
this
can
see
the
storage
and
the
it
also
can
help
somehow
in
the
performance.
So
this-
and
we
also
have
some
against
the
opinions
that
the
storage
itself
is
cheap.
A
You
to
knowledge,
II
need
to
see
that
and
if
there's
no
for
later
there's
there's
all
in
snapshots,
the
ADA
will
require
more
operations
to
query
the
transactions.
So
it's
not
not
very
efficient.
So
how
many
guys
is
about
this
Avenue
whoa
more
than
ten
I,
can't
can't
so
how
many
against
only
one
okay?
A
A
However,
we
we
do
not
have
a
native
solution
for
these
issues.
We
have
some
walk
around
that
you
can
store
the
blockchain
files
externally.
You
do
not
need
to
put
it
inside
into
the
storage
system
without
the
peer
nodes
and
when
there's
some
credibility
to
the
block
kinda
transitions,
you
can
access
the
external
storage,
but
that's
illegal
work
around
some
ongoing
work
related
to
f
AP
106,
and
hopefully
they
can
be
done
out
this
year.
A
A
Okay,
almost
ten,
how
many
guys
against
two
this
feature:
one
two:
three:
four:
okay,
five.
Okay,
so
we
agree
that
this
is
not
very
common
requirement
related
to
enterprise
protein,
but
for
some
specific
scenarios,
like
some
specifics,
everything
inside
of
the
finance
area,
this
is
required.
So
there
is
not.
There
is
no
practical
solution
for
this
problem
yet,
but
I
believe
we
can
have
some.
We
can
take
advantage
of
some
reasons:
Damir
work,
including
the
first
one,
it's
the
al-quran,
maybe
no
it's
the
10
image.
Maybe
someone
guys
have
heard
of
that.
It's.
B
A
The
performance
is
good,
it
can
achieve
more
than
thousand
TPS
and
basically
it's
simplified
practical
editing
for
tolerance,
algorithm,
a
mathematician
and
eight
select
the
part
of
the
Rangers
inside
of
the
network
and
each
one
just
take
the
seconds
to
become
to
purpose
old,
to
propose
request
into
a
network
and
also
when
there's
some
will
change
happening.
Everyone
is
consist.
We
are
somewhat
you
empty
block.
A
Another
worker
is
the
a
Legrand
it's
where
new
work
and
proposal
by
MIT
lab
it's
it's.
The
motivation
is
to
enhance
and
to
increase
the
throughput
come
here
with
Bitcoin.
For
guess,
it
claims
to
be
30
times
faster
than
bill
:,
and
it
also
adopted
some
create
capital
adidas
track
functionality.
It
named
were
wherever
we
were
viable
random
function
and
to
select
all
in
parts
of
the
members
they
can
become
the
proposer
of
the
network,
and
it
requires
a
very
strong,
very
strong
synchronizing
network
condition.
A
A
And
the
first
one
is
the
hot
stuff
you
guys
may
have
heard
this
name
into
liberal
white
paper.
It's
where
a
new
work.
It's
a
change
that
actually
I
think
this
work
changed
the
consensus
model,
it
increased
change,
82
from
two
faiths
consensus:
confirmation
to
a
three
phase;
confirmation,
and
also
it.
A
Rely
on
some
central
north
central,
some
leader
node,
to
reduce
the
Commission
cost
it
is,
it
is
thought
very
promising
in
improved
performance
out,
Carissa's,
okay,
those
are
the
academia
works.
So
how
about
today's
hot
leather
fabric?
How
about
the
BFG
work
there?
Actually,
since
the
wedding
earlier
version
of
0.26,
we
already
have
some
PFT
vanishing
in
fabric,
but
it's
very
buggy
and
the
performance
and
scalability
not
good
at
the
UN
Punto
and
at
the
1.03
version.
There's
a
no
official
support
or
PFT
in
Hubli
refer
brain,
but
we
have
some
third-party
limitation
of
that.
A
The
how
Java,
based
plugging
to
work
with
fabric
adhere,
and
but
all
these
versions
are
have
some
loss
of
real
issues,
for
example,
are
related
to
the
conflict.
They
do
not
support
the
channel
configuration
there
and
there
some
also
some
ongoing
work
to
re-implement
these
PFT
smarts
library
using
Gulen
I,
wanted
to
and
wanted
to
integrate
into
the
have
Allegra
Fiber
project.
Maybe
this
year,
maybe
this
year.
A
C
A
A
Okay,
almost
ten?
How
many
guys
against
that?
Okay,
everyone
needs
this
feature.
Yes,
we
agree
that
we
think
it
is
very
important
to
bring
this
governing
functionality
into
the
enterprise
poaching.
You
know
when
you
negotiate
includes
operations
cause
organized
organization.
It
is
very,
very,
is
very,
very
low
if
it's
in
theory
and
we're
being
commended
in
practice-
and
there
are
several
potential
solutions
there.
First,
the
wire
is-
we
have
some
special
governing
channel.
A
You
know
in
today's
hub
later
fabrica
we
already
have
a
system
channel,
it's
used
to
governing
those
other
in
service,
so
we
can
extend
that
or
we
can
have
another
specific
channel
to
governing
our
other
operations
course.
Regions
course
organizations
and
the
second
way
is
we
can
implement
this
functionality
using
a
new
system
chain
code.
I
will
call
that
the
governing
system
change
SEC
and
the
third
one.
A
If
you
do
not
want
to
change
any
of
the
fabric
code,
you
do
not
want
to
hack
that
then
you
can
use
the
some
external
dinner
service
like
a
gateway
to
doing
these
governing
operations.
I
will
see
the
second
option
using
some
system
Cinque
it's
very
promising
here,
because
in
the
new
release
the
fabric
2.0,
you
may
see
the
the
lifecycle
of
the
chain
code
has
already
been
routed
using
some
livestock
whole
system
chain
code.
So
we
we
can
watch
for
the
lifecycle
of
the
chain
code.
A
Suddenly
we
can
extend
that
to
with
wood
for
other
things
like
who
can
join
the
channel
and
which
other
service
can
be
used,
and
also
we
have
a
fabricant
of
team.
We
have
done
some
experiments
to
test
the
interoperability
between
Oracle,
IBM
and
other
vendors
blockchain
platforms,
and
we
used
a
similar
system
chain
code
to
to
tween
to
two
Toads
governing
our
functionalities.
So
we
think
this
way
is
very
promising
here.
A
Ok,
the
next
one
is
performance,
lots
of
guys
talking
about
Palmas.
So
do
you
guys
think
the
customers
really
care
about
performance
and
there's
some
supposin
a
support
of
ten
years?
That,
because
everyone
is
talking
about
set
performance
and
and
also
from
the
academia,
we
can
see
lots
of
paper.
The
published
numbers
L
from
K
200
K
through
put
out
the
distributed
systems,
but
there
is
also
some
against
the
opinions
that,
when
talking
about
the
performance,
most
of
the
cosmati
do
not
really
understand
what
foremost
they
are
talking
about.
A
I'm
talking
about
the
report,
the
latency,
the
scalability
you're
talking
about
the
configuration
I
talking
about
applique
application
layer
or
the
system
layer
and
for
real
scenarios,
I
will
see
for
most
of
the
cases
less
than
1000
through
TPS
is
good
enough.
We
only
see
a
few
few
real
scenarios.
They
require
more
than
1000
GPS.
A
A
And
when
your
customer
think
they
need
you
to
do
some
moments
of
meditation
work
for
them
and
I
think
the
first
issue
you
need
to
figure
out
is
what
performance
you
are
talking
about,
and
the
next
question
is:
what
is
the
goal
for
the
next
five
years?
You
know
today
due
to
the
rapid
developments
approaching
technology,
and
you
can.
You
cannot
actually
design
system
that
can
work
perfectly
in
more
than
five
years
and
in
terms
of
optimizing,
the
capillary
fabric
based
plow
chain
platform.
Actually,
so
it's
a!
Where
is
this
magic
work?
You
can
see.
A
We
have
the
chain
code,
we
have
the
ledger.
We
have
the
consistence,
we
have
a
security
and
the
require
different
missions-
resources,
for
example.
The
chain
code
typically
requires
it
needs
a
lot
so
calculation,
so
it's
compute,
compute,
sensitive
and
the
ledger
will
require
or
faster
and
very
large
volume
of
storage
and
the
consistent
is
where,
since
the
network,
I/o
latency
right,
it's
differently
things
here.
A
Another
issue
I
want
to
mention
here
is
when
people
are
talking
about
the
performance.
They
only
care
about
some
member
for
very
small
skill
or
very
specific
topology,
but
they
ignore
Manning
real
issues.
For
example,
you
may
notice
that
the
with
more
and
more
ledger
data
accumulated
the
performance
when
you
run
chenko,
the
same
transaction
will
degrade
so
why
the
problem
is
that
for
today's
blockchain
in
order
to
resolve
those
turbulent
problem,
each
transaction
will
be
checked
with
this
high
D,
whether
it
happened
already
inside
the
database.
A
We
need
to
keep
privacy
protection,
however,
when
people
when
they
want
to
use
blockchain
the
actually
it's
you
want
to
share
data
among
multi
parties
right
different
members
and
another
option
is
that
not
other
opinion
is
that
you
should
not
put
those
sensitive
data
on
to
the
blockchain
and
all
existing
solutions
based
on
cryptography.
They
have
some
trade-offs
in
practice,
so
how
many
guys
in
sport
is
feature
okay,
more
than
ten,
how
many
guys
against?
A
A
A
In
the
2.0
release
there
are
some
new
features.
It
called
the
implements
implicit
collection.
I
will
see
it's
very
promising
and
I
see
the
fundaments
for
furthers
flexible
private
collection
design.
You
can
pay
attention
to
this
feature,
however,
for
all
these
private
collections,
there's
a
steel
hashed
recall
the
inside
of
the
public
chains,
so
I
remember
inside
of
the
channel.
A
They
all
see
these
hashed
without
so
my
suggestion
is
that
you,
if
you
clear
about
the
dealer
priority
seriously,
please
just
encrypt
the
data
before
you
put
it
onto
the
chain,
or
at
least
that
you
can
add
some
salt
into
the
data
increased
to
to
avoid
other
members.
Get
statistical
results
from
the
hostel
result.
A
However,
there
is
some
also
against
veneer,
like
today's
enterprise.
Consume
is
typically,
is
not
very
large
right,
most
of
them
less
than
30
members
so
and
that
they
are
focusing
on
the
same
business
narrow.
So
why
not
they
just
using
the
same
technology.
So
how
many
of
you
are
support?
This
ok,
almost
20,
how
many
against
who
against
ok.
A
So
I
will
see.
We
totally
agree
that
this
is
very
critical
requirements
and
when
you
think
about
that,
you
need
to
think
that
what
that's
internetwork
reversion
mean.
So
it's
all
in
support
of
real
data
or
update
data,
or
just
you
know,
across
different
approach
in
toner.
Did
you
just
choose
some
consensus?
They
are
totally
I
will
see
different
requirements.
A
Today
we
have
seen
many
platforms
build
on
top
of
hubba
later,
especially
the
fabric
and
also
tears
easier
remedies
when
the
quorum
is
one
and
there
are
some
standards
for
our
course
network
or
intro
later,
but
so
there's
still
no
easy
solution.
I
will
see
most
of
those
practical
solutions.
They
require
some
extra
effort
likely
to
introduce
data
gateway
and
which
will
become
centralized
somehow.
A
Okay,
the
next
one
is
the
flexible,
crypto
implementation,
so
you
may
know
today's
tablet
or
fabric
it
using
the
Golan
official
crypto
library
to
implemented
international
standards.
So
it's
not
difficult.
It's
not
easy
to
support
other
capital
standards
so
especially
in
the
Asia
area,
and
we
have
seen
that
China,
Russia
and
particular
youth.
They
have
some
policy
requirements
to
align
with
the
original
capital
standards.
A
A
A
Is
that
all
you
think
blockchain
dissolve
it
does
not
need,
does
not
need
auditing,
because
it
is
already
there
because
bludgeon
ledger
recalls
all
the
history
of
the
data,
all
the
history
of
the
transactions,
and
so
another
opinion
is
that
the
auditing
can
be
implemented
outside
low
changes
and
also
block
change
works
to
the
auditing
to
serve
all
the
auditing
and
especially
with
the
fabric,
hard
leather
fabric.
We
have
the
original
API
aggressive
API,
so
we
can
get
a
lot
of
studies
to
get
data
from
it.
C
A
You
I
will
see.
Auditing
itself
is
highly
demanded,
special
industry
like
finance
and
the
auditing
functionality.
Its
is
different
from
logging
from
different
from
of
recent
statistics
and
also
different
from
so
if
we
can
have
some
auditing
API
from
the
hub
Allegra
fabric
code
base.
That
will
see
well
also
time
or
currently,
people
can
still
used
operational,
metrics
API,
and
we
are
the
logging
analysis.
They
can't
get
part
of
the
ordinary
that
they
want,
and
this
already
is
checked
with
the
very
earlier
fabric
issue.
A
A
A
C
A
D
D
A
D
No
wait
a
second.
We
can
have
other
algorithms
that
implement
consensus
differently,
right
that
provide
protection,
randomize
voting
and
all
kinds
of
stuff,
but
I'd
argue.
The
same
thing
is
true
of
EFT:
it's
just
a
strategy
to
solve
a
problem,
and
if
you
look
at
the
problem,
we
should
look
at
other
strategies.
B
A
Well,
it's
is
I,
will
see,
cities
multi-layer
ponds,
you
can
talk
about
strategy
from
business
and
the
technologies
of
coast
right
and
there's
some
so
I.
Don't
think
it's
a
technical
issue
here,
but
the
slide
the
topic
I
Carol
here
compared
with
EFT
and
safety,
is
just
from
the
technical
side,
so
algorithm
yeah.
But
there
that's
a
good
point.