►
From YouTube: CNCF SIG Storage – An Intro to Cloud Native Storage - Alex Chircop, StorageOS & Erin Boyd, Red Hat
Description
Don’t miss out! Join us at our upcoming events: EnvoyCon Virtual on October 15 and KubeCon + CloudNativeCon North America 2020 Virtual from November 17-20. Learn more at https://kubecon.io. The conferences feature presentations from developers and end users of Kubernetes, Prometheus, Envoy, and all of the other CNCF-hosted projects.
CNCF SIG Storage – An Intro to Cloud Native Storage - Alex Chircop, StorageOS & Erin Boyd, Red Hat
The CNCF SIG Storage (https://github.com/cncf/sig-storage) collaborates to explore and understand how different storage technologies are used in cloud-native environments. Topics include block stores, file systems, object stores, key-value stores and databases, amongst others. Different architectural approaches (centralized, distributed, sharded etc) are compared in terms of key attributes like availability, scalability, performance, data consistency, durability, fault tolerance, ease of development and operational complexity. In the intro, we'll provide an overview of the work currently underway and on the roadmap. You'll get the meet the people leading these efforts, and find out how best to get involved and contribute. We'll present our work thus far, and host discussions around the findings of how cloud native storage is approached in the real world, including some representative successes and failures. Our hope is that others can learn from, and expand upon those experiences.
https://sched.co/Zevi
A
B
Is
aaron
boyd?
I
work
for
red
hat,
as
mentioned,
I'm
also
the
co-chair
for
the
cncf
storage,
sig
and
sorry
we're
not
able
to
join
everyone
in
person,
but
we're
glad
that
we
can
provide
this
introductory
session.
So
you
can
learn
more
about
how
to
get
involved
or
how
to
donate
your
project
to
the
cncf.
B
And
so
with
that
we're
just
going
to
go
over,
why
do
we
have
sigs
and
the
reason
why
sigs
were
were
generally
created
was
so
that
we
could
scale
contributions
to
the
cloud
native
computing
foundation
where
we
have
both
the
technical
and
user
community
and
retaining
the
integrity
and
quality
of
the
ones
being
committed
to
the
cncf
and
their
mission?
This
allowed
more
people
to
be
able
to
provide
a
due
diligence
and
interface
to
many
members
of
the
community.
B
So
what
exactly
are
the
objectives
of
the
sig?
It's
meant
to
be
created
so
that
we
can
strengthen
the
project
ecosystem.
This
means
that
we
can
find
projects
that
aren't
currently
well
represented
within
the
cncf
and
foster
those
into
the
environment
by
identifying
those
gaps
in
the
portfolio
and
attracting
projects
to
fill
those
gaps.
It's
also
meant
to
educate
and
inform
users
what
different
projects
do
where
the
gaps
are
and
where
we
can
grow
the
different
technologies.
We're
meant
to
be
unbiased,
effective
and
provide
practical
information
to
our
user
community.
B
B
Process
the
cncf
sig
also
helps
us
clarify
the
relationship
between
projects,
because
there
are
many
different
sigs
within
the
cncf.
This
allows
us
to
cross-communicate
between
different
pieces
of
the
technology,
where
we
may
have
a
very
storage
focused
project.
It
also
has
aspects
of
perhaps
security,
so
this
allows
us
to
talk
with
other
experts
within
that
field
and
communicate
and
understand
how
the
project
fits
across
all
of
these
different
pillars.
It
also
allows
us
to
engage
more
communities
into
on-ramp,
so
we
have
an
effective
communication
to
the
technical
oversight
committee.
B
B
Neutral,
so
with
our
responsibilities,
we
start
off
with
handling
the
project
by
understanding
what
the
project
does
by
documenting
a
high-level
roadmap
within
the
space
and
identifying
the
gaps
within
the
project
landscape.
Again,
the
project
landscape
can
be
found
on
the
cncf
website.
The
projects
that
fall
within
that
we
then
go
through
and
perform
health
checks.
Is
it
a
vibrant
project?
Is
it
growing?
Does
it
have
a
wide
base
of
community
or
is
it
a
new
project?
That's
looking
to
find
new
members.
These
are
all
questions
that
we
ask
ourselves.
B
B
B
This
presentation
is
then
recorded
so
that
it
can
be
passed
on
to
the
toc
and
they
can
watch
it
when
they
have
time
for
it.
This
allows
the
toc
to
scale
much
more
effectively
rather
than
having
rigid
meetings.
The
toc
then
meets
and
reviews
a
recommendation
that
the
sig
creates
and
then
decides
whether
or
not
the
project
is
appropriate
and
goes
forward
through
incubation
sponsorship,
and
with
that
the
a
toc
member
is
then
assigned,
along
with
the
sig
to
do
a
due
diligence.
B
The
due
diligence
consists
of
looking
at
the
technology,
the
governance
and
talking
to
different
people
in
the
community
to
understand
more
about
projects
once
the
due
diligence
is
done,
then
the
project
goes
up
for
public
comment
and
then
the
toc
vote.
If
the
vote
is
confirmed,
the
project
is
accepted
into
incubation.
B
Another
one
of
our
responsibilities
within
the
sig
is
end
user
education.
We
aim
to
provide
up-to-date,
high
quality
and
unbiased
easy
to
understand
material.
This,
in
that,
how
do
we
choose
to
use
a
technology
within
our
workflow
and
effectively
adopt
these
different
cloud-native
technologies
when
it
comes
to
storage?
What
are
the
best
practices?
B
Some
of
the
things
that
we've
created
to
help
users
facilitate
doing
this?
Are
white
papers,
presentations,
videos
and
other
forms
of
training
to
talk
about
to
clarify
the
technologies
to
compare
different
approaches
again,
since
we're
vendor
neutral,
we
tend
not
to
focus
on
one
or
the
other,
but
what
is
the
best
way
to
go
about
it,
given
your
particular
application
or
workload?
B
Technologies
and
lastly,
we
look
to
the
community
to
understand
what
their
needs
are.
Part
of
this
is
gathering
input
from
them.
What
are
their
expectations
when
they're
looking
to
the
cncf
storage
sig?
What
information
do
they
expect
to
learn
and
glean
from
us?
What
are
the
pain
points
that
they're
having
when
they're
integrating
persistent
storage
into
their
environment?
B
B
B
And
lastly,
the
sig
responsibilities
are
meant
to
be
the
trusted
advisors
to
the
toc,
since
the
toc
is
a
very
limited
number
of
people.
Sometimes
different
technologies
are
maybe
not
fully
represented
by
the
toc,
so
this
is
meant
to
expand
on
their
experience
as
an
sme
were
to
assist
in
performing
the
technical
due
diligence.
Given
that
we're
subject
matter
experts
in
those
areas
on
both
new
and
graduating
projects
and
advise
the
toc
on
our
findings
and
be
involved
or
periodically
check
in
on
those
projects.
A
Thanks
erin,
so
a
little
bit
now
about
some
of
the
work
that
we're
that
we've
been
doing
in
sig
storage,
so
six
storage
was
was
founded
in
kubecon
back
in
barcelona,
and
it
was
one
of
the
first
things
that
was
put
together.
The
reason.
The
reason
why
we
think
you
know
storage
is
is
important
to
consider
when
we're
looking
at
the
the
overall
cloud
native
landscape
is
because
you
know,
contrary
to
popular
belief.
There
is
no
such
thing
as
a
as
a
stateless
architecture.
A
A
So
it's
more
than
ever
before
it's
important
to
understand
the
different
attributes
of
a
storage
system
like
availability
and
scalability
performance,
durability,
consistency
and
we've
tried
to
to
to
cover
that
detail
and
to
give
end
users
the
ability
to
to
understand
the
requirements
and
to
therefore
choose
the
appropriate
storage
systems
for
their
for
their
use.
Cases
in
the
storage
landscape
white
paper,
which
which
we
recently
published
the
version
two
of
this.
A
So
this
has
been
an
ongoing
and
ongoing
project
that
that
we've
that
we've
been
keeping
up
to
date,
as
as
things
advance,
the
sig
itself
is
made
up
of
a
number
of
coaches
and
a
number
of
tech
leads
and
is,
and
and
the
toc
liaison
for
the
sick
is
who's.
A
top
member
and
he's
also
a
contributor
to
the
to
the
and
and
author
of
the
xcd
project.
A
So
we've
got,
we've
got
a
healthy
mix
of
different
sets
of
expertise
and
we're
always
looking
for
additional
tech
leads.
So
if
you
think
that
that
you
can
help
with
the
sig
and
work
with
projects
and
evaluating
or
contributing
to
the
documents
that
we're
providing,
please
let
yourself
be
known,
we
always
welcome
new
tech
leads.
A
Some
of
the
specific
groups
that
we
also
interact
with
to
to
which
have
a
strong
storage
influence
are
the
obviously
the
kubernetes
storage
sake.
In
fact,
a
number
of
the
tech
leads
and
coaches
are
also
members
of
the
of
the
storage,
kubernetes
storage
sig.
Similarly,
the
the
csi
project,
which
is
which
is
an
interface
that
allows
kubernetes
to
to
interface
in
a
standard
way,
with
a
number
of
different
storage
providers.
A
Again
we
have
a
number
of
of
tech
leads,
who
are
also
on
the
csi
project.
The
security
sig,
as
as
as
aaron
mentioned
is,
is
I
think
that
we
work
with
for
security
issues
related
to
storage
projects.
Similarly,
the
app
sigs
and
the
service
catalog
sigs
a
little
bit
of
some
of
the
stuff
we've
been
working
on
since
the
last
cube
gone.
A
So
we've
been
helping
the
toc
review
a
number
of
new
or
existing
projects,
the
the
reviews
aren't
only
for
for
new
projects.
Sometimes
we
we
have
we're
reviewing
projects
that
are
going
through
a
graduation
process,
so
moving
from,
say
a
sandbox
to
incubation
or
or
an
incubation
stage
to
to
graduation.
A
We've
also
created
some
surveys
to
help
with
prioritizing
prioritizing
the
work
that
we're
doing
and
we've
been
updating
the
the
the
the
documents
that
we've
published
in
the
past.
So,
for
example,
we've
recently
published
the
version
two
of
the
storage
landscape
white
paper,
which
has
new
database
sections
and
updated
management
interfaces,
focus
focusing
on
csi
we're
we're
working
on
a
number
of
documents
for
for
community
use
cases
and
we've
also
drafted
a
white
paper
focusing
on
performance
and
benchmarking
within
storage.
A
There
there's
been
a
wide
variety
of
doc
of
projects
that
have
presented
to
the
sig.
The
presentations
are
really
useful
because
they're
they're
they're
they're
helpful
to
understand
the
project.
It
gets
better
community
adoption
and
and
widens
the
the
the
scope
of
the
project
and
obviously,
if
the
project
is
going
through
a
cncf
review,
those
recordings
and
and
the
and
the
q
a
process
that
that
typically
follows.
A
I'd
like
to
I'd
like
to
focus
just
a
little
bit
on
the
storage
landscape
white
paper.
This.
This
covers
a
number
of
different
things.
As
we
mentioned,
you
know
we
want
to
give
end
users
and
developers
the
ability
to
to
understand
the
attributes
of
a
storage
system
and
also
to
understand
the
different
layers
within
a
storage
system
and
how
they
impact
the
the
different
attributes
and
how
they
work
and
then,
finally,
we're
also
trying
to
focus.
A
The
reason
why
it's
important
to
to
to
understand
the
storage
attributes
is
because
you
they
help
to
define
what's
needed
out
of
the
storage
environment
based
on
the
attributes
that
are
required
by
the
application.
So
so
we've
we've
classified
five
different
attributes,
availability,
scalability
performance,
consistency
and
durability,
and
those
are
actually
measured
across
a
number
of
a
number
of
different
metrics.
So,
for
example,
availability
can
can
cover
things
from
the
ability
of
a
storage
system
to
fail
over
or
to
move
data
across
nodes,
as
well
as
data
protection
and
redundancy
capabilities.
A
Again,
scalability
can
be
measured
through
a
number
of
different
options,
often
through
the
number
of
operations
or
the
number
of
throughputs,
but
but
sometimes
also
the
number
of
clients
that
they
support
and
similarly
performance
can
be,
can
be
measured
with
a
number
of
different
parameters
such
as
latency
operations
or
throughput.
One
thing
to
consider
here
is
that
that
often
there
are
compromises
to
be
made
and
it's
not
possible
to
hit
high
values
of
all
of
these
metrics.
A
Data
to
to
a
non-volatile
layer
and,
what's
particularly
important,
is,
is
to
understand
that
the
way
that
you
access,
the
storage
from
your
operating
system
from
your
application,
doesn't
necessarily
define
anymore
the
types
of
attributes
of
that
storage
system.
So
so
you
know
we
we
sometimes
think
of
you
know.
Broad
generalizations,
like
file
systems
are
excellent
for
throughput
or
block-based
devices
might
be
excellent
for
low
latency,
but
in
reality
is
we.
A
The
key
thing
here
is
to
give
people
the
tools
to
to
make
their
own
comparisons.
It's
very
hard.
In
fact,
it's
almost
pointless
to
try
and
compare
published
results
from
different
vendors
or
different
projects
without
a
deep
understanding
of
of
how
the
test
is
run
and
how
it
applies
to
your
application.
So
long
story
short,
always
try
and
run
your
own
tests,
and
the
good
news
is
that,
with
you
know,
systems
like
kubernetes,
it's
it's
possible
to
to
make
self-service
declarative
tests
fairly
straightforward
too.
A
A
The
zoom
calls
are
open
the
the
repo
containing
the
work
that
we
do
and
the
meeting
minutes
are
also
open
and
every
meeting
is
recorded,
which
is
particularly
useful
to
to
to
go
back
and
and
be
able
to
watch
project
presentations.
For
example,
if
you
want
to
to
join
the
sig,
the
easiest
thing
to
do
is
to
join
the
mailing
list
and
and
hop
onto
the
on
to
the
next
call.
We
would
be
happy
to
welcome
you
specifically,
as
I
mentioned
before,
we
look.
A
We're
always
looking
for
for
tech,
leads
to
help
work
on
project
reviews
and
and
and
whether
they're
new
projects
or
or
reviewing
the
health
of
existing
projects
and
we'll
also
like
help
in
documents
like
the
performance
white
paper.
So
again,
please
do
join
and
we'd
love
to
have
you.
The
the
sig
is,
is
made
up
of
over
30
representatives
from
different
vendors
and
as
well
as
a
number
of
independent
developers
and
community
members,
and
that
takes
us
to
the
end.
A
A
B
A
Of
course,
the
the
projects
that
that
we
evaluate
and
that
we're
that
we're
looking
at
are
covered
in
our
six
sessions,
so
please
feel
free
to
to
either
join
those
sessions
and
contribute
and
ask
questions
and
participate
in
the
projects
or
all
of
the
sessions
are
also
recorded.
So
it's
it's.
It's
a
good
idea
to
to
have
a
look
at
those
recordings
to
see
perhaps
past
project
presentations.
B
A
B
Yep,
let's
go
ahead
and
take
the
next
one,
so
alvaro
says:
is
the
sig
storage
only
interested
or
concerned
about
cncf
projects
or
its
activity
may
also
extend
to
other
open
source
projects
related
to
sick
interests?
So
I
think
the
sig
is
open
to
discussing
projects
that
haven't
yet
been
accepted.
I
mean
that's,
it's
part
of
our
process
is
understanding
the
ecosystem
projects
that
aren't
currently
in
the
cncf,
but
maybe
want
to
join
and
part
of
that
would
be
discussing
any
storage,
related
interests,
use
cases.
That's
certainly
part
of
it.
B
A
Yeah
definitely,
I
think,
I
think
the
project
presentations
are
some
of
the
most
informative
and
useful
sessions
on
on
the
sig,
and
one
of
the
things
that
we're
always
on
the
lookout
for
is
is
to
identify
potential
gaps
in
the
landscape.
Of
course,
the
the
cncf
projects
covering
storage
are
are
quite
broad.
We
have,
we
have
sections
covering,
you,
know
everything
from
distributed
file,
systems
to
block
stores
and
and
and
also
things
like
key
value
stores.
A
B
And
the
next
one
from
owen
says
I'm
particularly
interested
in
security
topics
in
relation
to
persistent
storage,
particularly
coming
from
a
traditional
non-cloud
native
background
into
the
world
of
k-8s.
I
think
both
alex-
and
I
touched
on
this
in
that
you
know,
though
the
sigs
are
focused
on
one
particular
area.
Far
as
being
storage
and
there's
a
separate
sig
for
security,
we
certainly
don't
operate
in
a
vacuum.
B
I
think
there
are
a
lot
of
security
concerns
that
we
look
at
when
we
talk
about
each
project
and
is
discussed
also
within
you
know
some
use
cases
and
white
papers,
so
we'd
love
to
hear
some
of
your
thoughts.
If
you
are
working
on
a
project
that
is
looking
through
this
or
have
any
questions
I
know
alex.
If
you
want
to
add
anything
to
that
piece.
A
Yeah
I
mean
when,
when
projects
go
through
graduation,
specifically
that
that
involves
a
security
audit,
so
that's
part
of
the
process.
That
said,
it's
probably
worth
noting
that
you
know
a
lot
of
storage
providers
that
that
work
with
the
csi
standards
also
have
integrations
to
the
to
the
kubernetes
security
model
and
and
volumes
tend
to
support
things
like
you
know,
name,
spaces
and
and
our
rack
etc.
So
so
you
you
may
find
that
that
the
security
requirements
are
are
catered
for
by
virtue
of
the
csi
integration.
B
We
felt
that
bias
crept
in
selecting
one
over
the
other,
so
we've
decided
to
kind
of
table
that
specific
approach
and
try
to
come
up
with
something
that
will
allow
users
to
make
smart
decisions,
and
maybe
that's
going
with
the
anti-patterns
of
what
not
to
do
and
where
not
to
use
things,
but
we're
certainly
open
to
suggestions
to
what
would
best
fit
the
user
community.
But
you
know
I,
I
guess
I'd
love
to
hear
feedback
of
what
people
are
looking
for
from
the
sig.
A
I
guess
the
different
attributes
of
their
application
and
how
that
translated
to
the
different
storage
systems-
and
you
know
there
are
therefore
some
examples
in
there
of
you-
know
different
different
characteristics
of
different
storage
systems
based
on
you
know
things
like
latency
or
consistency
or
or
things
like
that.
That
said,
you
know,
as
aaron
said,
we
we
have
a
fairly
strict
rule
about
not
being
king
making
within
within
the
sig
and
therefore
providing
more
detailed
use.
B
Okay,
the
next
question
is
around:
did
you
work
with
vmfs
plug-in
to
k-8s
there's
some
problems
with
the
vms
when
they're
shut
down
that
the
vm
stay
locked
until
it
started?
Do
you
know
about
this
problem?
Is
there
a
road
map
to
fix
it?
I
can't
speak
in
particular
to
this
particular
bug
and
we
don't
necessarily
look
at
k8s
at
that
level
within
the
sick.
A
B
Yep,
I
was
going
gonna
say
the
exact
same
thing,
so
they
typically
meet
bi-monthly
on
thursday.
You
can
also
find,
if
you
go
to
the
google
groups
for
kubernetes
you'll,
be
able
to
find
that
on
that
particular
calendar
or
also
send
it
to
their
mailing
list.
They
might
be
able
to
provide
you
better
status
around
that
and
if
there
isn't
a
bug,
please
open
an
issue
around
this.
A
Indeed,
so
on
that,
on
that
note,
we're
we're
we're
at
time,
so
I'll
I'll,
just
quickly
wrap
this
up
by
saying
we're
we're
always
super
keen
to
have
new
members
trying
to
take.
So
please,
if
you
do
consider
joining
one
of
the
signals
and
maybe
volunteering
to
to
help
with
the
documents
that
we
provide
or
or
to
to
to
volunteer,
to
become
a
tech
lead
and
and
help
with
project
reviews.
We're
we're
always
super
keen
on
having
new
members
join.