<s>
In	O
data	B-General_Concept
storage	I-General_Concept
,	O
disk	B-General_Concept
mirroring	I-General_Concept
is	O
the	O
replication	B-Architecture
of	O
logical	B-General_Concept
disk	I-General_Concept
volumes	O
onto	O
separate	O
physical	O
hard	B-Device
disks	I-Device
in	O
real	B-General_Concept
time	I-General_Concept
to	O
ensure	O
continuous	B-General_Concept
availability	I-General_Concept
.	O
</s>
<s>
A	O
mirrored	B-General_Concept
volume	I-General_Concept
is	O
a	O
complete	O
logical	O
representation	O
of	O
separate	O
volume	O
copies	O
.	O
</s>
<s>
In	O
a	O
disaster	B-General_Concept
recovery	I-General_Concept
context	O
,	O
mirroring	O
data	O
over	O
long	O
distance	O
is	O
referred	O
to	O
as	O
storage	O
replication	B-Architecture
.	O
</s>
<s>
Depending	O
on	O
the	O
technologies	O
used	O
,	O
replication	B-Architecture
can	O
be	O
performed	O
synchronously	O
,	O
asynchronously	O
,	O
semi-synchronously	O
,	O
or	O
point-in-time	O
.	O
</s>
<s>
Replication	B-Architecture
is	O
enabled	O
via	O
microcode	B-Device
on	O
the	O
disk	B-Device
array	I-Device
controller	I-Device
or	O
via	O
server	B-Application
software	I-Application
.	O
</s>
<s>
It	O
is	O
typically	O
a	O
proprietary	O
solution	O
,	O
not	O
compatible	O
between	O
various	O
data	B-General_Concept
storage	I-General_Concept
device	I-General_Concept
vendors	O
.	O
</s>
<s>
Asynchronous	O
replication	B-Architecture
can	O
achieve	O
an	O
RPO	O
of	O
just	O
a	O
few	O
seconds	O
while	O
the	O
remaining	O
methodologies	O
provide	O
an	O
RPO	O
of	O
a	O
few	O
minutes	O
to	O
perhaps	O
several	O
hours	O
.	O
</s>
<s>
Disk	B-General_Concept
mirroring	I-General_Concept
differs	O
from	O
file	B-Application
shadowing	I-Application
that	O
operates	O
on	O
the	O
file	O
level	O
,	O
and	O
disk	B-Application
snapshots	I-Application
where	O
data	O
images	O
are	O
never	O
re-synced	O
with	O
their	O
origins	O
.	O
</s>
<s>
Typically	O
,	O
mirroring	O
is	O
provided	O
in	O
either	O
hardware	O
solutions	O
such	O
as	O
disk	B-General_Concept
arrays	I-General_Concept
,	O
or	O
in	O
software	O
within	O
the	O
operating	O
system	O
(	O
such	O
as	O
Linux	O
mdadm	B-Application
and	O
device	O
mapper	O
)	O
.	O
</s>
<s>
Additionally	O
,	O
file	B-Application
systems	I-Application
like	O
Btrfs	B-Operating_System
or	O
ZFS	B-Application
provide	O
integrated	O
data	B-General_Concept
mirroring	I-General_Concept
.	O
</s>
<s>
There	O
are	O
additional	O
benefits	O
from	O
Btrfs	B-Operating_System
and	O
ZFS	B-Application
,	O
which	O
maintain	O
both	O
data	O
and	O
metadata	O
integrity	O
checksums	O
,	O
making	O
themselves	O
capable	O
of	O
detecting	O
bad	O
copies	O
of	O
blocks	O
,	O
and	O
using	O
mirrored	O
data	O
to	O
pull	O
up	O
data	O
from	O
correct	O
blocks	O
.	O
</s>
<s>
In	O
a	O
hot	B-Device
swap	I-Device
system	O
,	O
in	O
the	O
event	O
of	O
a	O
disk	O
failure	O
,	O
the	O
system	O
itself	O
typically	O
diagnoses	O
a	O
disk	O
failure	O
and	O
signals	O
a	O
failure	O
.	O
</s>
<s>
Sophisticated	O
systems	O
may	O
automatically	O
activate	O
a	O
hot	B-General_Concept
standby	I-General_Concept
disk	O
and	O
use	O
the	O
remaining	O
active	O
disk	O
to	O
copy	O
live	O
data	O
onto	O
this	O
disk	O
.	O
</s>
<s>
For	O
remote	O
disaster	B-General_Concept
recovery	I-General_Concept
systems	O
,	O
this	O
mirroring	O
may	O
not	O
be	O
done	O
by	O
integrated	O
systems	O
but	O
simply	O
by	O
additional	O
applications	O
on	O
primary	O
and	O
secondary	O
machines	O
.	O
</s>
<s>
In	O
addition	O
to	O
providing	O
an	O
additional	O
copy	O
of	O
the	O
data	O
for	O
the	O
purpose	O
of	O
redundancy	O
in	O
case	O
of	O
hardware	O
failure	O
,	O
disk	B-General_Concept
mirroring	I-General_Concept
can	O
allow	O
each	O
disk	O
to	O
be	O
accessed	O
separately	O
for	O
reading	O
purposes	O
.	O
</s>
