<s>
The	O
Chandra	B-Operating_System
–	I-Operating_System
Toueg	I-Operating_System
consensus	I-Operating_System
algorithm	I-Operating_System
,	O
published	O
by	O
Tushar	O
Deepak	O
Chandra	O
and	O
Sam	O
Toueg	O
in	O
1996	O
,	O
is	O
an	O
algorithm	O
for	O
solving	O
consensus	B-Operating_System
in	O
a	O
network	O
of	O
unreliable	O
processes	O
equipped	O
with	O
an	O
eventually	O
strong	O
failure	B-Operating_System
detector	I-Operating_System
.	O
</s>
<s>
The	O
failure	B-Operating_System
detector	I-Operating_System
is	O
an	O
abstract	O
version	O
of	O
timeouts	O
;	O
it	O
signals	O
to	O
each	O
process	O
when	O
other	O
processes	O
may	O
have	O
crashed	O
.	O
</s>
<s>
An	O
eventually	O
strong	O
failure	B-Operating_System
detector	I-Operating_System
is	O
one	O
that	O
never	O
identifies	O
some	O
specific	O
non-faulty	O
process	O
as	O
having	O
failed	O
after	O
some	O
initial	O
period	O
of	O
confusion	O
,	O
and	O
,	O
at	O
the	O
same	O
time	O
,	O
eventually	O
identifies	O
all	O
faulty	O
processes	O
as	O
failed	O
(	O
where	O
a	O
faulty	O
process	O
is	O
a	O
process	O
which	O
eventually	O
fails	O
or	O
crashes	O
and	O
a	O
non-faulty	O
process	O
never	O
fails	O
)	O
.	O
</s>
<s>
The	O
Chandra	B-Operating_System
–	I-Operating_System
Toueg	I-Operating_System
consensus	I-Operating_System
algorithm	I-Operating_System
assumes	O
that	O
the	O
number	O
of	O
faulty	O
processes	O
,	O
denoted	O
by	O
,	O
is	O
less	O
than	O
n/2	O
(	O
i.e.	O
</s>
<s>
Each	O
process	O
waits	O
(	O
1	O
)	O
to	O
receive	O
(	O
r	O
,	O
preference	O
)	O
from	O
the	O
coordinator	O
,	O
or	O
(	O
2	O
)	O
for	O
its	O
failure	B-Operating_System
detector	I-Operating_System
to	O
identify	O
the	O
coordinator	O
as	O
crashed	O
.	O
</s>
<s>
An	O
algorithm	O
which	O
"	O
solves	O
"	O
the	O
consensus	B-Operating_System
problem	I-Operating_System
must	O
ensure	O
the	O
following	O
properties	O
:	O
</s>
<s>
Before	O
arguing	O
that	O
the	O
Chandra	B-Operating_System
–	I-Operating_System
Toueg	I-Operating_System
consensus	I-Operating_System
algorithm	I-Operating_System
satisfies	O
the	O
three	O
properties	O
above	O
,	O
recall	O
that	O
this	O
algorithm	O
requires	O
=	O
2*	O
+	O
1	O
processes	O
,	O
where	O
at	O
most	O
f	O
of	O
which	O
are	O
faulty	O
.	O
</s>
<s>
Furthermore	O
,	O
note	O
that	O
this	O
algorithm	O
assumes	O
the	O
existence	O
of	O
eventually	O
strong	O
failure	B-Operating_System
detector	I-Operating_System
(	O
which	O
are	O
accessible	O
and	O
can	O
be	O
used	O
to	O
detect	O
the	O
crash	O
of	O
a	O
node	O
)	O
.	O
</s>
<s>
An	O
eventually	O
strong	O
failure	B-Operating_System
detector	I-Operating_System
is	O
one	O
that	O
never	O
identifies	O
some	O
specific	O
non-faulty	O
(	O
or	O
correct	O
)	O
process	O
as	O
having	O
failed	O
,	O
after	O
some	O
initial	O
period	O
of	O
confusion	O
,	O
and	O
,	O
at	O
the	O
same	O
time	O
,	O
eventually	O
identifies	O
all	O
faulty	O
processes	O
as	O
failed	O
.	O
</s>
<s>
Termination	O
holds	O
because	O
eventually	O
the	O
failure	B-Operating_System
detector	I-Operating_System
stops	O
suspecting	O
some	O
non-faulty	O
process	O
p	O
and	O
eventually	O
p	O
becomes	O
the	O
coordinator	O
.	O
</s>
