<s>
Shortest	B-Operating_System
remaining	I-Operating_System
time	I-Operating_System
,	O
also	O
known	O
as	O
shortest	B-Operating_System
remaining	I-Operating_System
time	I-Operating_System
first	O
(	O
SRTF	B-Operating_System
)	O
,	O
is	O
a	O
scheduling	O
method	O
that	O
is	O
a	O
preemptive	B-Operating_System
version	O
of	O
shortest	B-Operating_System
job	I-Operating_System
next	I-Operating_System
scheduling	O
.	O
</s>
<s>
In	O
this	O
scheduling	O
algorithm	O
,	O
the	O
process	B-Operating_System
with	O
the	O
smallest	O
amount	O
of	O
time	O
remaining	O
until	O
completion	O
is	O
selected	O
to	O
execute	O
.	O
</s>
<s>
Since	O
the	O
currently	O
executing	O
process	B-Operating_System
is	O
the	O
one	O
with	O
the	O
shortest	O
amount	O
of	O
time	O
remaining	O
by	O
definition	O
,	O
and	O
since	O
that	O
time	O
should	O
only	O
reduce	O
as	O
execution	O
progresses	O
,	O
the	O
process	B-Operating_System
will	O
either	O
run	O
until	O
it	O
completes	O
or	O
get	O
preempted	O
if	O
a	O
new	O
process	B-Operating_System
is	O
added	O
that	O
requires	O
a	O
smaller	O
amount	O
of	O
time	O
.	O
</s>
<s>
Shortest	B-Operating_System
remaining	I-Operating_System
time	I-Operating_System
is	O
advantageous	O
because	O
short	O
processes	O
are	O
handled	O
very	O
quickly	O
.	O
</s>
<s>
The	O
system	O
also	O
requires	O
very	O
little	O
overhead	O
since	O
it	O
only	O
makes	O
a	O
decision	O
when	O
a	O
process	B-Operating_System
completes	O
or	O
a	O
new	O
process	B-Operating_System
is	O
added	O
,	O
and	O
when	O
a	O
new	O
process	B-Operating_System
is	O
added	O
the	O
algorithm	O
only	O
needs	O
to	O
compare	O
the	O
currently	O
executing	O
process	B-Operating_System
with	O
the	O
new	O
process	B-Operating_System
,	O
ignoring	O
all	O
other	O
processes	O
currently	O
waiting	O
to	O
execute	O
.	O
</s>
<s>
Like	O
shortest	B-Operating_System
job	I-Operating_System
next	I-Operating_System
,	O
it	O
has	O
the	O
potential	O
for	O
process	B-Operating_System
starvation	I-Operating_System
:	O
long	O
processes	O
may	O
be	O
held	O
off	O
indefinitely	O
if	O
short	O
processes	O
are	O
continually	O
added	O
.	O
</s>
<s>
This	O
threat	O
can	O
be	O
minimal	O
when	O
process	B-Operating_System
times	O
follow	O
a	O
heavy-tailed	O
distribution	O
.	O
</s>
<s>
A	O
similar	O
algorithm	O
which	O
avoids	O
starvation	B-Operating_System
at	O
the	O
cost	O
of	O
higher	O
tracking	O
overhead	O
is	O
highest	B-Operating_System
response	I-Operating_System
ratio	I-Operating_System
next	I-Operating_System
(	O
HRRN	B-Operating_System
)	O
.	O
</s>
<s>
Like	O
shortest	B-Operating_System
job	I-Operating_System
next	I-Operating_System
scheduling	O
,	O
shortest	B-Operating_System
remaining	I-Operating_System
time	I-Operating_System
scheduling	I-Operating_System
is	O
rarely	O
used	O
outside	O
of	O
specialized	O
environments	O
because	O
it	O
requires	O
accurate	O
estimates	O
of	O
the	O
runtime	O
of	O
each	O
process	B-Operating_System
.	O
</s>
