<s>
Grid	B-Architecture
computing	I-Architecture
is	O
the	O
use	O
of	O
widely	O
distributed	O
computer	O
resources	B-General_Concept
to	O
reach	O
a	O
common	O
goal	O
.	O
</s>
<s>
A	O
computing	B-Architecture
grid	I-Architecture
can	O
be	O
thought	O
of	O
as	O
a	O
distributed	B-Architecture
system	I-Architecture
with	O
non-interactive	O
workloads	O
that	O
involve	O
many	O
files	O
.	O
</s>
<s>
Grid	B-Architecture
computing	I-Architecture
is	O
distinguished	O
from	O
conventional	O
high-performance	B-Architecture
computing	I-Architecture
systems	O
such	O
as	O
cluster	B-Architecture
computing	I-Architecture
in	O
that	O
grid	O
computers	O
have	O
each	O
node	B-Data_Structure
set	O
to	O
perform	O
a	O
different	O
task/application	O
.	O
</s>
<s>
Grid	O
computers	O
also	O
tend	O
to	O
be	O
more	O
heterogeneous	O
and	O
geographically	O
dispersed	O
(	O
thus	O
not	O
physically	O
coupled	O
)	O
than	O
cluster	B-Architecture
computers	I-Architecture
.	O
</s>
<s>
Grids	O
are	O
often	O
constructed	O
with	O
general-purpose	O
grid	O
middleware	B-General_Concept
software	O
libraries	O
.	O
</s>
<s>
Grids	O
are	O
a	O
form	O
of	O
distributed	B-Architecture
computing	I-Architecture
composed	O
of	O
many	O
networked	O
loosely	B-Application
coupled	I-Application
computers	O
acting	O
together	O
to	O
perform	O
large	O
tasks	O
.	O
</s>
<s>
For	O
certain	O
applications	O
,	O
distributed	O
or	O
grid	B-Architecture
computing	I-Architecture
can	O
be	O
seen	O
as	O
a	O
special	O
type	O
of	O
parallel	B-Operating_System
computing	I-Operating_System
that	O
relies	O
on	O
complete	O
computers	O
(	O
with	O
onboard	O
CPUs	O
,	O
storage	O
,	O
power	O
supplies	O
,	O
network	B-Architecture
interfaces	O
,	O
etc	O
.	O
)	O
</s>
<s>
connected	O
to	O
a	O
computer	B-Architecture
network	I-Architecture
(	O
private	O
or	O
public	O
)	O
by	O
a	O
conventional	O
network	B-Protocol
interface	I-Protocol
,	O
such	O
as	O
Ethernet	O
.	O
</s>
<s>
This	O
is	O
in	O
contrast	O
to	O
the	O
traditional	O
notion	O
of	O
a	O
supercomputer	B-Architecture
,	O
which	O
has	O
many	O
processors	O
connected	O
by	O
a	O
local	O
high-speed	O
computer	B-General_Concept
bus	I-General_Concept
.	O
</s>
<s>
This	O
technology	O
has	O
been	O
applied	O
to	O
computationally	O
intensive	O
scientific	O
,	O
mathematical	O
,	O
and	O
academic	O
problems	O
through	O
volunteer	B-Operating_System
computing	I-Operating_System
,	O
and	O
it	O
is	O
used	O
in	O
commercial	O
enterprises	O
for	O
such	O
diverse	O
applications	O
as	O
drug	O
discovery	O
,	O
economic	O
forecasting	O
,	O
seismic	O
analysis	O
,	O
and	O
back	O
office	O
data	O
processing	O
in	O
support	O
for	O
e-commerce	O
and	O
Web	O
services	O
.	O
</s>
<s>
Grid	B-Architecture
computing	I-Architecture
combines	O
computers	O
from	O
multiple	O
administrative	B-Protocol
domains	I-Protocol
to	O
reach	O
a	O
common	O
goal	O
,	O
to	O
solve	O
a	O
single	O
task	O
,	O
and	O
may	O
then	O
disappear	O
just	O
as	O
quickly	O
.	O
</s>
<s>
The	O
size	O
of	O
a	O
grid	O
may	O
vary	O
from	O
small	O
—	O
confined	O
to	O
a	O
network	B-Architecture
of	O
computer	O
workstations	O
within	O
a	O
corporation	O
,	O
for	O
example	O
—	O
to	O
large	O
,	O
public	O
collaborations	O
across	O
many	O
companies	O
and	O
networks	O
.	O
</s>
<s>
"	O
The	O
notion	O
of	O
a	O
confined	O
grid	O
may	O
also	O
be	O
known	O
as	O
an	O
intra-nodes	O
cooperation	O
whereas	O
the	O
notion	O
of	O
a	O
larger	O
,	O
wider	O
grid	O
may	O
thus	O
refer	O
to	O
an	O
inter-nodes	O
cooperation	O
"	O
.	O
</s>
<s>
Coordinating	O
applications	O
on	O
Grids	O
can	O
be	O
a	O
complex	O
task	O
,	O
especially	O
when	O
coordinating	O
the	O
flow	O
of	O
information	O
across	O
distributed	B-Architecture
computing	I-Architecture
resources	B-General_Concept
.	O
</s>
<s>
Grid	B-Application
workflow	I-Application
systems	O
have	O
been	O
developed	O
as	O
a	O
specialized	O
form	O
of	O
a	O
workflow	O
management	O
system	O
designed	O
specifically	O
to	O
compose	O
and	O
execute	O
a	O
series	O
of	O
computational	O
or	O
data	O
manipulation	O
steps	O
,	O
or	O
a	O
workflow	O
,	O
in	O
the	O
grid	O
context	O
.	O
</s>
<s>
“	O
Distributed	O
”	O
or	O
“	O
grid	O
”	O
computing	O
in	O
general	O
is	O
a	O
special	O
type	O
of	O
parallel	B-Operating_System
computing	I-Operating_System
that	O
relies	O
on	O
complete	O
computers	O
(	O
with	O
onboard	O
CPUs	O
,	O
storage	O
,	O
power	O
supplies	O
,	O
network	B-Architecture
interfaces	O
,	O
etc	O
.	O
)	O
</s>
<s>
connected	O
to	O
a	O
network	B-Architecture
(	O
private	O
,	O
public	O
or	O
the	O
Internet	O
)	O
by	O
a	O
conventional	O
network	B-Protocol
interface	I-Protocol
producing	O
commodity	O
hardware	O
,	O
compared	O
to	O
the	O
lower	O
efficiency	O
of	O
designing	O
and	O
constructing	O
a	O
small	O
number	O
of	O
custom	O
supercomputers	B-Architecture
.	O
</s>
<s>
This	O
arrangement	O
is	O
thus	O
well-suited	O
to	O
applications	O
in	O
which	O
multiple	O
parallel	B-Operating_System
computations	I-Operating_System
can	O
take	O
place	O
independently	O
,	O
without	O
the	O
need	O
to	O
communicate	O
intermediate	O
results	O
between	O
processors	O
.	O
</s>
<s>
The	O
high-end	O
scalability	B-Architecture
of	O
geographically	O
dispersed	O
grids	O
is	O
generally	O
favorable	O
,	O
due	O
to	O
the	O
low	O
need	O
for	O
connectivity	O
between	O
nodes	B-Data_Structure
relative	O
to	O
the	O
capacity	O
of	O
the	O
public	O
Internet	O
.	O
</s>
<s>
It	O
can	O
be	O
costly	O
and	O
difficult	O
to	O
write	O
programs	O
that	O
can	O
run	O
in	O
the	O
environment	O
of	O
a	O
supercomputer	B-Architecture
,	O
which	O
may	O
have	O
a	O
custom	O
operating	B-General_Concept
system	I-General_Concept
,	O
or	O
require	O
the	O
program	O
to	O
address	O
concurrency	B-Operating_System
issues	O
.	O
</s>
<s>
If	O
a	O
problem	O
can	O
be	O
adequately	O
parallelized	B-Operating_System
,	O
a	O
“	O
thin	O
”	O
layer	O
of	O
“	O
grid	O
”	O
infrastructure	O
can	O
allow	O
conventional	O
,	O
standalone	O
programs	O
,	O
given	O
a	O
different	O
part	O
of	O
the	O
same	O
problem	O
,	O
to	O
run	O
on	O
multiple	O
machines	O
.	O
</s>
<s>
This	O
makes	O
it	O
possible	O
to	O
write	O
and	O
debug	O
on	O
a	O
single	O
conventional	O
machine	O
and	O
eliminates	O
complications	O
due	O
to	O
multiple	O
instances	O
of	O
the	O
same	O
program	O
running	O
in	O
the	O
same	O
shared	O
memory	B-General_Concept
and	O
storage	O
space	O
at	O
the	O
same	O
time	O
.	O
</s>
<s>
One	O
feature	O
of	O
distributed	O
grids	O
is	O
that	O
they	O
can	O
be	O
formed	O
from	O
computing	B-General_Concept
resources	I-General_Concept
belonging	O
to	O
one	O
or	O
multiple	O
individuals	O
or	O
organizations	O
(	O
known	O
as	O
multiple	O
administrative	B-Protocol
domains	I-Protocol
)	O
.	O
</s>
<s>
This	O
can	O
facilitate	O
commercial	O
transactions	O
,	O
as	O
in	O
utility	B-Operating_System
computing	I-Operating_System
,	O
or	O
make	O
it	O
easier	O
to	O
assemble	O
volunteer	B-Operating_System
computing	I-Operating_System
networks	O
.	O
</s>
<s>
This	O
often	O
involves	O
assigning	O
work	O
randomly	O
to	O
different	O
nodes	B-Data_Structure
(	O
presumably	O
with	O
different	O
owners	O
)	O
and	O
checking	O
that	O
at	O
least	O
two	O
different	O
nodes	B-Data_Structure
report	O
the	O
same	O
answer	O
for	O
a	O
given	O
work	O
unit	O
.	O
</s>
<s>
Discrepancies	O
would	O
identify	O
malfunctioning	O
and	O
malicious	O
nodes	B-Data_Structure
.	O
</s>
<s>
However	O
,	O
due	O
to	O
the	O
lack	O
of	O
central	O
control	O
over	O
the	O
hardware	O
,	O
there	O
is	O
no	O
way	O
to	O
guarantee	O
that	O
nodes	B-Data_Structure
will	O
not	O
drop	O
out	O
of	O
the	O
network	B-Architecture
at	O
random	O
times	O
.	O
</s>
<s>
Some	O
nodes	B-Data_Structure
(	O
like	O
laptops	O
or	O
dial-up	O
Internet	O
customers	O
)	O
may	O
also	O
be	O
available	O
for	O
computation	O
but	O
not	O
network	B-Architecture
communications	O
for	O
unpredictable	O
periods	O
.	O
</s>
<s>
These	O
variations	O
can	O
be	O
accommodated	O
by	O
assigning	O
large	O
work	O
units	O
(	O
thus	O
reducing	O
the	O
need	O
for	O
continuous	O
network	B-Architecture
connectivity	O
)	O
and	O
reassigning	O
work	O
units	O
when	O
a	O
given	O
node	B-Data_Structure
fails	O
to	O
report	O
its	O
results	O
in	O
expected	O
time	O
.	O
</s>
<s>
Another	O
set	O
of	O
what	O
could	O
be	O
termed	O
social	O
compatibility	O
issues	O
in	O
the	O
early	O
days	O
of	O
grid	B-Architecture
computing	I-Architecture
related	O
to	O
the	O
goals	O
of	O
grid	O
developers	O
to	O
carry	O
their	O
innovation	O
beyond	O
the	O
original	O
field	O
of	O
high-performance	B-Architecture
computing	I-Architecture
and	O
across	O
disciplinary	O
boundaries	O
into	O
new	O
fields	O
,	O
like	O
that	O
of	O
high-energy	O
physics	O
.	O
</s>
<s>
The	O
impacts	O
of	O
trust	O
and	O
availability	O
on	O
performance	O
and	O
development	O
difficulty	O
can	O
influence	O
the	O
choice	O
of	O
whether	O
to	O
deploy	O
onto	O
a	O
dedicated	O
cluster	B-Architecture
,	O
to	O
idle	O
machines	O
internal	O
to	O
the	O
developing	O
organization	O
,	O
or	O
to	O
an	O
open	O
external	O
network	B-Architecture
of	O
volunteers	O
or	O
contractors	O
.	O
</s>
<s>
In	O
many	O
cases	O
,	O
the	O
participating	O
nodes	B-Data_Structure
must	O
trust	O
the	O
central	O
system	O
not	O
to	O
abuse	O
the	O
access	O
that	O
is	O
being	O
granted	O
,	O
by	O
interfering	O
with	O
the	O
operation	O
of	O
other	O
programs	O
,	O
mangling	O
stored	O
information	O
,	O
transmitting	O
private	O
data	O
,	O
or	O
creating	O
new	O
security	O
holes	O
.	O
</s>
<s>
Other	O
systems	O
employ	O
measures	O
to	O
reduce	O
the	O
amount	O
of	O
trust	O
“	O
client	O
”	O
nodes	B-Data_Structure
must	O
place	O
in	O
the	O
central	O
system	O
such	O
as	O
placing	O
applications	O
in	O
virtual	O
machines	O
.	O
</s>
<s>
Public	O
systems	O
or	O
those	O
crossing	O
administrative	B-Protocol
domains	I-Protocol
(	O
including	O
different	O
departments	O
in	O
the	O
same	O
organization	O
)	O
often	O
result	O
in	O
the	O
need	O
to	O
run	O
on	O
heterogeneous	O
systems	O
,	O
using	O
different	O
operating	B-General_Concept
systems	I-General_Concept
and	O
hardware	B-General_Concept
architectures	I-General_Concept
.	O
</s>
<s>
With	O
many	O
languages	O
,	O
there	O
is	O
a	O
trade-off	O
between	O
investment	O
in	O
software	O
development	O
and	O
the	O
number	O
of	O
platforms	O
that	O
can	O
be	O
supported	O
(	O
and	O
thus	O
the	O
size	O
of	O
the	O
resulting	O
network	B-Architecture
)	O
.	O
</s>
<s>
Cross-platform	B-Operating_System
languages	O
can	O
reduce	O
the	O
need	O
to	O
make	O
this	O
tradeoff	O
,	O
though	O
potentially	O
at	O
the	O
expense	O
of	O
high	O
performance	O
on	O
any	O
given	O
node	B-Data_Structure
(	O
due	O
to	O
run-time	O
interpretation	O
or	O
lack	O
of	O
optimization	O
for	O
the	O
particular	O
platform	O
)	O
.	O
</s>
<s>
Various	O
middleware	B-General_Concept
projects	O
have	O
created	O
generic	O
infrastructure	O
to	O
allow	O
diverse	O
scientific	O
and	O
commercial	O
projects	O
to	O
harness	O
a	O
particular	O
associated	O
grid	O
or	O
for	O
the	O
purpose	O
of	O
setting	O
up	O
new	O
grids	O
.	O
</s>
<s>
BOINC	B-Operating_System
is	O
a	O
common	O
one	O
for	O
various	O
academic	O
projects	O
seeking	O
public	O
volunteers	O
;	O
more	O
are	O
listed	O
at	O
the	O
end	O
of	O
the	O
article	O
.	O
</s>
<s>
In	O
fact	O
,	O
the	O
middleware	B-General_Concept
can	O
be	O
seen	O
as	O
a	O
layer	O
between	O
the	O
hardware	O
and	O
the	O
software	O
.	O
</s>
<s>
On	O
top	O
of	O
the	O
middleware	B-General_Concept
,	O
a	O
number	O
of	O
technical	O
areas	O
have	O
to	O
be	O
considered	O
,	O
and	O
these	O
may	O
or	O
may	O
not	O
be	O
middleware	B-General_Concept
independent	O
.	O
</s>
<s>
Example	O
areas	O
include	O
SLA	O
management	O
,	O
Trust	O
,	O
and	O
Security	O
,	O
Virtual	B-Operating_System
organization	I-Operating_System
management	O
,	O
License	O
Management	O
,	O
Portals	O
and	O
Data	O
Management	O
.	O
</s>
<s>
For	O
the	O
segmentation	O
of	O
the	O
grid	B-Architecture
computing	I-Architecture
market	O
,	O
two	O
perspectives	O
need	O
to	O
be	O
considered	O
:	O
the	O
provider	O
side	O
and	O
the	O
user	O
side	O
:	O
</s>
<s>
These	O
are	O
the	O
grid	O
middleware	B-General_Concept
market	O
,	O
the	O
market	O
for	O
grid-enabled	O
applications	O
,	O
the	O
utility	B-Operating_System
computing	I-Operating_System
market	O
,	O
and	O
the	O
software-as-a-service	B-Architecture
(	O
SaaS	B-Architecture
)	O
market	O
.	O
</s>
<s>
Grid	O
middleware	B-General_Concept
is	O
a	O
specific	O
software	O
product	O
,	O
which	O
enables	O
the	O
sharing	O
of	O
heterogeneous	O
resources	B-General_Concept
,	O
and	O
Virtual	B-Operating_System
Organizations	I-Operating_System
.	O
</s>
<s>
Major	O
grid	O
middlewares	B-General_Concept
are	O
Globus	O
Toolkit	O
,	O
gLite	B-Operating_System
,	O
and	O
UNICORE	B-General_Concept
.	O
</s>
<s>
Utility	B-Operating_System
computing	I-Operating_System
is	O
referred	O
to	O
as	O
the	O
provision	O
of	O
grid	B-Architecture
computing	I-Architecture
and	O
applications	O
as	O
service	O
either	O
as	O
an	O
open	O
grid	O
utility	B-Operating_System
or	O
as	O
a	O
hosting	O
solution	O
for	O
one	O
organization	O
or	O
a	O
VO	B-Operating_System
.	O
</s>
<s>
Major	O
players	O
in	O
the	O
utility	B-Operating_System
computing	I-Operating_System
market	O
are	O
Sun	O
Microsystems	O
,	O
IBM	O
,	O
and	O
HP	O
.	O
</s>
<s>
This	O
is	O
made	O
possible	O
by	O
the	O
use	O
of	O
grid	O
middleware	B-General_Concept
,	O
as	O
pointed	O
out	O
above	O
.	O
</s>
<s>
Software	B-Architecture
as	I-Architecture
a	I-Architecture
service	I-Architecture
(	O
SaaS	B-Architecture
)	O
is	O
“	O
software	O
that	O
is	O
owned	O
,	O
delivered	O
and	O
managed	O
remotely	O
by	O
one	O
or	O
more	O
providers.	O
”	O
(	O
Gartner	O
2007	O
)	O
Additionally	O
,	O
SaaS	B-Architecture
applications	O
are	O
based	O
on	O
a	O
single	O
set	O
of	O
common	O
code	O
and	O
data	O
definitions	O
.	O
</s>
<s>
They	O
are	O
consumed	O
in	O
a	O
one-to-many	O
model	O
,	O
and	O
SaaS	B-Architecture
uses	O
a	O
Pay	O
As	O
You	O
Go	O
(	O
PAYG	O
)	O
model	O
or	O
a	O
subscription	O
model	O
that	O
is	O
based	O
on	O
usage	O
.	O
</s>
<s>
Providers	O
of	O
SaaS	B-Architecture
do	O
not	O
necessarily	O
own	O
the	O
computing	B-General_Concept
resources	I-General_Concept
themselves	O
,	O
which	O
are	O
required	O
to	O
run	O
their	O
SaaS	B-Architecture
.	O
</s>
<s>
Therefore	O
,	O
SaaS	B-Architecture
providers	O
may	O
draw	O
upon	O
the	O
utility	B-Operating_System
computing	I-Operating_System
market	O
.	O
</s>
<s>
The	O
utility	B-Operating_System
computing	I-Operating_System
market	O
provides	O
computing	B-General_Concept
resources	I-General_Concept
for	O
SaaS	B-Architecture
providers	O
.	O
</s>
<s>
For	O
companies	O
on	O
the	O
demand	O
or	O
user	O
side	O
of	O
the	O
grid	B-Architecture
computing	I-Architecture
market	O
,	O
the	O
different	O
segments	O
have	O
significant	O
implications	O
for	O
their	O
IT	O
deployment	O
strategy	O
.	O
</s>
<s>
CPU-scavenging	O
,	O
cycle-scavenging	O
,	O
or	O
shared	O
computing	O
creates	O
a	O
“	O
grid	O
”	O
from	O
the	O
idle	O
resources	B-General_Concept
in	O
a	O
network	B-Architecture
of	O
participants	O
(	O
whether	O
worldwide	O
or	O
internal	O
to	O
an	O
organization	O
)	O
.	O
</s>
<s>
Typically	O
,	O
this	O
technique	O
exploits	O
the	O
'	O
spare	O
 '	O
instruction	B-General_Concept
cycles	I-General_Concept
resulting	O
from	O
the	O
intermittent	O
inactivity	O
that	O
typically	O
occurs	O
at	O
night	O
,	O
during	O
lunch	O
breaks	O
,	O
or	O
even	O
during	O
the	O
(	O
comparatively	O
minuscule	O
,	O
though	O
numerous	O
)	O
moments	O
of	O
idle	O
waiting	O
that	O
modern	O
desktop	O
CPU	O
's	O
experience	O
throughout	O
the	O
day	O
(	O
when	B-General_Concept
the	I-General_Concept
computer	I-General_Concept
is	I-General_Concept
waiting	I-General_Concept
on	I-General_Concept
IO	I-General_Concept
from	I-General_Concept
the	I-General_Concept
user	I-General_Concept
,	I-General_Concept
network	I-General_Concept
,	I-General_Concept
or	I-General_Concept
storage	I-General_Concept
)	O
.	O
</s>
<s>
In	O
practice	O
,	O
participating	O
computers	O
also	O
donate	O
some	O
supporting	O
amount	O
of	O
disk	O
storage	O
space	O
,	O
RAM	O
,	O
and	O
network	B-Architecture
bandwidth	O
,	O
in	O
addition	O
to	O
raw	O
CPU	O
power	O
.	O
</s>
<s>
Many	O
volunteer	B-Operating_System
computing	I-Operating_System
projects	O
,	O
such	O
as	O
BOINC	B-Operating_System
,	O
use	O
the	O
CPU	O
scavenging	O
model	O
.	O
</s>
<s>
Since	O
nodes	B-Data_Structure
are	O
likely	O
to	O
go	O
"	O
offline	O
"	O
from	O
time	O
to	O
time	O
,	O
as	O
their	O
owners	O
use	O
their	O
resources	B-General_Concept
for	O
their	O
primary	O
purpose	O
,	O
this	O
model	O
must	O
be	O
designed	O
to	O
handle	O
such	O
contingencies	O
.	O
</s>
<s>
For	O
instance	O
,	O
HTCondor	B-Operating_System
(	O
the	O
open-source	O
high-throughput	B-Operating_System
computing	I-Operating_System
software	O
framework	O
for	O
coarse-grained	O
distributed	O
rationalization	O
of	O
computationally	O
intensive	O
tasks	O
)	O
can	O
be	O
configured	O
to	O
only	O
use	O
desktop	O
machines	O
where	O
the	O
keyboard	O
and	O
mouse	O
are	O
idle	O
to	O
effectively	O
harness	O
wasted	O
CPU	O
power	O
from	O
otherwise	O
idle	O
desktop	O
workstations	O
.	O
</s>
<s>
Like	O
other	O
full-featured	O
batch	O
systems	O
,	O
HTCondor	B-Operating_System
provides	O
a	O
job	O
queueing	O
mechanism	O
,	O
scheduling	O
policy	O
,	O
priority	O
scheme	O
,	O
resource	O
monitoring	O
,	O
and	O
resource	O
management	O
.	O
</s>
<s>
It	O
can	O
be	O
used	O
to	O
manage	O
workload	O
on	O
a	O
dedicated	O
cluster	B-Architecture
of	O
computers	O
as	O
well	O
or	O
it	O
can	O
seamlessly	O
integrate	O
both	O
dedicated	O
resources	B-General_Concept
(	O
rack-mounted	O
clusters	O
)	O
and	O
non-dedicated	O
desktop	O
machines	O
(	O
cycle	O
scavenging	O
)	O
into	O
one	O
computing	O
environment	O
.	O
</s>
<s>
The	O
term	O
grid	B-Architecture
computing	I-Architecture
originated	O
in	O
the	O
early	O
1990s	O
as	O
a	O
metaphor	O
for	O
making	O
computer	O
power	O
as	O
easy	O
to	O
access	O
as	O
an	O
electric	O
power	O
grid	O
.	O
</s>
<s>
This	O
was	O
preceded	O
by	O
decades	O
by	O
the	O
metaphor	O
of	O
utility	B-Operating_System
computing	I-Operating_System
(	O
1961	O
)	O
:	O
computing	O
as	O
a	O
public	O
utility	B-Operating_System
,	O
analogous	O
to	O
the	O
phone	O
system	O
.	O
</s>
<s>
CPU	O
scavenging	O
and	O
volunteer	B-Operating_System
computing	I-Operating_System
were	O
popularized	O
beginning	O
in	O
1997	O
by	O
distributed.net	O
and	O
later	O
in	O
1999	O
by	O
SETI	B-Application
@home	I-Application
to	O
harness	O
the	O
power	O
of	O
networked	O
PCs	O
worldwide	O
,	O
in	O
order	O
to	O
solve	O
CPU-intensive	O
research	O
problems	O
.	O
</s>
<s>
The	O
ideas	O
of	O
the	O
grid	O
(	O
including	O
those	O
from	O
distributed	B-Architecture
computing	I-Architecture
,	O
object-oriented	O
programming	O
,	O
and	O
Web	O
services	O
)	O
were	O
brought	O
together	O
by	O
Ian	O
Foster	O
and	O
Steve	O
Tuecke	O
of	O
the	O
University	O
of	O
Chicago	O
,	O
and	O
Carl	O
Kesselman	O
of	O
the	O
University	O
of	O
Southern	O
California	O
's	O
Information	O
Sciences	O
Institute	O
.	O
</s>
<s>
The	O
toolkit	O
incorporates	O
not	O
just	O
computation	O
management	O
but	O
also	O
storage	B-Application
management	I-Application
,	O
security	O
provisioning	O
,	O
data	O
movement	O
,	O
monitoring	O
,	O
and	O
a	O
toolkit	O
for	O
developing	O
additional	O
services	O
based	O
on	O
the	O
same	O
infrastructure	O
,	O
including	O
agreement	O
negotiation	O
,	O
notification	O
mechanisms	O
,	O
trigger	O
services	O
,	O
and	O
information	O
aggregation	O
.	O
</s>
<s>
In	O
2007	O
the	O
term	O
cloud	B-Architecture
computing	I-Architecture
came	O
into	O
popularity	O
,	O
which	O
is	O
conceptually	O
similar	O
to	O
the	O
canonical	O
Foster	O
definition	O
of	O
grid	B-Architecture
computing	I-Architecture
(	O
in	O
terms	O
of	O
computing	B-General_Concept
resources	I-General_Concept
being	O
consumed	O
as	O
electricity	O
is	O
from	O
the	O
power	O
grid	O
)	O
and	O
earlier	O
utility	B-Operating_System
computing	I-Operating_System
.	O
</s>
<s>
In	O
November	O
2006	O
,	O
Seidel	O
received	O
the	O
Sidney	O
Fernbach	O
Award	O
at	O
the	O
Supercomputing	B-Architecture
Conference	O
in	O
Tampa	O
,	O
Florida	O
.	O
</s>
<s>
"	O
For	O
outstanding	O
contributions	O
to	O
the	O
development	O
of	O
software	O
for	O
HPC	O
and	O
Grid	B-Architecture
computing	I-Architecture
to	O
enable	O
the	O
collaborative	O
numerical	O
investigation	O
of	O
complex	O
problems	O
in	O
physics	O
;	O
in	O
particular	O
,	O
modeling	O
black	O
hole	O
collisions.	O
"	O
</s>
<s>
As	O
of	O
April	O
7	O
,	O
2020	O
,	O
BOINC	B-Operating_System
–	O
29.8	O
PFLOPS	O
.	O
</s>
<s>
As	O
of	O
March	O
2020	O
,	O
Folding	B-Application
@home	I-Application
–	O
1.1	O
exaFLOPS	O
.	O
</s>
<s>
As	O
of	O
February	O
2018	O
,	O
Einstein	B-Operating_System
@Home	I-Operating_System
–	O
3.489	O
PFLOPS	O
.	O
</s>
<s>
As	O
of	O
April	O
7	O
,	O
2020	O
,	O
SETI	B-Application
@Home	I-Application
–	O
1.11	O
PFLOPS	O
.	O
</s>
<s>
As	O
of	O
April	O
7	O
,	O
2020	O
,	O
MilkyWay	B-Operating_System
@Home	I-Operating_System
–	O
1.465	O
PFLOPS	O
.	O
</s>
<s>
Also	O
,	O
as	O
of	O
March	O
2019	O
,	O
the	O
Bitcoin	B-General_Concept
Network	I-General_Concept
had	O
a	O
measured	O
computing	O
power	O
equivalent	O
to	O
over	O
80,000	O
exaFLOPS	O
(	O
Floating-point	O
Operations	O
Per	O
Second	O
)	O
.	O
</s>
<s>
This	O
measurement	O
reflects	O
the	O
number	O
of	O
FLOPS	O
required	O
to	O
equal	O
the	O
hash	O
output	O
of	O
the	O
Bitcoin	B-General_Concept
network	I-General_Concept
rather	O
than	O
its	O
capacity	O
for	O
general	O
floating-point	O
arithmetic	O
operations	O
,	O
since	O
the	O
elements	O
of	O
the	O
Bitcoin	B-General_Concept
network	I-General_Concept
(	O
Bitcoin	B-Protocol
mining	O
ASICs	O
)	O
perform	O
only	O
the	O
specific	O
cryptographic	O
hash	O
computation	O
required	O
by	O
the	B-General_Concept
Bitcoin	I-General_Concept
protocol	I-General_Concept
.	O
</s>
<s>
Grid	B-Architecture
computing	I-Architecture
offers	O
a	O
way	O
to	O
solve	O
Grand	B-Device
Challenge	I-Device
problems	I-Device
such	O
as	O
protein	O
folding	O
,	O
financial	O
modeling	O
,	O
earthquake	O
simulation	O
,	O
and	O
climate/weather	O
modeling	O
,	O
and	O
was	O
integral	O
in	O
enabling	O
the	O
Large	O
Hadron	O
Collider	O
at	O
CERN	O
.	O
</s>
<s>
Grids	O
offer	O
a	O
way	O
of	O
using	O
the	O
information	O
technology	O
resources	B-General_Concept
optimally	O
inside	O
an	O
organization	O
.	O
</s>
<s>
They	O
also	O
provide	O
a	O
means	O
for	O
offering	O
information	O
technology	O
as	O
a	O
utility	B-Operating_System
for	O
commercial	O
and	O
noncommercial	O
clients	O
,	O
with	O
those	O
clients	O
paying	O
only	O
for	O
what	O
they	O
use	O
,	O
as	O
with	O
electricity	O
or	O
water	O
.	O
</s>
<s>
As	O
of	O
October	O
2016	O
,	O
over	O
4	O
million	O
machines	O
running	O
the	O
open-source	O
Berkeley	B-Operating_System
Open	I-Operating_System
Infrastructure	I-Operating_System
for	I-Operating_System
Network	I-Operating_System
Computing	I-Operating_System
(	O
BOINC	B-Operating_System
)	O
platform	O
are	O
members	O
of	O
the	O
World	B-Operating_System
Community	I-Operating_System
Grid	I-Operating_System
.	O
</s>
<s>
One	O
of	O
the	O
projects	O
using	O
BOINC	B-Operating_System
is	O
SETI	B-Application
@home	I-Application
,	O
which	O
was	O
using	O
more	O
than	O
400,000	O
computers	O
to	O
achieve	O
0.828	O
TFLOPS	O
as	O
of	O
October	O
2016	O
.	O
</s>
<s>
As	O
of	O
October	O
2016	O
Folding	B-Application
@home	I-Application
,	O
which	O
is	O
not	O
part	O
of	O
BOINC	B-Operating_System
,	O
achieved	O
more	O
than	O
101	O
x86-equivalent	O
petaflops	O
on	O
over	O
110,000	O
machines	O
.	O
</s>
<s>
According	O
to	O
the	O
project	O
fact	O
sheet	O
,	O
their	O
mission	O
is	O
“	O
to	O
establish	O
effective	O
routes	O
to	O
foster	O
the	O
adoption	O
of	O
grid	B-Architecture
computing	I-Architecture
across	O
the	O
EU	O
and	O
to	O
stimulate	O
research	O
into	O
innovative	O
business	O
models	O
using	O
Grid	O
technologies	O
”	O
.	O
</s>
<s>
The	O
Enabling	B-Operating_System
Grids	I-Operating_System
for	I-Operating_System
E-sciencE	I-Operating_System
project	O
,	O
based	O
in	O
the	O
European	O
Union	O
and	O
included	O
sites	O
in	O
Asia	O
and	O
the	O
United	O
States	O
,	O
was	O
a	O
follow-up	O
project	O
to	O
the	O
European	O
DataGrid	O
(	O
EDG	O
)	O
and	O
evolved	O
into	O
the	O
European	O
Grid	O
Infrastructure	O
.	O
</s>
<s>
This	O
,	O
along	O
with	O
the	O
LHC	O
Computing	B-Architecture
Grid	I-Architecture
(	O
LCG	O
)	O
,	O
was	O
developed	O
to	O
support	O
experiments	O
using	O
the	O
CERN	O
Large	O
Hadron	O
Collider	O
.	O
</s>
<s>
The	O
NASA	B-Device
Advanced	I-Device
Supercomputing	I-Device
facility	I-Device
(	O
NAS	O
)	O
ran	O
genetic	B-Algorithm
algorithms	I-Algorithm
using	O
the	O
Condor	B-Operating_System
cycle	I-Operating_System
scavenger	I-Operating_System
running	O
on	O
about	O
350	O
Sun	O
Microsystems	O
and	O
SGI	O
workstations	O
.	O
</s>
<s>
In	O
2001	O
,	O
United	B-Operating_System
Devices	I-Operating_System
operated	O
the	O
United	B-Operating_System
Devices	I-Operating_System
Cancer	O
Research	O
Project	O
based	O
on	O
its	O
Grid	B-Operating_System
MP	I-Operating_System
product	O
,	O
which	O
cycle-scavenges	O
on	O
volunteer	O
PCs	O
connected	O
to	O
the	O
Internet	O
.	O
</s>
<s>
Today	O
there	O
are	O
many	O
definitions	O
of	O
grid	B-Architecture
computing	I-Architecture
:	O
</s>
<s>
Computing	B-General_Concept
resources	I-General_Concept
are	O
not	O
administered	O
centrally	O
.	O
</s>
<s>
IBM	O
defines	O
grid	B-Architecture
computing	I-Architecture
as	O
“	O
the	O
ability	O
,	O
using	O
a	O
set	O
of	O
open	O
standards	O
and	O
protocols	O
,	O
to	O
gain	O
access	O
to	O
applications	O
and	O
data	O
,	O
processing	O
power	O
,	O
storage	O
capacity	O
and	O
a	O
vast	O
array	O
of	O
other	O
computing	B-General_Concept
resources	I-General_Concept
over	O
the	O
Internet	O
.	O
</s>
<s>
A	O
grid	O
is	O
a	O
type	O
of	O
parallel	O
and	O
distributed	B-Architecture
system	I-Architecture
that	O
enables	O
the	O
sharing	O
,	O
selection	O
,	O
and	O
aggregation	O
of	O
resources	B-General_Concept
distributed	O
across	O
‘	O
multiple’	O
administrative	B-Protocol
domains	I-Protocol
based	O
on	O
their	O
(	O
resources	B-General_Concept
)	O
availability	O
,	O
capacity	O
,	O
performance	O
,	O
cost	O
and	O
users	O
 '	O
quality-of-service	O
requirements	O
”	O
.	O
</s>
<s>
An	O
earlier	O
example	O
of	O
the	O
notion	O
of	O
computing	O
as	O
the	O
utility	B-Operating_System
was	O
in	O
1965	O
by	O
MIT	O
's	O
Fernando	O
Corbató	O
.	O
</s>
<s>
Corbató	O
and	O
the	O
other	O
designers	O
of	O
the	O
Multics	O
operating	B-General_Concept
system	I-General_Concept
envisioned	O
a	O
computer	O
facility	O
operating	O
“	O
like	O
a	O
power	O
company	O
or	O
water	O
company	O
”	O
.	O
</s>
<s>
Buyya/Venugopal	O
define	O
grid	O
as	O
"	O
a	O
type	O
of	O
parallel	O
and	O
distributed	B-Architecture
system	I-Architecture
that	O
enables	O
the	O
sharing	O
,	O
selection	O
,	O
and	O
aggregation	O
of	O
geographically	O
distributed	O
autonomous	O
resources	B-General_Concept
dynamically	O
at	O
runtime	O
depending	O
on	O
their	O
availability	O
,	O
capability	O
,	O
performance	O
,	O
cost	O
,	O
and	O
users	O
 '	O
quality-of-service	O
requirements	O
"	O
.	O
</s>
