vo-centric ganglia catalin lucian dumitrescu, mike...
Post on 12-Sep-2018
217 Views
Preview:
TRANSCRIPT
VO
-cen
tric
Gan
glia
Cat
alin
Lu
cian
Du
mit
resc
u, M
ike
Wil
de*
, Ro
b G
ard
ner
Th
e U
niv
ersi
ty o
f C
hic
ago
* A
rgon
ne
Na
tion
al
La
bor
ato
ryJu
ly -
Au
gust
2k2
Wo
rk G
oal
sG
lob
al R
eso
urc
e U
sage
Vie
wer
(an
d n
ot
on
ly)
at t
he
Gri
d L
evel
Gri
d a
nd
VO
leve
l vie
ws
of
reso
urc
e ac
tivi
tyu
tili
zati
on
per
form
ance
pol
icy
poi
nt
of v
iew
: ach
ieve
d v
ersu
s p
rovi
ded
Ter
ms
Gri
d: c
oll
ecti
on
of r
eso
urc
e fr
acti
on
s (s
et o
f sit
es p
rovi
ded
for
usa
ge)
Exa
mp
le: A
TL
AS
grid
, CM
S gr
idV
O: c
oll
ecti
on
of i
den
tifi
able
en
titi
es e
nti
tled
to
use
gri
d r
eso
urc
esE
xam
ple
: AT
LA
S V
O (
use
rs, g
rou
ps)
, CM
S V
O (
use
rs, g
rou
ps)
Req
uir
emen
tsE
ven
t lo
ggin
gs a
nd
sta
tist
ic v
iew
s fo
rm
ach
ines
, clu
ster
s, s
ites
, gri
ds
use
rs, g
rou
ps,
VO
s, w
ork
typ
esjo
bs
Op
erat
ion
s(t
o a
chie
ve s
cala
bil
ity
in p
rovi
din
g m
ean
ingf
ul i
nfo
rma
tion
)re
trie
ve /
co
llec
tag
greg
ate
sto
revi
zual
ize
en
d u
ser
/ in
telli
ge
nt
job
sub
mis
sio
n p
oin
t /
po
licy
en
forc
em
en
t
Pre
sen
tati
on
Ou
tlin
e1.
In
tro
du
ctio
n, G
oal
s &
Req
uir
emen
ts
2. E
xist
ing
Wo
rk O
verv
iew
3. A
cco
mp
lish
ed W
ork
a. D
ata
Su
mm
ati
on a
nd
Hie
rarc
hic
Rep
orti
ng
b. M
DS
Ad
dre
ss P
ub
lish
ing
c. L
oca
l Sen
sors
Sp
ecif
ica
tion
d. D
ata
Ret
riev
al f
rom
Na
gios
4. S
cree
nsh
ots
/ D
emo
in a
Sim
ula
ted
En
viro
nm
ent
5. N
agio
s W
eb I
nte
rfac
e In
corp
ora
tio
n
6. C
urr
ent
Stat
us
& C
on
clu
sio
ns
& R
emai
nin
g &
Fu
ture
Wo
rk
Gan
glia
Ove
rvie
wR
eal-
tim
e m
on
ito
rin
g (a
nd
exe
cuti
on
en
viro
nm
ent)
run
s o
n a
sin
gle
ho
stco
llec
ts d
ata
fro
m o
ther
ho
st-l
evel
mo
nit
ori
ng
too
ls (
e.g.
, gm
on
d, r
ock
s,n
agi
os-e
xpor
ter)
or
even
oth
er g
angl
ia d
aem
on
sd
ata
is m
anip
ula
ted
in X
ML
fo
rmat
and
sto
red
loca
lly
in a
RR
D fo
r st
atis
tics
pu
rpo
ses
use
s h
ost
-lev
el c
oll
ecto
rs (
gmo
nd
s) f
or
loa
d,
me
mo
ry,
swa
p,
dis
k,n
etw
ork
, u
ser
de
fin
ed
me
tric
s
com
pu
te n
od
es
com
pu
te n
od
es
RR
D
qu
erie
s
gan
glia
da
emon
gmon
d d
aem
ons
gmon
d d
aem
ons
Web
In
terf
ace
Ove
rvie
wP
rovi
des
an
eas
y-to
-use
an
d p
ow
erfu
l mec
han
ism
to
vie
wgr
aph
ical
ly g
ath
ered
info
rmat
ion
Op
erat
ion
s:lo
cal d
ata
retr
ieva
lm
eta-
clu
ster
su
mm
atio
ns
dif
fere
nt
view
s (m
eta-
clu
ster
vie
ws,
clu
ster
ove
rvie
ws,
ho
st d
etai
ls)
Lim
itat
ion
s:o
nly
tw
o le
vel:
clu
ster
, ho
std
ata
mu
st r
esid
e o
n t
he
web
serv
er n
od
e
Mo
re in
form
atio
n:
htt
p:/
/gan
glia
.so
urc
efo
rge.
net
/
Nag
ios(
TM
) as
Alt
ern
ativ
eA
few
wo
rds
abo
ut:
Cen
tral
ized
mo
nit
ori
ng
too
l fo
r se
rvic
e st
atu
s d
ata
gat
her
ing
Co
mp
lex
spec
ific
atio
n f
iles
(ca
n b
e au
tom
atic
ally
gen
erat
ed)
Po
wer
ful w
eb in
terf
ace,
wic
h p
rovi
des
su
pp
ort
fo
r re
mo
teco
mm
and
s
Pre
sen
tati
on
Ou
tlin
e1.
In
tro
du
ctio
n, G
oal
s &
Req
uir
emen
ts
2. E
xist
ing
Wo
rk O
verv
iew
3. A
cco
mp
lish
ed W
ork
a. D
ata
Su
mm
ati
on a
nd
Hie
rarc
hic
Rep
orti
ng
b. M
DS
Ad
dre
ss P
ub
lish
ing
c. L
oca
l Sen
sors
Sp
ecif
ica
tion
d. D
ata
Ret
riev
al f
rom
Na
gios
4. S
cree
nsh
ots
/ D
emo
in a
Sim
ula
ted
En
viro
nm
ent
5. N
agio
s W
eb I
nte
rfac
e In
corp
ora
tio
n
6. C
urr
ent
Stat
us
& C
on
clu
sio
ns
& R
emai
nin
g &
Fu
ture
Wo
rk
Hie
rarc
hic
Rep
ort
ing
Th
e B
ig P
ictu
re:
gm
on
dgm
on
dgm
on
d
gan
glia
RR
D
gmo
nd
gmo
nd
gmo
nd
gan
glia
RR
D
gmo
nd
gmo
nd
gmo
nd
gan
glia
RR
D
gan
glia ga
ngl
ia
gan
glia
...
EU
gri
d
US
grid
Dat
a Su
mm
atio
nP
rob
lem
:gr
ids
are
larg
e co
llec
tio
ns
of c
om
pu
ters
mo
nit
ori
ng
info
rmat
ion
is d
iffi
cult
to
be
pas
sed
aro
un
dst
and
ard
gan
glia
do
es n
ot
pro
vid
e su
pp
ort
fo
r la
yere
d r
epo
rtin
g
Ap
pro
ach
ed s
olu
tio
n:
gath
er in
form
atio
n, s
um
an
d p
rovi
de
on
ly a
n o
verv
iew
Ho
w (
seve
ral t
ypes
of
dat
a):
stri
ngs
nu
mer
ics
sum
(m
emo
ry, c
pu
nu
mb
ers,
..)
aver
age
(cp
u_l
oad
, ...)
Lo
cal S
enso
rsA
s a
mea
ns
for:
cap
turi
ng
add
itio
nal
info
rmat
ion
:ga
teke
eper
s /
clu
ster
s /
site
s /
grid
s sp
ecif
ic d
ata,
e.g
.:ga
teke
eper
ad
dre
sses
an
d t
ypes
po
licy
info
rmat
ion
use
rs, g
rou
ps,
an
d V
Os
usa
ge s
tati
stic
s
min
imal
inte
ract
ion
wit
h o
ther
loca
l so
ftw
are
com
po
nen
ts
dis
tin
ctio
n o
f in
form
atio
n p
ub
lish
ed t
hro
ugh
XM
L
In a
nu
tsh
ell,
no
t a
req
uir
emen
t, b
ut
an u
sefu
l fea
ture
MD
S A
dd
ress
/Ret
riev
alM
oti
vati
on
:h
ard
to
mai
nta
in c
entr
aliz
ed c
on
figu
rati
on
fil
es (
Nag
ios)
imp
oss
ible
to
cap
ture
dyn
amic
ch
ange
s o
f th
e en
viro
nm
ent
real
co
nn
ecti
on
wit
h t
he
Gri
d c
om
po
nen
ts
Tw
o o
per
atio
ns
wer
e in
tro
du
ced
:G
angl
ia D
aem
on
Ad
dre
ss P
ub
lish
ing:
by
mea
ns
of
a gr
id p
ub
lish
erG
angl
ia A
dd
ress
es R
etri
eval
: by
mea
ns
of
a sp
ecia
l mo
du
le t
hat
qu
erie
sG
RIS
/GII
S se
rver
s
Inte
rfac
ing
wit
h N
agio
sP
rob
lem
:d
iffe
ren
t si
tes/
grid
hav
e sp
ent
con
sid
erab
le a
mo
un
t o
f w
ork
tim
e in
imp
rovi
ng
oth
er m
on
ito
rin
g to
ols
(e.
g., N
agio
s) f
or
thei
r p
arti
cula
r n
eed
s
Fo
cus
on
Nag
ios:
info
rmat
ion
co
llec
ted
by
Nag
ios
is s
till
ava
ilab
le t
o V
O-c
entr
ic G
angl
iaid
ea: s
imu
late
a g
mo
nd
co
llec
tor
wit
h a
sim
ple
dae
mo
n p
roce
ss
Imp
rove
d W
eb I
nte
rfac
eSt
and
ard
inte
rfac
e m
od
ifie
d t
o h
and
le n
ew t
ypes
A s
eco
nd
vie
w t
ho
rugh
a s
imp
lifi
ed N
agio
s in
terf
ace
ach
ieve
d b
y ex
po
rtin
g in
tern
al d
ata
in N
agio
s fo
rmat
Pre
sen
tati
on
Ou
tlin
e1.
In
tro
du
ctio
n, G
oal
s &
Req
uir
emen
ts
2. E
xist
ing
Wo
rk O
verv
iew
3. A
cco
mp
lish
ed W
ork
a. D
ata
Su
mm
ati
on a
nd
Hie
rarc
hic
Rep
orti
ng
b. M
DS
Ad
dre
ss P
ub
lish
ing
c. L
oca
l Sen
sors
Sp
ecif
ica
tion
d. D
ata
Ret
riev
al f
rom
Na
gios
4. S
cree
nsh
ots
/ D
emo
in a
Sim
ula
ted
En
viro
nm
ent
5. N
agio
s W
eb I
nte
rfac
e In
corp
ora
tio
n
6. C
urr
ent
Stat
us
& C
on
clu
sio
ns
& R
emai
nin
g &
Fu
ture
Wo
rk
Use
d C
on
figu
rati
on
blu
e.c
olo
rsre
d.c
olo
rsm
agen
.co
lors
yell
ow
.co
lors
pu
rple
.co
lors
Fer
mi C
lust
er @
Uo
fCP
hys
Sci S
ite
@U
ofC
Clu
ster
@A
NL
Gri
d @
Ch
icag
o
Site
@U
ofC
gmo
nd
gmo
nd
gmo
nd
gmet
adgm
etad
gmm
etad
2gm
met
ad2
gmm
etad
2
Scre
ensh
ot:
Mai
n P
age
Scre
ensh
ot:
Mai
n P
age
(2)
Scre
ensh
ot:
Ove
rvie
w
Scre
ensh
ot:
Det
ails
Scre
ensh
ot:
Det
ails
(m
em)
Scre
ensh
ot:
Sen
sors
In
fo
Pre
sen
tati
on
Ou
tlin
e1.
In
tro
du
ctio
n, G
oal
s &
Req
uir
emen
ts
2. E
xist
ing
Wo
rk O
verv
iew
3. A
cco
mp
lish
ed W
ork
a. D
ata
Su
mm
ati
on a
nd
Hie
rarc
hic
Rep
orti
ng
b. M
DS
Ad
dre
ss P
ub
lish
ing
c. L
oca
l Sen
sors
Sp
ecif
ica
tion
d. D
ata
Ret
riev
al f
rom
Na
gios
4. S
cree
nsh
ots
/ D
emo
in a
Sim
ula
ted
En
viro
nm
ent
5. N
agio
s W
eb I
nte
rfac
e In
corp
ora
tio
n
6. C
urr
ent
Stat
us
& C
on
clu
sio
ns
& R
emai
nin
g &
Fu
ture
Wo
rk
Scre
ensh
ot:
Ho
st D
etai
ls
Scre
ensh
ot:
Ser
vice
s D
etai
ls
Scre
ensh
ot:
Sta
tus
Map
Use
d C
on
figu
rati
on
(2)
blu
e.co
lors
red
.co
lors
mag
en.c
olo
rsye
llo
w.c
olo
rsp
urp
le.c
olo
rs
Fer
mi C
lust
er @
Uo
fCP
hys
Sci S
ite
@U
ofC
Clu
ster
@A
NL
Gri
d @
Ch
icag
o
Site
@U
ofC
nag
ios
gmo
nd
gmo
nd
gmet
adgm
etad
exp
ort
ergm
met
ad2
gmm
etad
2
Inte
rfac
e w
ith
Nag
ios
Pre
sen
tati
on
Ou
tlin
e1.
In
tro
du
ctio
n, G
oal
s &
Req
uir
emen
ts
2. E
xist
ing
Wo
rk O
verv
iew
3. A
cco
mp
lish
ed W
ork
a. D
ata
Su
mm
ati
on a
nd
Hie
rarc
hic
Rep
orti
ng
b. M
DS
Ad
dre
ss P
ub
lish
ing
c. L
oca
l Sen
sors
Sp
ecif
ica
tion
d. D
ata
Ret
riev
al f
rom
Na
gios
4. S
cree
nsh
ots
/ D
emo
in a
Sim
ula
ted
En
viro
nm
ent
5. N
agio
s W
eb I
nte
rfac
e In
corp
ora
tio
n
6. C
urr
ent
Stat
us
& C
on
clu
sio
ns
& R
emai
nin
g &
Fu
ture
Wo
rk
Cu
rren
t St
atu
sE
very
thin
g is
dev
elo
ped
an
d t
este
d o
n m
y si
mu
late
d c
lust
er
Th
e so
ftw
are
is d
istr
ibu
ted
as
two
pac
kag
es a
vail
able
at:
htt
p:/
/peo
ple
.cs.
uch
icag
o.e
du
/~cl
du
mit
r/so
ft/m
on
i/v2
Pac
man
per
son
al c
ach
e at
:h
ttp
://p
eop
le.c
s.u
chic
ago
.ed
u/~
cld
um
itr/
pac
man
Inst
alat
ion
Ste
ps
Co
re (
5 st
eps)
:au
toco
nf
./co
nfi
gure
--p
refi
x=D
IRm
ake
mak
e in
stal
led
it b
y h
and
co
nfi
gura
tio
n f
iles
(in
stru
ctio
ns
insi
de
pac
kage
)
Intr
efac
e (i
ncl
ud
es b
oth
ext
end
ed a
nd
nag
ios
lik
e vi
ews
- 4
step
s):
auto
con
f./
con
figu
re -
-pre
fix=
DIR
2 --
wit
h-g
met
ad=
DIR
mak
em
ake
inst
all
Co
ncl
usi
on
s
Fu
ture
Wo
rkW
rite
oth
er s
enso
rs a
bo
ut
VO
res
ou
rce
usa
ges
Stat
isti
cs c
lass
ific
atio
n b
y:re
sou
rce
con
sum
pti
on
VO
s, g
rou
ps,
use
rsW
ork
Typ
ere
sou
rces
:h
ost
, clu
ster
, sit
e, a
nd
gri
d
Inte
grat
e w
ith
th
e ru
le-b
ased
po
licy
mo
du
le
top related