W hat

’saGr
aphDat
abaseandW hyShoul
dYouCar
e?

WeUseGr
aphsEv
eryDay
Youus egr aphdat abas esev er
ydayandpr obabl ydon’tk nowi t
.
Everytimey ouvisitLi nkedInands eef i
rst-,
second- ,ort hird-degree
connections ,you’r
eget ti
ngr esultsfrom t hesoci alnetwor ki
ngs it
e ’
s
prof
ess i
onal net
wor kgraphbui ltonagr aphdat abas e.Facebook ,
Inst
agram, andTwi tt erallusegraphdat abas esandanal yticsto
understandhowus er srel
atetoeachot herandconnectt hem wi th
theri
ghtcont ent.Ev eryt i
mey oudoaGoogl es earch, y
ou’ ret apping
i
ntoGoogl e ’
sKnowl edgeGr aph. Thos epr oductr ecommendat i
ons
onAmaz on- -“peopl ewhoboughtt hisi
t em als
obought …”or“ t
hes e
i
temsar eof t
enboughtt ogether”?Thatcomesf rom agr aph
analyt
icsquer ytoo.

Whati
saGr
aphDat
abase?
Graphanalyti
csisthetechnol
ogythatal
lowsf ort
hedeepexplor
ati
onofcompl exi
nter
rel
ati
onshipsamongv ar
iousenti
ti
es–organizations,
people,t
ransacti
ons,pr
ettymuchanydat athatcanbemodeledandquer i
ed.Gart
nerest
imatesthatt
hegraphdatabaseandanalyt
icsmar ket
wil
l gr
owat100per centannuall
ythr
ough2022, makingi
toneofthefastestgrowi
ngmarketsindataandanalyt
ics.
Accordi
ngtoGar tner,
“Graphanalyt
icswil
lgrowi nt
henextfewy ear
sduetotheneedtoas kcomplexquesti
onsacrosscomplexdata,whichi
snotal
way spr acti
calor
evenpos si
bl
eats cal
eus i
ngSQLquer i
es.
”

Rel
ati
onal
andNoSQLDat
abasesAr
en’
tUpt
otheTask
Relationaldatabas
esalsofall
shortbecaus etheirarchi
tectur
es implyisn’tdesi
gnedforthi
slev elofanalyt
ics.RDBsstorethedataforeach
busines sentit
ysuchascus t
omer ,order,product,andpay mentdat ainsepar at
edatabasetables .
Tounder standandanalyzerel
ati
onships
acros sthebus i
nessenti
ti
es,r
elat
ional databasesr equi
retablejoins,
whi chcant akehours,
evenday sforthecomplexjoinsandare
comput ati
onall
yexpensi
veasthes i
zeoft hedat agrows .
NoSQLdat abas ess t
oreall
ofthedatai nas i
ngletable.
Thismeanst hatt
he
rel
ations hi
panalysi
srequi
resscanningahuget ablewit
hmi l
l
ionsorbi l
l
ionsofr ows,mak i
ngitv er
ydiff
icul
ttoperfor
m adeeperanal ysi
sof
therelationshi
psbeyondtwoort hreelevels .

Lear
nmoreinourNati
veParal
lel
Gr aphseBookht
tps:/
/www.
ti
gergr
aph.com/ebook
anddownloadt
hefr
eeTiger
GraphDeveloperEdi
ti
onathttps:
//
www.t
iger
graph.com/downl
oad

sal
es@t
iger
graph.
com
 W hat
’saGr
aphDat
abaseandW hyShoul
dYouCar
e?

Ol
derGr
aphTechnol
ogyFal
l
sShor
t
Gr
aphdat abasesgobacktotheear
ly2000sandhavelongbeenseenasagreatwaytoans
werquest
ionsaboutcomplexr
elat
ions
hipsi
n
l
argedatasets.However
,foral
ongti
met heyst
ruggl
edtoperf
orm wel
lwhendatavol
umesgrewl
argeandwhent heanswersmustbe
pr
ovidedinrealt
ime.
Fi
rst-
gener
ati
ongraphdatabas
eswer ebuil
twi
thnat
iv
egr aphst
oragebutwerenotmadetohandl
elar
gedataorqueryv
olumesorper
for
m
beyondthr
eelevel
sorconnect
ions–k nownashops--i
nsidet
hegraph.Wit
heveryhopi
nagraph,t
hescopeofthesear
chexpands
dramati
cal
lyandthei
nsight
sgleanedbecomedeeper
.
Second-gener
ationgraphdat abaseswerebui
ltontopofNoSQLs tor
age,whichal
lowedt hem toloadl
argeamount sofdat
a.Howev er,
they
sti
l
ldonots caleforqueri
esinvolv
ingthr
eeormor ehops.Ol
dergr aphdatabasesals
ot ypical
lydonotsupport“
databaseshar
ding”–
part
it
ioni
ngofdat aacrossanumberofs er
verstoincr
easescalabil
it
y--whichmeans ,
al argegraphwit
hteraby
tesofdatacan’
tbedi st
ri
buted.
Theselegacygraphdat abasesareil
l-
equi
ppedtoscaleuptotoday ’
sreal
-worldr
equir
ement s
,whichcal
lforasyst
em thatcanperform many
hopseffi
cient
lyandinpar al
l
el t
odeliv
ersub-
secondqueryperformanceonbi gdata.

“
Wequi
ckl
yrani
ntoprobl
emsscali
ngwithourori
ginalgraphdatabase–loadi
ngthedat
atookal
otoft
ime
andonceitwasl
oaded,cal
cul
ati
onsei t
herdidn’
tfi
nishoritwasext
remel
ysl
ow.”
-NoelGomez,Dat
aSciencesLeader,Amgen
Wat
cht
het
est
imoni
alatht
tps:
//
ti
nyur
l.
com/
TGt
est
imoni
al
s

LeapFor
war
dwi
thaNat
ivePar
all
elGr
aphDat
abase-Ti
ger
Graph
Tiger
Graphisanewkindofgr
aphdatabase,
anativ
eparal
l
elgraphdatabasepur pose-
buil
tforl
oadi
ngmassi
veamountsofdat
a(t
eraby
tes
)
i
nhour sandanal
yzi
ngasmanyas10ormor ehopsdeepi
ntorelat
ions
hipsinreal-t
ime.Ti
gerGr
aphsuppor
tst
rans
act
ionalaswel
las
analy
ti
calworkl
oads
,isACIDcompl
iant,
scal
esupandoutwithdatabasesharding.
Ti
gerGr
aph’
sprov
entechnol
ogysupport
sappl
icat
ionssuchasfr
auddetecti
on,cus
tomer360,
IoT
,AI
andmachi
nel
ear
ningt
omak
esens
e
ofev
er-
changi
ngbigdat
a,andi
sus edbycust
omer si
ncl
udingVi
sa,I
ntuit
,ChinaMobi
le,
WishandZi
l
low.

Lear
nmoreinourNati
veParal
lel
Gr aphseBookht
tps:/
/www.
ti
gergr
aph.com/ebook
anddownloadt
hefr
eeTiger
GraphDeveloperEdi
ti
onathttps:
//
www.t
iger
graph.com/downl
oad

sal
es@t
iger
graph.
com