• No se han encontrado resultados

Entrevista y respuestas

In document Natalia Vélez Daniela Falla (página 76-79)

7!1 8b&ective 8b&ective measurementsmeasurements

7ach <'ery as r'n 10 t!es on each "aa5ase an" e8ec'ton t!es ere !eas're" in secon"s $s%. The 7ach <'ery as r'n 10 t!es on each "aa5ase an" e8ec'ton t!es ere !eas're" in secon"s $s%. The longes an" shores t!es ere "roppe" an" he re!aining eigh t!es ere aerage". The res'ls o he longes an" shores t!es ere "roppe" an" he re!aining eigh t!es ere aerage". The res'ls o he ess he!seles can 5e o'n" in #ppen"i8 K an" he !eas're" t!es can 5e o'n" in #ppen"i8 KI. No ess he!seles can 5e o'n" in #ppen"i8 K an" he !eas're" t!es can 5e o'n" in #ppen"i8 KI. No a:e!ps ere !a"e o opt!i9e he =aa KM, he +osgreSQL "aa5ase serice or he SQL <'eries. The a:e!ps ere !a"e o opt!i9e he =aa KM, he +osgreSQL "aa5ase serice or he SQL <'eries. The 5en

5ench!ch!ar(ar(s s or Neoor Neo; ; an" +osan" +osgregreSQL ill SQL ill 5e r'n 5e r'n 3o'3o'  o o he 5o83he 5o83, , iih h a a 3na3na'r'ral al synsyna8a83, an" 3, an" 3as3as "oc'!ene"3 or all ypes o <'eries.

"oc'!ene"3 or all ypes o <'eries.

Beore operatng on +osGIS or Neo;, each "aa5ase ha" a 3ar!ing 'p3 proce"'re r'n on i. In +osGIS, a Beore operatng on +osGIS or Neo;, each "aa5ase ha" a 3ar!ing 'p3 proce"'re r'n on i. In +osGIS, a 3S7L7/T b @-M geo!a5les3 as e8ec'e" an" all o he res'ls ere ierae" hro'gh. In Neo;, eery 3S7L7/T b @-M geo!a5les3 as e8ec'e" an" all o he res'ls ere ierae" hro'gh. In Neo;, eery ere8 in he graph as ierae" hro'gh an" he o'going e"ges o each ere8 ere reriee". The ess ere8 in he graph as ierae" hro'gh an" he o'going e"ges o each ere8 ere reriee". The ess ere r'n in he olloing se<'ence sar operatng syse! J sar "aa5ase J ar!ing 'p J r'n es J ere r'n in he olloing se<'ence sar operatng syse! J sar "aa5ase J ar!ing 'p J r'n es J sh'"on operatng syse!. This as "one o ens're ha any caching or syse! process actiy o'l" no sh'"on operatng syse!. This as "one o ens're ha any caching or syse! process actiy o'l" no a4ec he t!ings.

a4ec he t!ings.

7!1!1

7!1!1 Storing Storing 8S* 8S* datadata

@o'r -SM "aases are i!pore" in Neo;6Spatal an" +osGIS. Bac(gro'n" inor!aton a5o' he ac<'ire" @o'r -SM "aases are i!pore" in Neo;6Spatal an" +osGIS. Bac(gro'n" inor!aton a5o' he ac<'ire" "aases, he soare 'se" an" he pro;ecton chosen is "escri5e" in paragraphs

"aases, he soare 'se" an" he pro;ecton chosen is "escri5e" in paragraphs C.2.2C.2.2,, C.2.1C.2.1 an"an" D.2.1 D.2.1. The. The res'l

res'ls o he i!por is s'!!ari9es o he i!por is s'!!ari9e" in" in Ta5le 1F.Ta5le 1F. The i!por ool or +osGIS is opt!i9e" an" operaes The i!por ool or +osGIS is opt!i9e" an" operaes <'ic(ly he res'ltng sorage space is accepa5le. The i!porer or Neo; is less opt!i9e", less as an" he <'ic(ly he res'ltng sorage space is accepa5le. The i!porer or Neo; is less opt!i9e", less as an" he t!e o i!por -SM "aa a(es longer as co!p'er poer is no eRcienly 'tli9e".

t!e o i!por -SM "aa a(es longer as co!p'er poer is no eRcienly 'tli9e". Table 1<: Resuls i!portng B$ /le

Table 1<: Resuls i!portng B$ /le "

"aammee 22$$ SSyysstteemm 99iimmee 99ooppoollooggyy 6S2->le

6S2->le "eo)*"eo)* Post+,SPost+,S "eo)*"eo)* Post+,SPost+,S "eo)*"eo)* Post+,SPost+,S

7nty

7nty $MB%$MB% $MB%$MB% $MB%$MB% $s%$s% $s%$s% $s%$s% $s%$s%

Me"e!5li(

Me"e!5li(   1D1D F,F, ,0,0 C,EDC,ED 1,012D1,012D

#!ser"a!

#!ser"a! 10H10H 21F21F 12F12F H00,DH00,D E,FE,F CD1,12DCD1,12D 2F,ED2F,ED

Norh6ollan"

Norh6ollan" 122F122F 2CCE2CCE 112112 1CHD,D1CHD,D HHC,1HHC,1 2,D2,D 2E1F0,CED2E1F0,CED

Neherlan"s

Neherlan"s 10HF110HF1    1DCH01DCH0    20C,H20C,H      

I!por

I!portng he Neherlantng he Neherlan"s "aase ih Neo; oo( !ore "s "aase ih Neo; oo( !ore han 1 han 1 ho'rs an" res'leho'rs an" res'le" in " in a "aa5ase haa "aa5ase ha gre o !ore han H0

gre o !ore han H0 giga5ygiga5ye $GB%. &ih +osGIS, i!pore $GB%. &ih +osGIS, i!portng oo( appro8i!tng oo( appro8i!aely  ho'rs, res'ltaely  ho'rs, res'ltng in ng in aa "aa5ase si9e o 1D GB. The si9e o he Norh6ollan" "aase see!s o 5e he li!i or Neo;, a leas "aa5ase si9e o 1D GB. The si9e o he Norh6ollan" "aase see!s o 5e he li!i or Neo;, a leas r'nning in his har"are eniron!en. )sing larger "aases, a (non pro5le! ih he L'cene in"e8 as r'nning in his har"are eniron!en. )sing larger "aases, a (non pro5le! ih he L'cene in"e8 as enco'nere" as "isc'sse" on heir or'!s $Neo; /o!!'niy 6 More spatal <'estons 2011%. The secon" enco'nere" as "isc'sse" on heir or'!s $Neo; /o!!'niy 6 More spatal <'estons 2011%. The secon" sage o he i!por consiss o rein"e8ing he "aa5ase an" 5'il"ing a L'cene in"e8. This process slos sage o he i!por consiss o rein"e8ing he "aa5ase an" 5'il"ing a L'cene in"e8. This process slos "on e8ponentally or larger "aases, a4ectng he scala5iliy o he i!porer.

"on e8ponentally or larger "aases, a4ectng he scala5iliy o he i!porer.

D1 D1

NoSQL spatal NoSQL spatal

+aragraph

+aragraph H.1.H.1. $page$page  D D% ill "isc'ss in so!e !ore "eail he correlaton 5eeen he -SM fle an" he% ill "isc'ss in so!e !ore "eail he correlaton 5eeen he -SM fle an" he har"are nee"e". Beca'se he i!por oo( so long an" creae" an 'naccepa5ly si9e" "aa5ase, he har"are nee"e". Beca'se he i!por oo( so long an" creae" an 'naccepa5ly si9e" "aa5ase, he Neherlan"s "aase as no incl'"e" in he oher ess. &hen -SM fle si9es re!ain s!aller han 1 GB, Neherlan"s "aase as no incl'"e" in he oher ess. &hen -SM fle si9es re!ain s!aller han 1 GB, 5oh "aa5ases 'ncton accepa5ly as o si9e an" spee". +osGIS has clear a"anages i!portng larger 5oh "aa5ases 'ncton accepa5ly as o si9e an" spee". +osGIS has clear a"anages i!portng larger -SM fles he i!por t!e is shor an" he "aa5ase si9e is s!aller.

-SM fles he i!por t!e is shor an" he "aa5ase si9e is s!aller.

7!1!#

7!1!# Creating Creating a a routing routing network network topolog+topolog+

#s e8plaine" in paragraph

#s e8plaine" in paragraph D.D.  $page  $page 22%, a ro'tng opology nee"s o 5e creae" or 5oh "aa5ases. The%, a ro'tng opology nee"s o 5e creae" or 5oh "aa5ases. The res'l is he opposie o soring -SM "aa Neo; is aser in creatng he sa!e ro'e neor( han +osGIS. res'l is he opposie o soring -SM "aa Neo; is aser in creatng he sa!e ro'e neor( han +osGIS. #gain, he "i4erences are larges 'sing he Norh6ollan" "aase, an" he Neherlan"s "aase is no #gain, he "i4erences are larges 'sing he Norh6ollan" "aase, an" he Neherlan"s "aase is no incl'"e" in he !eas're!ens.

incl'"e" in he !eas're!ens. #s shon in

#s shon in Ta5le 1FTa5le 1F, he s!alles "aase ro! Me"e!5li( shos an a"anage or +osGIS. @or larger, he s!alles "aase ro! Me"e!5li( shos an a"anage or +osGIS. @or larger "aases, Neo; 5enefs ro! aser ro'e neor( creatng t!es.

"aases, Neo; 5enefs ro! aser ro'e neor( creatng t!es. @ig're 22@ig're 22 is'ali9es he t!e in secon"sis'ali9es he t!e in secon"s in a col'!n char.

in a col'!n char.

#pparenly he graph "aa5ase nee"s less t!e o fn" he correspon"ing sar no"es an" en" no"es. The #pparenly he graph "aa5ase nee"s less t!e o fn" he correspon"ing sar no"es an" en" no"es. The "i4erences are o'n" in he sage ha fn"s he correspon"ing no"es. #s e8plaine" in paragraph

"i4erences are o'n" in he sage ha fn"s he correspon"ing no"es. #s e8plaine" in paragraph D.D. $page $page 2

2%, he pgo'tng SQL proce"'re%, he pgo'tng SQL proce"'re assignverteidassignverteid flls in he so'rce an" arge col'!ns o he neor( flls in he so'rce an" arge col'!ns o he neor( a5le in s'ch a ay ha connece" geo!eries share ere8 i"s. This sage o he SQL scrip is hrea"6 a5le in s'ch a ay ha connece" geo!eries share ere8 i"s. This sage o he SQL scrip is hrea"6 inensie an" !a(es 'p F0 o processing t!e.

inensie an" !a(es 'p F0 o processing t!e.

D2 D2

Figure #1: easure!ens I i!portng B$ daa Figure #1: easure!ens I i!portng B$ daa

Figure ##: easure!ens I creatng a roue opology  Figure ##: easure!ens I creatng a roue opology 

Me

Medemdemblik blik AmstAmsterderdam am NoNorthrth-H-Hollaolland nd NetNetherherlalandsnds 0 0 2000 2000 4000 4000 6000 6000 8000 8000 10000 10000 12000 12000 14000 14000 16000 16000 Datastore size (MB) Datastore size (MB) !M-"ile !M-"ile Neo4# Neo4# $ost%&! $ost%&! M

Mededememblb li k ik A mA ms ts tere rdadam m NoNor tr th-h -HoHol ll lanandd 0'0 0'0 2000'0 2000'0 4000'0 4000'0 6000'0 6000'0 8000'0 8000'0 10000'0 10000'0 12000'0 12000'0 14000'0 14000'0 &mort time (s) &mort time (s) Neo4# (s) Neo4# (s) $ost%&! (s) $ost%&! (s) M

Meeddeemmbblli k i k AAmmsstteerrddaam m NNoorrtthh--HHoollllaanndd 0 0 000 000 10000 10000 1000 1000 20000 20000 2000 2000 *0000 *0000 +

+oolo, .reation oolo, .reation time (s)time (s)

Neo4# Neo4# $ost%&! $ost%&!

7!1!)

7!1!) Spatial Spatial operationsoperations

The aera

The aerage" res'lge" res'ls s o o he spatahe spatal l opeoperatratons ons are s'!!aare s'!!ari9ri9e" e" in in he a5le he a5le 5e5elolo. . a a resres'l'ls s areare presene" in #ppen"i8 KI. The olloing sectons ill "isc'ss he res'ls o eery partc'lar es in "eail. presene" in #ppen"i8 KI. The olloing sectons ill "isc'ss he res'ls o eery partc'lar es in "eail. Table #): Resuls spatal operatons

Table #): Resuls spatal operatons "

"aammee BBmmppttyy ccaallll CCoo ccoouunntt C/ C/

C

Coo ggmml l ++// ;;lloosseesstt ppooiinnt t ;;// SShhoorrtteesstt ppaatthhPP//

"eo)*

"eo)* Post+,SPost+,S "eo)*"eo)* Post+,SPost+,S "eo)*"eo)* Post+,SPost+,S "eo)*"eo)* Post+,SPost+,S "eo)*"eo)* Post+,SPost+,S

7nty

7nty $s%$s% $s%$s% $s%$s% $s%$s% $s%$s% $s%$s% $s%$s% $s%$s% $s%$s% $s%$s%

Me"e!5li(

Me"e!5li( 0,020,02 0,0C0,0C 0,FC0,FC 0,00,0 C,1CC,1C 0,DH0,DH 0,120,12 0,00,0 0,10,1 0,ED0,ED

#!ser"a!

#!ser"a! 0,020,02 0,0C0,0C 1,ED1,ED 1,011,01 111,0ED111,0ED 10,E010,E0 0,DC0,DC 0,D10,D1 1,21,2 C,DC,D

Norh6ollan"

Norh6ollan" 0,020,02 0,0C0,0C DE,1CDE,1C 2,HF2,HF E01,EDE01,ED 1,FH1,FH 2,012,01 0,FC0,FC D,1D,1 2F,2F,

Spa%al ounding o opera%ons Spa%al ounding o opera%ons

These operatons are "efne" o fn" all he geo!eries ihin a specifc recangle area. Tes B e8ec'es a These operatons are "efne" o fn" all he geo!eries ihin a specifc recangle area. Tes B e8ec'es a co'n operaton ihin a recangle area hile es G re<'e

co'n operaton ihin a recangle area hile es G re<'ess he "aa in GML or!a. The applie" co"e isss he "aa in GML or!a. The applie" co"e is sae" in paragraph

sae" in paragraph  D.D.2 D.D.2, na!e", na!e"   Tes B J 5o'n"ing 5o8 co'n  Tes B J 5o'n"ing 5o8 co'n, an" paragraph, an" paragraph  D.D.C D.D.C, na!e", na!e"  Tes G J Tes G J 5o'n"ing 5o8 ge

5o'n"ing 5o8 ge. )n"o'5e"ly, +osGIS o'peror!s he graph "aa5ase ih 5o'n"ing 5o8 operatons. )n"o'5e"ly, +osGIS o'peror!s he graph "aa5ase ih 5o'n"ing 5o8 operatons he larger he "aase, he larger he "i4erences in operaton t!es. This es "e!onsraes he eRciency o he larger he "aase, he larger he "i4erences in operaton t!es. This es "e!onsraes he eRciency o he

he relrelatoatonal nal "a"aa5aa5ase. se. # # <'e<'ery ry on on a a s!as!alleller r "a"aasease  $Me$Me"e!"e!5li5li( ( or or #!s#!serer"a!"a!% % has has co!co!parpara5la5lee operaton t!es o he sa!e <'ery on

operaton t!es o he sa!e <'ery on he larger Norh6ollan" "aase.he larger Norh6ollan" "aase. Neo

Neo; ; conconsissisenenly ly perperor!or!s s slosloer er 's'sing ing he sa!e he sa!e 5o'5o'n"in"ing ng 5o85o8es. es. ooeeer, er, in in NeNeo;o;, , hehen n aa s'5se<'en 5o'n"ing 5o8 operaton is e8ec'e" aer he inital r'n, he operaton 5eco!es !'ch aser. s'5se<'en 5o'n"ing 5o8 operaton is e8ec'e" aer he inital r'n, he operaton 5eco!es !'ch aser. @or e8a!ple, he frs6calle" 5o'n"ing 5o8 es in he #!ser"a! area a(es appro8i!aely 1D secon"s. @or e8a!ple, he frs6calle" 5o'n"ing 5o8 es in he #!ser"a! area a(es appro8i!aely 1D secon"s. 'nning he sa!e es again ill a(e a5o'  secon"s. This opic ill 5e "isc'sse" o so!e e8en laer in 'nning he sa!e es again ill a(e a5o'  secon"s. This opic ill 5e "isc'sse" o so!e e8en laer in his paragraph.

his paragraph.

The "i4erences in 5o'n"ary t!es are noeorhy. Boh "aa5ases 'se an 6re

The "i4erences in 5o'n"ary t!es are noeorhy. Boh "aa5ases 'se an 6ree in"e8 or aser geo!erye in"e8 or aser geo!ery selecton. So!e "aa5ase opt!i9atons ill res'l in aser operaton t!e, 5' he oerall res'ls sho'l" 5e selecton. So!e "aa5ase opt!i9atons ill res'l in aser operaton t!e, 5' he oerall res'ls sho'l" 5e ro'ghly he sa!e 5eca'se o he 6ree in"e8. The large "i4erences in <'ery t!e are "'e o he ay in ro'ghly he sa!e 5eca'se o he 6ree in"e8. The large "i4erences in <'ery t!e are "'e o he ay in hich Neo; 'ses <'eries. L'cene is he in"e8 'se" or <'erying, hich as originally "esigne" as a or" hich Neo; 'ses <'eries. L'cene is he in"e8 'se" or <'erying, hich as originally "esigne" as a or" "ens

"ensiy search engine li5rary. &hile neer ersioniy search engine li5rary. &hile neer ersions "o s "o allo n'!erallo n'!eric <'eries, 'n"er he hoo" ic <'eries, 'n"er he hoo" all "aa all "aa isis reae" as e8. 78ec'tng a <'ery on non6ineger n'!eric "aa, s'ch as "o'5les, res'ls in een orse reae" as e8. 78ec'tng a <'ery on non6ineger n'!eric "aa, s'ch as "o'5les, res'ls in een orse peror!ance or Neo;. The spee" iss'es concerning searching or n'!5ers in Neo;3s in"e8 are relae" o peror!ance or Neo;. The spee" iss'es concerning searching or n'!5ers in Neo;3s in"e8 are relae" o L'cene since conersions !'s 5e peror!e". This is a pro5le! (non o Neo;3s "eelopers an in"e8 L'cene since conersions !'s 5e peror!e". This is a pro5le! (non o Neo;3s "eelopers an in"e8 "e"icae" o n'!eric al'es is

"e"icae" o n'!eric al'es is s'ppose" o res'l in aser 5o'n"ary s'ppose" o res'l in aser 5o'n"ary operatons.operatons.

&hen 5o'n"ary <'eries are he 'se case or an applicaton, +osGIS clearly proi"es a !ore opt!al &hen 5o'n"ary <'eries are he 'se case or an applicaton, +osGIS clearly proi"es a !ore opt!al sol'ton.

sol'ton.

;losest point opera%ons ;losest point opera%ons The olloi

The olloing es is ng es is "esig"esigne" o ne" o fnfn" a " a no"e in he "aa5ase hich is closes o he no"e in he "aa5ase hich is closes o he coor"icoor"inae proi"nae proi"e",e", an" o calc'lae he "isance ro! no"e o coor"inae. The applie" co"e is e8plaine" in paragraph

an" o calc'lae he "isance ro! no"e o coor"inae. The applie" co"e is e8plaine" in paragraph D.D.D.D. an"an" na!e"

na!e" Tes / J closes no"e Tes / J closes no"e. +osGIS shos so!e oerhea" t!e in he Me"e!5li( "aase in general, he. +osGIS shos so!e oerhea" t!e in he Me"e!5li( "aase in general, he operaton o fn" hese no"es yiel"s c

operaton o fn" hese no"es yiel"s co!para5le res'ls 5eeen he o syse!s.o!para5le res'ls 5eeen he o syse!s. Shortest path opera%ons

Shortest path opera%ons

Ne8 are he shores pah operatons 5ase" on *i;(sra $1FDF%. The applie" co"e is sae" in paragraph Ne8 are he shores pah operatons 5ase" on *i;(sra $1FDF%. The applie" co"e is sae" in paragraph

DC DC

NoSQL spatal NoSQL spatal D.D.D

D.D.D an" na!e" an" na!e" Tes + J shores pahTes + J shores pah . @or hese <'eries, Neo; as clearly aser, appro8i!aely 5y a. @or hese <'eries, Neo; as clearly aser, appro8i!aely 5y a ac

acor o or o appappro8ro8i!ai!aelely y C, as C, as "e"eailaile" ine" in Ta5Ta5le le 2020 an" #ppen"i8 KI. # <'ery on a s!aller "aasean" #ppen"i8 KI. # <'ery on a s!aller "aase $Me"e!5li( or #!ser"a!% has co!para5le operaton t!es o he sa!e <'ery on he larger Norh6 $Me"e!5li( or #!ser"a!% has co!para5le operaton t!es o he sa!e <'ery on he larger Norh6 ollan" "aase. Shores pah <'eries are he ypical raersals Neo; is opt!i9e" or, hereas relatonal ollan" "aase. Shores pah <'eries are he ypical raersals Neo; is opt!i9e" or, hereas relatonal "aa5ases are no "esigne" o "o raersals. Beca'se graph "aa5ases !ainain "irec reerences 5eeen "aa5ases are no "esigne" o "o raersals. Beca'se graph "aa5ases !ainain "irec reerences 5eeen relae" "aa, hey are !os eRcien hen peror!ing his ype o local "aa analysis. Generally, relatonal relae" "aa, hey are !os eRcien hen peror!ing his ype o local "aa analysis. Generally, relatonal "aa5ases 'se ;oins in or"er o !oe 5eeen a5les ha are lin(e" 5y cerain col'!ns. Traersing a "aa5ases 'se ;oins in or"er o !oe 5eeen a5les ha are lin(e" 5y cerain col'!ns. Traersing a relatonal "aa5ase !'s 5e inerre" hro'gh a series o ;oin operatons on "i4eren a5les. &here a s'5se relatonal "aa5ase !'s 5e inerre" hro'gh a series o ;oin operatons on "i4eren a5les. &here a s'5se o he "aa !ay 5e "esire", ;oin operatons re<'ire all "aa in all <'erie" a5les o 5e e8a!ine" in or"er o o he "aa !ay 5e "esire", ;oin operatons re<'ire all "aa in all <'erie" a5les o 5e e8a!ine" in or"er o e8rac he "esire" s'5se. This li!iaton is sole" 5y he "aa sr'c're pgo'tng 'ses. The neor( ha e8rac he "esire" s'5se. This li!iaton is sole" 5y he "aa sr'c're pgo'tng 'ses. The neor( ha 'se" consiss o one a5le conaining all neor( "aa. The res'ls o pgo'tng are ery goo", 5' "o no 'se" consiss o one a5le conaining all neor( "aa. The res'ls o pgo'tng are ery goo", 5' "o no co!e near Neo;3s. This can 5e e8plaine" 5y he ac ha pgo'tng nee"s o peror!s loo('ps o he ne8 co!e near Neo;3s. This can 5e e8plaine" 5y he ac ha pgo'tng nee"s o peror!s loo('ps o he ne8 possi5le no"e in he neor(, hile raersing an e"ge in he graph "aa5ase "oes no re<'ire loo('ps a possi5le no"e in he neor(, hile raersing an e"ge in he graph "aa5ase "oes no re<'ire loo('ps a

In document Natalia Vélez Daniela Falla (página 76-79)

Documento similar