Sie sind auf Seite 1von 36

Designing EIM Interfaces from a Performance Perspective

DRAFT 1 Published by BT Retail T&T

Siebel EIM Best Practice

IN CONFIDENCE

Siebel EIM Best Practice IN CONFIDENCE

About this document...

Author
The auth r ! this d cu"e#t "ay be c #tacted at$

Paul Du#ca# %ME&'( Ale)a#der Bai# & use *#d Fl r %++ *DA,( 1- . r/ Street 01AS0O2 0* 31A Tel$ M bile$ E"ail$ 41'1 **4 354' 45676 *1- -'1 Paul878Du#ca#9bt8c "

Distribution List
BT Siebel EIM I#ter!ace De:el +"e#t Tea"s$ &i;h 1e:el Desi;#ers 1 < 1e:el Desi;#ers De:el +"e#t Tea"s BT Siebel Per! r"a#ce Testi#; a#d Ma#a;e"e#t

Intended Audience
This d cu"e#t c #tai#s detailed tech#ical i#! r"ati # relati#; t Siebel E#ter+rise I#te;rati # Ma#a;er %EIM( a#d Oracle8 A le:el ! u#dersta#di#; ! these tech# l ;ies is assu"ed8

Filing
The !ili#; re!ere#ce ! r this d cu"e#t is$ TBD

Pa;e * ! 7= IN CONFIDENCE

Siebel EIM Best Practice IN CONFIDENCE

History
Issue 481 Date 43>A?0>*44' Author Paul Du#ca# Reason I#itial Creati #

Pa;e 7 ! 7= IN CONFIDENCE

Siebel EIM Best Practice IN CONFIDENCE

Contents
1 OVERVIEW.......................................................................................................................................................................5 1.1 PURPOSE........................................................................................................................................................................5 1.2 WHATS NOT INCLUDED................................................................................................................................................5 1.3 DOCUMENT CONVENTIONS...........................................................................................................................................6 2 EIM BEST PRACTICE....................................................................................................................................................7 2.1 INTRODUCTION..............................................................................................................................................................7 2.2 EIM CONFIGURATION...................................................................................................................................................7 2.2.1 Overview................................................................................................................................................................7 2.2.2 IFB Generic Parameters.......................................................................................................................................9
2.2.2.1 Only !"# T!$l#".............................................................................................................................................................% 2.2.2.2 I&n'(# !"# T!$l#"..........................................................................................................................................................% 2.2.2.3 I&n'(# !"# C'l)*n".....................................................................................................................................................1+ 2.2.2., D#-!)l. C'l)*n"............................................................................................................................................................1+ 2.2.2.5 F/0#1 C'l)*n"...............................................................................................................................................................1+ 2.2.2.6 U"# In1#0 H/n."2 U"# E""#n./!l In1#0 H/n."...................................................................................................................11 2.2.2.7 S#""/'n S3L..................................................................................................................................................................12 2.2.2.4 U"/n& Syn'ny*"............................................................................................................................................................ 12 2.2.2.% R'll$!56 On E(('(2 C'**/. E!57 P!""2 C'**/. E!57 T!$l#.........................................................................................13 2.2.2.1+ U81!.# A-.#( F!/l#1 In"#(.............................................................................................................................................1,

2.2.3 IFB Import Process Specific Parameters............................................................................................................15


2.2.3.1 Only !"# C'l)*n"....................................................................................................................................................... 15 2.2.3.2 In"#(. R'9"2 U81!.# R'9".............................................................................................................................................15 2.2.3.3 N#. C7!n&#....................................................................................................................................................................16 2.2.3., O(!5l# In"#(. A88#n1 M'1#...........................................................................................................................................17 2.2.3.5 T(/* S8!5#"...................................................................................................................................................................14

2.2.4 EI

!e"ete Process #onfi$%ration ....................................................................................................................19

2.2.,.1 D#l#.# All R'9"............................................................................................................................................................. 1% 2.2.,.2 D#l#.# E0!5.................................................................................................................................................................... 1% 2.2.,.3 D#l#.# M!.57#".............................................................................................................................................................. 2+

2.2.5 EI er$e Process #onfi$%ration.....................................................................................................................22 2.2.& Ot'er Parameters................................................................................................................................................24 3 RUNNING EIM...............................................................................................................................................................26 3.1 DATA...........................................................................................................................................................................26 3.2 IF FILE AND EIM PROCESSES...................................................................................................................................27 3.3 SCHEDULING................................................................................................................................................................24 3.3.1 !e$ree of Para""e"ism()a*o%t.............................................................................................................................2+ 3.3.2 Batc'si,e..............................................................................................................................................................2+ 3.3.3 EI Streamin$ #onfi$%rations...........................................................................................................................2+ 3.3.4 #ontro"- Sc'e.%"in$.........................................................................................................................................29 3., EIM TA LE MAINTENANCE........................................................................................................................................31 4 VOLUME 4 LESSONS LEARNED...............................................................................................................................32 ,.1 ENHANCEMENTS TO EIM PROCESSES.........................................................................................................................32 4.1.1 /e.%n.ant 0p.ates.............................................................................................................................................32 4.1.2 Sp"ittin$ 0p.ates.................................................................................................................................................33 4.1.3 1ints an. Optimiser o.e..................................................................................................................................35 ,.2 DATA ASE CONFIGURATION.......................................................................................................................................35 4.2.1 In.e2es.................................................................................................................................................................35 ,.3 ODM ADAPTER FUNCTIONALIT:................................................................................................................................36

Pa;e ' ! 7= IN CONFIDENCE

Siebel EIM Best Practice IN CONFIDENCE

!vervie"
1.1 Purpose
This d cu"e#t is i#te#ded t be used as a ;uide ! r the desi;# a#d de+l y"e#t ! EIM i#ter!aces t su++ rt the deli:ery ! Siebel based +r @ects at BT8 It <ill be ! +articular rele:a#ce t the deli:ery ! O#eSiebel %a/a Acc u#t Ma#a;ed Plat! r"A AMP( a#d O#eBie< %a/a B lu"e Siebel Release =(8 This d cu"e#t is s+lit i#t t< "ai# secti #s as ! ll <s$ 18 Siebel EIM Best Practice This secti # details tech#iCues t be ad +ted t e#sure that EIM i#ter!aces a#d their c #!i;urati #s adhere t Siebel a#d BT Best Practices8 *8 O#eBie< Release ' data l ad 1ess #s 1ear#ed This secti # details the /ey less #s lear#ed !r " the O#eBie< %a/a B lu"e SiebelA Bi; DataA CCC( bul/ data l ad ! Billi#; Acc u#tsA Assets a#d C #su"er ?+dates i#t the O#eBie< Siebel i#sta#ce8 This d cu"e#t has bee# asse"bled !r " e)+erie#ce !r " BTDs eB l:e a#d B lu"e +lat! r"s8 These a++licati #s are !ed data :ia ODM c #tr lled data l ads8 2here +r cesses ther tha# ODM <ill dri:e EIM i#ter!aces %e8;8 C #:er;e t O#eSiebel "i;rati # @ bs( the# s "e ODM related c #te#t <ill still be rele:a#t a#d s all secti #s sh uld be read8

1.2 Whats not included


This d cu"e#t is i#te#ded t share Best Practice a#d 1ess #s 1ear#ed a#d is #either a &i;h 1e:el # r a 1 < 1e:el Desi;#8 This d cu"e#t sh uld be used t !eed i#t the ab :e "e#ti #ed desi;#s8 Details ! Oracle>1e:el tu#i#; is # t i#cluded alth u;h s "e tech#iCues "ay be "e#ti #ed i# the c #te)t ! ther EIM s+eci!ic +ara"eters r c #!i;urati #s8 This is # t a EO#e St + Sh +D !r " <hich t lear# EIM8 I#steadA this d cu"e#t is i#te#ded t c ""u#icate EIM Best Practice8

Pa;e - ! 7= IN CONFIDENCE

Siebel EIM Best Practice IN CONFIDENCE

1.3 Document Conventions


Fey + i#ts <ithi# this d cu"e#t ha:e bee# i#cluded i# call> ut b )es as ! ll <s$ This i#dicates a c #:e#ti # that sh uld be ! ll <ed by De:el +"e#t Tea"s <ithi# BT8 A <ar#i#; si;# i#dicates that s+eci!ic care sh uld be ta/e#A that a +iece ! < r/ is reCuired t deter"i#e the best c #!i;urati # r that a c #:e#ti # r +ara"eter sh uld # t be ! ll <ed8 A# i#! r"ati # sy"b l is used t i#dicate additi #al + i#ts ! i#terestA # t #ecessarily rele:a#t t s+eci!ic i"+le"e#tati #s at BT8 These D cu"e#t C #:e#ti #s a++ear ! r " st /ey + i#t %i# the EIM Best Practice secti #( a#d s the d cu"e#tDs rec ""e#dati #s c uld be su""arised by readi#; @ust the illustrated + i#ts8

Pa;e = ! 7= IN CONFIDENCE

Siebel EIM Best Practice IN CONFIDENCE

# EIM $est Practice


2.1 Introduction
SiebelDs E#ter+rise I#te;rati # Ma#a;er %EIM( +erati # is deter"i#ed by a c #!i;urati # !ile8 This !ile is called the IFB !ile a#dA c u+led <ith l adi#; data i#t a Siebel EIMGI#ter!ace table %a sta;i#; table that Siebel uses s that it ca# build the l ;ical data relati #shi+s as de!i#ed <ithi# SiebelDs Re+ sit ry( deter"i#es <hat +erati #s EIM carries ut8 As a ;e#ericA ut> !>the>b ) i#ter!ace "echa#is"A EIMDs +er! r"a#ce ca# be strea"li#ed by reduci#; the a" u#t ! < r/ it has t d 8 This is d #e by i"+ si#; strict IFB ;uideli#es8 N t all +ara"eters are discussed i# detail8 A si#;le I#ter!ace Table "a+s t "ulti+le Base Tables %t cl sely ! ll < the #ature ! Siebel Busi#ess Ob@ects a#d Busi#ess C "+ #e#ts(8 All tables a#d related ! rei;# /eys are +r cessed by de!ault u#less +ara"eters are s+eci!ied i# the i!b !ile8 This d cu"e#t describes h < t alter EIMDs de!ault +erati # by s+eci!yi#; +ara"eters i# the i!b !ile8 S "e ! these cha#;es "a/e !u#cti #al i"+acts %i8e8 cha#;e the <ay i# <hich how EIM +r cesses the data( <hereas thers ca# i"+r :e +er! r"a#ce %# r"ally by li"iti#; the a" u#t ! work EIM a#d the u#derlyi#; database has t d (8

2.2 EIM Con i!uration


#%#% !vervie"
EIM su++ rts the ! ll <i#; ' " des ! +erati #$ o o o o I"+ rt Delete Mer;e E)+ rt

The EIM I"+ rt @ b is actually a# ?+sert +erati # a#d s this d cu"e#t <ill re!er t b th I"+ rt %I#sert( a#d ?+date EIM +erati #s8 There are a #u"ber ! ;e#eric i!b +ara"eters that are :alid ! r all " des ! +erati # as <ell as th se that are +erati # s+eci!ic8 This d cu"e#t <ill e)+l re I"+ rtA Delete a#d Mer;e @ bs i# " re detail8 EIM E)+ rts <ill # t be c :eredA as they are # t %;e#erally( used <ithi# BT i#steadA P1GS,1 % !te# c #tr lled by Ab I#iti ;ra+hs( is used t Cuery the Siebel data tables directly i# rder t e)+ rt i#! r"ati #8

Pa;e 5 ! 7= IN CONFIDENCE

Siebel EIM Best Practice IN CONFIDENCE S that s "e ! the ter"i# l ;y <ithi# this d cu"e#t is clearA hereDs a# e)a"+le IFB !ile %H i#dicates li#es re" :ed ! r clarity($ [Siebel Interface Manager] LOG TRANSACTIONS TO FILE = FALSE USING SYNONYMS = FALSE header secti #

[U!"ate#$ol%&e#'ill#Accnt] TY(E = IM(ORT 'ATC) = *'atc+Range TA'LE = EIM#ACCOUNT ONLY 'ASE TA'LES = S#(ARTY, S#ORG#E-T INSERT RO.S = S#(ARTY, FALSE U(/ATE RO.S = S#(ARTY, FALSE INSERT RO.S = S#ORG#E-T, FALSE NET C)ANGE = FALSE ONLY 'ASE COLUMNS = S#(ARTY0(ARTY#UI/, 1 S#(ARTY0(ARTY#TY(E#C/, 1 S#ORG#E-T0NAME, 1 S#ORG#E-T0LOC, 1 S#ORG#E-T0'U#I/, 1 N te$ a bac/slash %EID( sh uld be used t esca+e #e< li#es ! r readability8

Pr cess secti #

Ma#y +ara"eters ca# be s+eci!ied i# the &eader secti #A the Pr cess secti # r b th8 The rules ! i#herita#ce are as ! ll <s$ I! a +ara"eter is set i# the &eader secti #A it is i#cluded i# all Pr cess secti #s8 I! a +ara"eter is used i# a Shell +r cessA it is i#cluded i# all Shell Sub +r cesses %detailed later(8 I! a +ara"eter is i# the &eader secti # a#d als i# the Pr cess secti #A the# the :alue i# the Pr cess secti # :er>rides that ! the &eader secti #8 The sa"e is true ! r Shell +r cess a#d Shell Sub +r cess8 I! a +ara"eter e)ists #ly i# a Pr cess secti # a#d # t i# the &eader secti #A the# it is #ly used i# that Pr cess8

Pa;e 3 ! 7= IN CONFIDENCE

Siebel EIM Best Practice IN CONFIDENCE

#%#%# IF$ &eneric Parameters


The +ara"eters here a++ly t all EIM @ b ty+esA <ith +r cess s+eci!ic %i8e8 I"+ rtA Delete r Mer;e( +ara"eters a#d tech#iCues bei#; s+eci!ied i# subseCue#t secti #s8 #%#%#% !nly $ase 'ables This is a Pr cess secti # +ara"eter8 ?sa;e$ ONLY 'ASE TA'LES = 'ASE#TA'LE2, 'ASE#TA'LE3 H<here BASEJTAB1E1 & BASEJTAB1E* are tables "a++ed t by the EIM table i# Cuesti #8 By #ly i#cludi#; the Base Table relati #shi+s that y u #eed t res l:eA EIMDs < r/ is reduced a#d s Per! r"a#ce ca# be i"+r :ed8 E)a"+le$ I# O#eBie<A a C #su"er ?+date @ b i#c rrectly carried ut u+dates t SJPART.A SJOR0JEKTA SJACCNTJPOSTN a#d SJOR0JB? des+ite the !act that the data i# SJACCNTJPOSTN a#d SJOR0JB? <as bei#; de!aulted by the C #su"er Ada+ter %Ab I#iti ( each ti"e8 By re" :i#; these tables !r " the ON1. BASE TAB1ES +ara"eterA the +er! r"a#ce ! the B lu"e Siebel Release ' C #su"er ?+date @ b <as ;reatly i"+r :ed8 S that +er! r"a#ce ;ai#s such as these ca# be "adeA detailed !u#cti #al a#alysis ! the EIM i#ter!ace "ust be carried ut i# rder t !ully u#dersta#d the data s urce8

?sa;e$

This +ara"eter sh uld be used ON1. BASE TAB1ES

#%#%#%# Ignore $ase 'ables This is a Pr cess secti # +ara"eter8

INGORE 'ASE TA'LES = 'ASE#TA'LE2, 'ASE#TA'LE3 H<here BASEJTAB1E1 & BASEJTAB1E* are tables "a++ed t by the EIM table i# Cuesti #8 As a# alter#ati:e t ON1. BASE TAB1ESA this +ara"eter ca# be used t detail s+eci!ic tables that sh uld be i;# red by the EIM +r cess8 F r clarityA the #ly i#sta#ce i# <hich this +ara"eter < uld be used i#stead ! ON1. BASE TAB1ES is <here the "a@ rity ! the tables "ust be i#cluded8 By usi#; this +ara"eterA it <ill be clear <hich tables <ill # t be used8

This +ara"eter <ill ;e#erally # t be used I0NORE BASE TAB1ES

Pa;e 6 ! 7= IN CONFIDENCE

Siebel EIM Best Practice IN CONFIDENCE #%#%#%( Ignore $ase )olumns This is a Pr cess secti # +ara"eter8 ?sa;e$ IGNORE 'ASE COLUMNS = 'ASE#TA'LE20'ASE#COLUMN2,1 'ASE#TA'LE20'ASE#COLUMN3,1 'ASE#TA'LE30'ASE#COLUMN2 H<here BASEJTAB1E1 a#d BASEJTAB1E* are "a++ed t by the EIM table i# Cuesti # a#d BASEJCO1?MN1 & BASEJCO1?MN* are c lu"#s # each Base Table8 This +ara"eter ca# be used t i#struct EIM t i;# re the +r cessi#; ! c lu"#s s+eci!ied8 This is " st use!ul ! r i;# ri#; u#used ! rei;# /ey8 This < uld result i# EIM # t atte"+ti#; t res l:e the ! rei;# /ey i# Cuesti #8 This +ara"eter < uld ;e#erally # t be used # O#eSiebel <ith the use ! ON1. BASE CO1?MNS bei#; +re!erred8

This +ara"eter <ill ;e#erally # t be used I0NORE BASE CO1?MNS N t su++ rted ! r Mer;e +r cesses8

#%#%#%* Default )olumns There are a #u"ber ! EIM c lu"#s <here it "ay be desirable t set a de!ault :alue ! r a c lu"#A i! # :alue is s+eci!ied8 There is a# +ti # t s+eci!y these :ia the i!b !ile %DEFA?1T CO1?MN L CO1?MNJNAMEA E:alueD( h <e:er this ;i:es EIM a# additi #al S,1 state"e#t t e)ecute8 I# tu#i#; EIMA the b@ecti:e is t reduce the < r/ that EIM has t d A there! re !i)ed & de!ault c lu"#s sh uld be set by the e)ter#al i#ter!ace architecture i8e8 Ab I#iti 8

This +ara"eter sh uld # t be used DEFA?1T CO1?MN

#%#%#%+ Fi,ed )olumns There are a #u"ber ! EIM c lu"#s <here it "ay be desirable t set a !i)ed :alue ! r a c lu"#8 There is a# +ti # t s+eci!y these :ia the i!b !ile %FIKED CO1?MN L CO1?MNJNAMEA E:alueD( h <e:er this ;i:es EIM a# additi #al S,1 state"e#t t e)ecute8 I# tu#i#; EIMA the b@ecti:e is t reduce the < r/ that EIM has t d A there! re !i)ed & de!ault c lu"#s sh uld be set by the e)ter#al i#ter!ace architecture i8e8 Ab I#iti 8

This +ara"eter sh uld # t be used DEFA?1T CO1?MN

Pa;e 14 ! 7= IN CONFIDENCE

Siebel EIM Best Practice IN CONFIDENCE #%#%#%- .se Inde, Hints/ .se Essential Inde, Hints These are Pr cess secti # i!b !ile +ara"eter8 ?sa;e$ USE IN/E- )INTS = TRUE4FALSE USE ESSENTIAL IN/E- )INTS = TRUE4FALSE The de!ault :alues ! these +ara"eters are i#dicated i# boldA "ea#i#; that EIM adds i#de) hi#ts t certai# S,1 state"e#ts8 There are t< le:els ! ;ra#ularity a:ailable8 By s+eci!yi#; ?SE INDEK &INTS L TR?EA the# additi #al i#de) hi#ts are used !r " the de!ault8 I! # hi#ts are desiredA the# set b th +ara"eters t FA1SE8 EIM d es # t al<ays su;;est the " st a++r +riate e)ecuti # +ath a#d s t deter"i#e <hether the %hard c ded( hi#ts i"+r :e +er! r"a#ce r # tA EIM @ bs sh uld be +er! r"a#ce tested <ith a#d <ith ut this set t "rue i# rder t ;au;e +er! r"a#ce characteristics8 2ithi# B lu"e " st i#ter!aces %but # t all( ha:e bee# ! u#d t +er! r" better <ith # i#de) hi#ts i# +lace <he# the Oracle C st Based O+ti"iMer %CBO( is used8 The actual i#de) hi#ts the"sel:es ca## t be s+eci!ied @ust a# ability t tur# #N !! i#de) hi#ts is +r :ided8 I# B lu"e Release ' this +ara"eter <as used i# c #@u#cti # <ith the SESSION S,1 L OAlter sessi # set +ti"iMerJ" de L allJr <sO alth u;h as O#eSiebel ha:e ch se# the Siebel 585 :ersi #A the# the database <ill already be ru##i#; i# C st Based " de8 It is i"+ rta#t that O#eSiebel +er! r"a#ce testi#; e)+l res this area8 As Siebel 585Ds # < su++ rts %i# !actA "a#dates( the Oracle C st Based O+ti"iMerA the# the ?SE INDEK &INTSA ?SE ESSENTIA1 INDEK &INTS "ust be th r u;hly e)a"i#ed by Per! r"a#ce Testi#;8 O# B lu"e it has bee# bser:ed that Siebel !te# adds i#a++r +riate i#de) hi#ts such as hi#ti#; the use ! SJOR0JEKTJ?* t res l:e ! rei;# /eys t SJOR0JEKT8 As this i#de) is led by B?JID <hich is # #>selecti:e i# the B lu"e a++licati # %all Acc u#ts bel #; t the EBT B lu"eD r;a#isati #(A this is a# i#a++r +riate i#de) t use8 By tur#i#; hi#ts !!A the " re a++r +riate SJOR0JEKTJ?1 i#de) % r better still a cust " i#de)( "ay be ch se# by Oracle8

Pa;e 11 ! 7= IN CONFIDENCE

Siebel EIM Best Practice IN CONFIDENCE #%#%#%0 1ession 12L This +ara"eter has already bee# i#tr duced i# the secti # # ?SE INDEK &INTSA ?SE ESSENTIA1 INDEK &INTS8 ?sa;e$ SESSION S5L = 6sql statement to be executed7 N te that # ter"i#ati#; se"i>c l # %P( sh uld be used e;8 SESSION S5L = 8Alter 9e99ion 9et o!ti&i:er#&o"e = all#ro;98 This +ara"eter ca# be used t e)ecute cust " S,1 at the be;i##i#; ! EIMDs sessi # %EIM is a si#;le threadedA si#;le database c ##ecti # c "+ #e#t(8 O#ly a si#;le state"e#t ca# be issued8 Sessi # S,1 sh uld NOT be used t " di!y data i# the Siebel sche"a directly as Siebel d es # t su++ rt direct S,1 %e)ce+t ! r Cueryi#;(8 Sessi # S,1 sh uld NOT be used t " di!y data i# the Siebel sche"a directly as Siebel d es # t su++ rt direct S,1 %e)ce+t ! r Cueryi#;(8 As Siebel 585Ds # < su++ rts %i# !actA "a#dates( the Oracle C st Based O+ti"iMerA the# setti#; ! +ti"iser " de :ia SESSION S,1 "ay # l #;er be a++r +riate8 The Per! r"a#ce Tea" <ill i#:esti;ate this area8 #%#%#%3 .sing 1ynonyms This is a# EIM Pr cess secti # +ara"eter8 ?sa;e$ USING SYNONYMS = TRUE4FALSE EIMDs de!ault +erati # is t l / u+ Acc u#t Sy# #y"s i# rder t res l:e ! rei;# /eys t SJOR0JEKT8 This reCuires that EIM d es a# additi #al l /u+ a#d s setti#; this +ara"eter t #alse ca# reduce :erall ru# ti"e8 The de!ault is "rue8 Alth u;h O#eSiebel "ay ha:e a reCuire"e#t t su++ rt Sy# #y"s %"radin! $s #a"es(A ODM <ill su++ly Siebel ?ser Feys ! the Acc u#t as de!i#ed i# SJOR0JEKT a#d s setti#; this +ara"eter t FA1SE <ill # t i"+act this !u#cti #ality8

This +ara"eter sh uld be used ?SIN0 S.NON.MS L FA1SE

Pa;e 1* ! 7= IN CONFIDENCE

Siebel EIM Best Practice IN CONFIDENCE #%#%#%4 Rollbac5 !n Error/ )ommit Each Pass/ )ommit Each 'able These +ara"eters deter"i#e <hat ha++e#s <he# a l < le:el EIM err r ccurs %such as a# Oracle err r(8 By de!aultA EIM c ""its each table a#d each +assA "ea#i#; that i! a# err r ccursA the result "ay be +artially i"+ rted data8 E)a"+le The B lu"e C #su"er e#tity c #sists ! a# e#try i# SJPART.A SJOR0JEKTA SJACCNTJPOSTNA SJOR0JB?8 I! a %ow %evel ccurs duri#; EIM I"+ rtA a# e#try "ay be i#serted i#t SJPART.A SJOR0JEKT but # t i#t ther tablesA "a/i#; this rec rd l ;ically i#:alid <ithi# O#eBie<8 I# rder t +r +erly r llbac/ tra#sacti #s i# the e:e#t ! a %ow %evel err rA the# this +ara"eter sh uld be used i# c #@u#cti # <ith COMMIT EAC& PASS L FA1SEA COMMIT EAC& TAB1E L FA1SE8 ?sa;e$ RO11BACF ON ERROR L TR?ENFAL1E COMMIT EAC& PASS L 'R.ENFA1SE COMMIT EAC& TAB1E L 'R.ENFA1SE The de!aults are i#dicated i# bold8 As +er Siebel Alert *64A these setti#;s sh uld be set t +er! r" a si#;le c ""it %RO11BACF ON ERROR L TR?EA COMMIT EAC& PASS L FA1SEA COMMIT EAC& TAB1E L FA1SE( ! r all Mer;e a#d Delete O+erati #s8 2hilst it "ay i""ediately a++ear be#e!icial t use these setti#;s ! r I#sert a#d ?+date +erati #sA there is a hi;h ris/ that this < uld i"+act c "+eti#; +r cesses %li/e Siebel A#alytics( by causi#; a# Oracle ES#a+sh t t ldD err r8 This c "bi#ati # ! +ara"eters sh uldA there! reA #ly be used ! r EIM Mer;e a#d Delete +r cesses8 As it d es # t "a/e se#se t use RO11BACF ON ERROR <ith ut usi#; b th COMMIT EAC& PASS L FA1SE a#d COMMIT EAC& TAB1E L FA1SEA the# this +ara"eter sh uld # t be i#cluded i# the header secti # ! the i!b !ile %as is the case curre#tly # B lu"e(8

Pa;e 17 ! 7= IN CONFIDENCE

Siebel EIM Best Practice IN CONFIDENCE #%#%#% 6 .pdate After Failed Insert

I! si#;le c ""it is # t i"+le"e#ted %a#d it sh uld ;e#erally # t be used ! r EIM I"+ rt +r cesses( i# O#eBie<A ODMDs TrueDi!! +r cess sh uld chec/ all tables i# a "ulti>table e#tity t e#sure that the &pdate a ter ailed Insert +erati # +er! r"s +r +erly8 EIM I"+ rt @ bs i#te#ded t I#sert> #ly i#t the Base Tables are # t ;e#erally c #!i;ured <ith ?PDATE RO2S L FA1SEA <hich "ay ther<ise be e)+ected %t a: id a Data ,uality Issue <here a# a# "aly has ccurred(8 This is s that i! the i#itial i#sert !ailsA the# se#di#; the rec rd a;ai# <ill result i# a# u+date t a +artial rec rd8 I# B lu"e at +rese#tA this is # t i"+le"e#ted +r +erly because the TrueDi!! +r cess %the ODM +r cess that ide#ti!ies <hich rec rds t i#sertA u+date r delete !r " Siebel based # c "+ari#; the data <ithi# Siebel <ith that <ithi# ODMDs datast res( d es # t c "+are c lu"#s i# all tables ! r e:ery e#tity8 As +er Siebel Alert *64A these setti#;s sh uld be set t +er! r" a si#;le c ""it %RO11BACF ON ERROR L TR?EA COMMIT EAC& PASS L FA1SEA COMMIT EAC& TAB1E L FA1SE( ! r all Mer;e a#d Delete O+erati #s8 I! si#;le c ""it is # t used %a#d it sh uld # t be(A ODMDs TrueDi!! +r cess sh uld be ca+able ! i#cludi#; all c lu"#s i# all Base tables i#: l:ed i# each EIM +r cess8

Pa;e 1' ! 7= IN CONFIDENCE

Siebel EIM Best Practice IN CONFIDENCE

#%#%( IF$ Import Process 1pecific Parameters


This secti # details th se +ara"eters that are s+eci!ic t EIM @ bs ! ty+e EI"+ rtD %b th I#sert a#d ?+date +erati #s(8 As <ell as the +ara"eters detailed <ithi# this secti #A th se +ara"eters i# the 0e#eric secti # ca# als be a++lied8 #%#%(% !nly $ase )olumns This is a Pr cess secti # +ara"eter8 ?sa;e$ IGNORE 'ASE COLUMNS = 'ASE#TA'LE20'ASE#COLUMN2,1 'ASE#TA'LE20'ASE#COLUMN3,1 'ASE#TA'LE30'ASE#COLUMN2 H<here BASEJTAB1E1 a#d BASEJTAB1E* are "a++ed t by the EIM table i# Cuesti # a#d BASEJCO?MN1 & BASEJCO1?MN* are c lu"#s # each Base Table8 2hereas the ;e#eric +ara"eter EI;# re Base C lu"#sD i#structs EIM t i;# re s+eci!ic c lu"#sA EO#ly Base C lu"#sD i#structs EIM t #ly +r cess th se c lu"#s s+eci!ically listed8 It d es # t "a/e se#se t use b th EI;# re Base C lu"#sD a#d EO#ly Base C lu"#sD i# the sa"e @ b8 Either s+eci!y #e r the ther8 ON1. BASE CO1?MNS sh uld be used ! r e:ery i#ter!ace t achie:e best +er! r"a#ce . u "ust i#clude F rei;# Fey c lu"#s %I#sert a#d ?+date(A reCuired !ields %I#sert +erati #( a#d a#y data c lu"#s t be l aded8

?sa;e$

ON1. BASE CO1?MNS sh uld be used ! r e:ery i#ter!ace t achie:e best +er! r"a#ce8

#%#%(%# Insert Ro"s/ .pdate Ro"s These t< Pr cess secti # +ara"eters i"+act u+ # b th Data ,uality a#d Per! r"a#ce8

INSERT RO.S = 'ASE#TA'LEN, TRUE4FALSE U(/ATE RO.S = 'ASE#TA'LEN, TRUE4FALSE As has already bee# statedA a# EIM table "a+s t "ulti+le base tables8 The use ! the ab :e +ara"eters c #tr ls <hat ty+e ! +erati # is all <ed # each s+eci!ied table8 The de!ault ! r b th +ara"eters is "rue8 As a# EIM I"+ rt @ b is actually a# &psert +erati #8 i8e8 i! the ?ser Feys + +ulated i# the EIM table "atches th se ! a# e)isti#; r < i# the Base TablesA the# the +erati # is a# ?+date8 I! the ?ser Feys + +ulated i# the EIM table d # t "atch th se ! a# e)isti#; r < i# the Base TablesA the# the +erati # is a# I#sert +erati #8 BT has i#:ested hea:ily i# the ODM layerDs ca+ability ! creati#; se+arate I#sert a#d ?+date tra#sacti #sA based ar u#d the unload !u#cti #ality <hich is #ecessary due t multi'masterin!8 As suchA ODM deter"i#es <hether the +erati # should be a# i#sert r a# u+date t SiebelDs data8 There! reA the auth r belie:es that e:ery EIM @ b sh uld s+eci!ically use INSERT RO2S Pa;e 1- ! 7= IN CONFIDENCE

Siebel EIM Best Practice IN CONFIDENCE a#d ?PDATE RO2S t s+eci!y the +erati #s that EIM is allowed t +er! r"8 This <ill +re:e#t a#y r <s !r " bei#; u+dated %<here ?ser Feys "atch i# err r( r i#serted %<he# ?ser Feys are # t l cated( u#i#te#ti #ally8 Per! r"a#ce ca# be i"+r :ed by +re:e#ti#; redu#da#t u+dates8 INSERT RO2S a#d ?PDATE RO2S sh uld be used t ti;hte# Data ,uality a#d i"+r :e +er! r"a#ce8 I! si#;le c ""it is # t usedA ODMDs TrueDi!! +r cess sh uld be ca+able ! chec/i#; all tables i#: l:ed i# a# EIM i#ter!ace8 This < uld "ea#A +ti"allyA that a# I#sert i#ter!ace c uld s+eci!y ?PDATE RO2S L FA1SE # base tables s that # err #e us u+date < uld be all <ed8 I! TrueDi!! is # t !ully a<are ! each Siebel e#tityA the# I#sert @ bs < uld #eed t be set t all < u+dates alth u;h this < uld#Dt al<ays !i) rec rds as i#te#ded %sel!>heali#; ca+ability(8 Alth u;h INSERT RO2S L TR?E is the de!aultA it is a ; s that the i#ter!aceDs i#te#ti # is # t a"bi;u us8 d idea t state this s+eci!ically

Full use ! INSERT RO2S a#d ?PDATE RO2S <ill "ea# that a#y ccurre#ces ! PARTIA11.JIMPORTED sh uld be treated as ri;ht!ul err rs8 #%#%(%( 7et )hange This +ara"eter is used t s+eci!y <hether c lu"#s ca# be u+dated t N?118 ?sa;e$ NET C)ANGE = TRUE4FALSE The de!ault :alue is sh <# i# bold8 2he# set t FA1SEA Base Table c lu"#s <ill be u+dated t N?11 i! the c rres+ #di#; EIM table c lu"# is N?11 a#d ON1. BASE CO1?MNS is used8 This is the #ly <ay t u+date a c lu"# t N?118 EIMDs de!ault beha:i ur is t s/i+ +r cessi#; ! N?11 c lu"#s i# the EIM table %e)ce+t <here ON1. BASE CO1?MNS is used(8 Because ! the <ay i# <hich ODM ada+ters are desi;#ed a#d <i#; t the !act that ON1. BASE CO1?MNS is used thr u;h+utA NET C&AN0E L FA1SE is used ! r all ?+date +r cesses8 2hilst the u#derlyi#; S,1 ;e#erated by NET C&AN0E L FA1SE i"+le"e#ts this by use ! a lar;e NB1 based state"e#tA this has # t bee# see# t cause a si;#i!ica#t +r cessi#; :erhead8

Pa;e 1= ! 7= IN CONFIDENCE

Siebel EIM Best Practice IN CONFIDENCE #%#%(%* !racle Insert Append Mode This is a #e< !eature i# Siebel 58-)8 ?sa;e$ ORACLE INSERT A((EN/ MO/E = TRUE4FALSE The de!ault is sh <# i# bold8 ?si#; this "ea#s that <he# i#serti#; data i#t the Base tablesA OracleDs I#sert A++e#d M de is used8 This i#serts data i#t #e< bl c/s ab :e the &i;h 2ater Mar/8 2hilst this "ay i"+r :e ;e#eral i#sert +er! r"a#ceA use ! this +ara"eter <ill lead t i#creased !ra;"e#tati # a#d i#creased st ra;e reCuire"e#ts8 The B lu"e Pr ducti # DBA tea" d # t su++ rt use ! this +ara"eter8 D # t use this +ara"eter as it ca# lead t i#creased Base Table !ra;"e#tati # as <ell as i#creased st ra;e reCuire"e#ts8

Pa;e 15 ! 7= IN CONFIDENCE

Siebel EIM Best Practice IN CONFIDENCE #%#%(%+ 'rim 1paces This is a# EIM Pr cess secti # +ara"eter8 ?sa;e$ TRIM S(ACES = TRUE4FALSE EIMDs de!ault +erati # is t ri;ht tri" <hite s+ace ar u#d data !ields$
U(/ATE SET SIE'EL0EIM#ACCOUNT IF#RO.#STAT#NUM = <, IF#RO.#STAT = =IN#(ROGRESS=, T#ORG#E-T##STA = <, T#ORG#E-T##RI/ = NULL, T#ORG#E-T##E-S = =N=, T#ORG#E-T##UN5 = =N=, T#ORG#E-T#'U#I/ = NULL, T#ORG#E-T#CURR(R = NULL, 000 000 EN/#/T = /ECO/E>EN/#/T,NULL,TO#/ATE>NULL?, TO#/ATE>TO#C)AR>EN/#/T,=YYYY@MM@//=?44= <<A<<A<<=,=YYYY@MM@// ))3BAMIASS=??, START#/T = /ECO/E>START#/T,NULL,TO#/ATE>NULL?, TO#/ATE>TO#C)AR>START#/T,=YYYY@MM@//=?44= <<A<<A<<=,=YYYY@MM@// ))3BAMIASS=??, NAME = RTRIM>NAME?, LOC = /ECO/E>LOC,NULL,NULL,RTRIM>LOC??, ALIAS#NAME = /ECO/E>ALIAS#NAME,NULL,NULL,RTRIM>ALIAS#NAME??, ASGN#()#AC = /ECO/E>ASGN#()#AC,NULL,NULL,RTRIM>ASGN#()#AC??, ASGN#()#CC = /ECO/E>ASGN#()#CC,NULL,NULL,RTRIM>ASGN#()#CC??, 'ASE#CURCY#C/ = /ECO/E>'ASE#CURCY#C/,NULL,NULL,RTRIM>'ASE#CURCY#C/??, 000 000 URL = /ECO/E>URL,NULL,NULL,RTRIM>URL??, $AT#REGN#NUM = /ECO/E>$AT#REGN#NUM,NULL,NULL,RTRIM>$AT#REGN#NUM??, RO.#STATUS = /ECO/E>RO.#STATUS,NULL,NULL,RTRIM>RO.#STATUS??, COMMENTS = /ECO/E>COMMENTS,NULL,NULL,RTRIM>COMMENTS??, STATUS = /ECO/E>STATUS,NULL,NULL,RTRIM>STATUS?? .)ERE >IF#RO.#'ATC)#NUM = 2?

N t all c lu"#s are tri""ed but this ca## t be c #tr lled8 By setti#; this +ara"eter t FalseA the# this +erati # is # t +er! r"ed duri#; the EI#itialiMe I#ter!ace TablesD EIM +ass8 2hilst this the retically sli"s d <# EIM +r cessi#;A the ;ai#s t by had by setti#; TRIM SPACES L FA1SE <ill be #e;li;ible <he# c "+ared t the i"+act ! ther EIM +ara"eters8 Additi #allyA b th the eB l:e a#d B lu"e de:el +"e#t tea"s ha:e e)+ressed that they +re!er t retai# the de!ault ! TR?E8

TRIM SPACES sh uld be le!t at the de!ault :alue ! TR?E8 It adds # :alue t s+eci!y this e)+licitly i# the i!b8

Pa;e 13 ! 7= IN CONFIDENCE

Siebel EIM Best Practice IN CONFIDENCE

#%#%* EIM Delete Process )onfiguration


This secti # details +ara"eters a#d tech#iCues as <ell as i"+ rta#t Tech#ical I#! r"ati # !r " Siebel Su++ rtA <hich is s+eci!ic t EIM @ bs ! ty+e EDeleteD a#d EMer;eD8 This secti # is structured di!!ere#tly tha# ther li/e secti #s <ithi# this d cu"e#t as this <al/s the reader thr u;h each ty+e ! Delete @ bA #a"ely$ o o o #%#%*% Delete All R <s Delete E)act Delete Matches Delete All Ro"s

This Pr cess secti # +ara"eter is used t s+eci!y that all r <s "a++ed t by the i#ter!ace table sh uld be deleted8 This +ara"eter sh uld there! re be used <ith e)tre"e cauti #Q ?sa;e$ /ELETE ALL RO.S = TRUE4FALSE The de!ault :alue is False8 N ther +ara"eters are really be :alid as all data i# the s+eci!ied Base Table <ill be deleted8

O#ce ru#A the ?ser Feys ! rec rds deleted by the +r cess <ill be l aded i#t the EIM tables a#d ca# be archi:ed r deleted8 This +ara"eter sh uld be used <ith e)tre"e cauti #Q It is # t +articularly rele:a#t t Pr ducti # Pr cesses8 #%#%*%# Delete E,act This Pr cess secti # +ara"eter is used t s+eci!y that EIM sh uld delete r <s i# the Base Tables <h se ?ser Feys "atch th se ! rec rds l aded i#t the I#ter!ace Tables8 ?sa;e$ /ELETE E-ACT = TRUE4FALSE As EIM #ly #eeds the ?ser FeysA the# #ly the ?ser Feys sh uld be l aded8 This +erati # is +ri"arily c #cer#ed <ith deleti#; r <s !r " the +ri"ary Base Tables "a++ed t by the ;i:e# EIM table8 Deleti # ! Child R <s ca# be c #tr lled by use ! ON1. BASE TAB1ES8 DE1ETE RO2S ca# be used t s+eci!y !r " <hich tables deletes are all <ed8 The de!ault :alue is True8 ?sa;e$ /ELETE RO.S = 'ASE#TA'LEN, TRUE4FALSE H<here BASEJTAB1EN is a Base Table "a++ed t by the s+eci!ic EIM table8 ?PDATE RO2S is als :alid ! r this @ b8 Pa;e 16 ! 7= IN CONFIDENCE

Siebel EIM Best Practice IN CONFIDENCE Setti#; DE1ETE RO2S a#dG r ?PDATE RO2S t FA1SE sh uld be d #e <ith e)tre"e cauti # as it "ay result i# da#;li#; ! rei;# /eys i8e8 ! rei;# /eys that # l #;er + i#t t a :alid rec rd8 O#ce ru#A # additi #al data is + +ulated i#t the i#ter!ace tables a#d s the ri;i#al r <s that <ere l aded i#t the EIM table "ay be archi:ed r deleted8 Setti#; DE1ETE RO2S a#dG r ?PDATE RO2S t FA1SE sh uld be d #e <ith e)tre"e cauti # as it "ay result i# da#;li#; ! rei;# /eys8 DE1ETE EKACT is the sa!est "eth d ! deleti#; data i# bul/ !r " Siebel a#d s is the ad +ted delete "eth d at BT8

#%#%*%( Delete Matches This Pr cess secti # +ara"eter is used t s+eci!y that EIM sh uld delete data !r " the Base Tables that "atches the where clause s+eci!ied i# the i!b8 e8;8 /ELETE MATC)ES = EIM#ACCOUNT, >CUST#STAT#C/ = CInactiDeE? I# the e)a"+le ab :eA all r <s <h se C?STJSTATJCD L EI#acti:eD < uld be deleted8 O#ce ru#A this EIM @ b <ill ha:e <ritte# the ?ser Feys ! the rec rds deleted i#t the EIM tables8 This data ca# the# be either archi:ed r deleted8 There is # c #ce+t ! BatchsiMe i# a Delete Matches @ b all r <s that "eet the criteria are +r cessed i# a si#;le ru#8 There is # c #ce+t ! batchsiMe i# a Delete Matches @ b8 There! reA se:eral @ bs "ay be reCuired t delete the i#te#ded rec rdsA +erha+s ;r u+i#; the" by a data ra#;e ! ra#;e ! ?#iCue Ide#ti!iers8

Pa;e *4 ! 7= IN CONFIDENCE

Siebel EIM Best Practice IN CONFIDENCE

2.2.4.3.1

Im ortant In!ormation !rom Siebel Su

ort

There are t< i"+ rta#t Siebel Alerts #eed t be /e+t i# "i#d ! r all Delete a#d Mer;e @ bs8 These $lerts sh uld be i"+le"e#ted %sub@ect t c #!ir"ati #Ga++r :al !r " Siebel( i# all Delete a#d Mer;e @ bs8 o There is a Siebel Alert %Alert (6-8 1pecify 97.M:IF'A$LE:L!AD:).'!FF ; 6 for EIM merge and delete( that states that y u "ust s+eci!y a +ara"eter %RN?MJIFTAB1EJ1OADJC?TOFFL4( ther<ise EIM <ill be (lind t s "e "etadata relati#; t ! rei;# /eys8 This c uld clearly cause D, issues as it "ay lea:e da#;li#; ! rei;# /eys8 2hilst this +ara"eter sh uld be able t be s+eci!ied i# the IFB &eader secti # as detailed <ithi# the AlertA this <as ! u#d t re+eatedly cause a c re du"+ # the S laris A++licati # Ser:er8 A Ser:ice ReCuest <as l ;;ed # Siebel Su++ rt2eb a#d the < r/ar u#d ! s+eci!yi#; this :ia E)te#ded Para"eters # the c ""a#d li#e <as ! u#d8 Or+ha#ed rec rds "ay be created i# the curre#t c #!i;urati # i# the e:e#t that a# Oracle err r % r ther si"ilarly se:ere err r( ccurs <hilst the EIM @ b is ru##i#;8 This is as a result ! a# ther Siebel Alert %Alert #648 EIM Merge and Delete processes should use only one commit per batch to avoid data integrity issues ( a#dA a;ai#A "ay cause D, issues8 T +re:e#t r+ha#s !r " bei#; created the ! ll <i#; setti#;s sh uld be s+eci!ied i# the Pr cess secti #$ COMMIT EAC& TAB1E L FA1SE COMMIT EAC& PASS L FA1SE RO11BACF ON ERROR L TR?E %I! a se:ere err r d es ccursA the e#tire batch is r lled bac/ but rec rds are le!t i# the EIM table <ith a# IFJRO2JSTAT ! EINJPRO0RESSD because Siebel d es a c ""it a!ter preparin! the inter ace ta(les %see "rim )paces secti # <ithi# this d cu"e#t ! r a# e)a"+le EIM l ; !ile sh <i#; this(8 There! reA i! rec rds are le!t <ith a# IFJRO2JSTAT L EINJPRO0RESSD a!ter all EIM @ bs ha:e c "+letedA the# it is clear that a l <>le:el err r has ccurred8( ?se the ! ll <i#; +ara"eters ! r all Delete a#d Mer;e @ bs$

COMMIT EAC& TAB1E L FA1SE COMMIT EAC& PASS L FA1SE RO11BACF ON ERROR L TR?E RN?MJIFTAB1EJ1OADJC?TOFF L 4 sh uld be s+eci!ied ! r all Delete a#d Mer;e @ bs8 ODMDs 0S1 % r a#y ther calli#; +r cess( "ust be able t s+eci!y this +ara"eter a#dA i# !actA it < uld be <ise that the 0S1 c uld build a !le)ible Ser:er Ma#a;er c ""a#d li#e " de call s that E*tended Parameters a#d ther s+ecialised attributes ca# be i#cluded8 This < uld reCuire that the curre#t ODM EIMJDETAI1 table a#d surr u#di#; !u#cti #ality be e#ha#ced8 I#cludi#; RN?MJIFTAB1EJ1OADJC?TOFF L 4 <ithi# the IFB !ileDs &eader secti # "ay cause c re du"+s # the S laris A++licati # Ser:er8 T < r/ar u#d thisA s+eci!y the +ara"eter :ia E)te#ded Para"eters # the c ""a#d li#e8

Pa;e *1 ! 7= IN CONFIDENCE

Siebel EIM Best Practice IN CONFIDENCE

#%#%+ EIM Merge Process )onfiguration


I# a Mer;e @ bA victim a#d survivor rec rds are s+eci!ied i# the EIM tables8 The IFJRO2JMER0EJID # the :icti" rec rd is + +ulated <ith the r <Jid %i# the I#ter!ace Table( ! the survivor rec rd e8;8 IF:R!<:$A')H:7.M 144 144 144 144 144 I# the e)a"+le ab :e$ +ne)ie(el a#d +ne,iew are sur:i: r rec rds %# IFJRO2JMER0EJID s+eci!ied(8 ,olume a#d -i! Data are "er;ed i#t +ne,iew8 $ccount Mana!ed Plat orm is "er;ed i#t +ne)ie(el8 R!<:ID 1 * 7 ' 7AME O#eSiebel O#eBie< B lu"e Acc u#t Ma#a;ed Plat! r" Bi; Data IF:R!<:MER&E:ID

* 1 *

A Mer;e @ b d es # t ha:e a#y s+eci!ic +ara"etersA h <e:er there are a #u"ber ! rec ""e#dati #s a#d bser:ati #s$ o o o A Mer;e @ b #ly #eeds the ?ser Feys a#d IFJRO2JMER0EJID # the victim rec rd t be s+eci!ied8 Siebel Alerts 74= a#d *46 %as detailed i# the secti # e#titled Important In ormation rom )ie(el )upport( sh uld be i"+le"e#ted8 ?PDATE RO2S is the #ly ther rele:a#t setti#;8 This c #tr ls <hether ! rei;# /eys i# a child table are u+dated t + i#t t <ards the survivor r # t8 Setti#; this +ara"eter t FA1SE ca# cause da#;li#; ! rei;# /eys a#d is # t rec ""e#ded8 The batchsiMe used sh uld be c #siderably s"aller tha# the batchsiMe usually ad +ted e8;8 # B lu"e the sta#dard EIM batchsiMe is 14A444 but a batchsiMe ! @ust *444 is used ! r Address Mer;e +erati #s8 2hilst EIM Mer;e ca# the retically be ru# i# +arallel i! rec rds are :icti" rec rds are "utually e)clusi:e acr ss all batchesA this is # t rec ""e#ded Oracle l c/s ccurred duri#; Per! r"a#ce Testi#; # a #u"ber ! ccasi #s a#d as database res urce is c #siderable e:e# ! r a si#;le EIM strea"A the# ru##i#; EIM "er;e i# +arallel is li/ely t saturate IO res urces :ery Cuic/ly a#d result i# a# own !oal8 As i# the e)a"+le ab :e i! " re tha# 1 :icti" is t be "er;ed i#t sur:i: rA this sh uld be d #e i# the sa"e batch %i! "utual e)clusi:ity ca# be ;uara#teed(8 Mer;e +erati #s sh uld NOT be s+lit i#t batches ! * c #tai#i#; @ust a sur:i: r a#d a :icti" %as <as the case # B lu"e at #e + i#t(8

Pa;e ** ! 7= IN CONFIDENCE

Siebel EIM Best Practice IN CONFIDENCE o B lu"e is curre#tly e)+erie#ci#; di!!iculties <ith EIM Mer;e @ bs due t Per! r"a#ce a#d Data ,uality Issues8 o Per! r"a#ce Due t the lar;e #u"ber ! relati #shi+s <ithi# the Siebel databaseA # t all rele:a#t ! rei;# /ey c lu"#s # the database ha:e bee# i#de)edA ut ! the b )8 This i#: /es Full Table Sca#s <hich are # t :iable a;ai#st O#eBie<Ds lar;est tables8 Siebel E)+ert ser:ices are curre#tly e#;a;ed t res l:e the issues a#d a res luti # is ! rthc "i#;8 o Data ,uality A Siebel d cu"e#ted issue <hereby duplicate columns are le!t behi#d i# i#tersecti # tables <as rece#tly e)+erie#ced by the B lu"e tea"8 2hilst this issue is re+ rtedly res l:ed i# 585A it "erits that due care be ta/e#8 ?se the ! ll <i#; +ara"eters ! r all Delete a#d Mer;e @ bs$

COMMIT EAC& TAB1E L FA1SE COMMIT EAC& PASS L FA1SE RO11BACF ON ERROR L TR?E RN?MJIFTAB1EJ1OADJC?TOFF L 4 sh uld be s+eci!ied ! r all Delete a#d Mer;e @ bs8 ODMDs 0S1 "ust be able t s+eci!y this +ara"eter a#dA i# !actA it < uld be <ise that the 0S1 c uld build a !le)ible Ser:er Ma#a;er c ""a#d li#e " de call s that E*tended Parameters a#d ther s+ecialised attributes ca# be i#cluded8 This < uld reCuire that the curre#t B lu"e EIMJDETAI1 table a#d surr u#di#; !u#cti #ality be e#ha#ced8 The B lu"e tea" e#c u#tered Data Cuality issues due t a# ut> !>the>b ) Siebel issue8 2hilst this has bee# re+ rtedly !i)ed i# 585A due care #eeds t be ta/e#8

Pa;e *7 ! 7= IN CONFIDENCE

Siebel EIM Best Practice IN CONFIDENCE

#%#%- !ther Parameters


The ! ll <i#; table i#cludes details # ther EIM +ara"eters ! r c "+lete#ess8 Para"eters reCuiri#; atte#ti # i#clude descri+ti # te)t i# bold8 N8B8 This ;uide <as +re+ared <ith re!ere#ce t b th Siebel 58-8* a#d Siebel 5858 Parameter Description Header 1ection !nly Parameters +ther parameters can !enerall. (e speci ied in either the header or the process section. CONNECT ODBC data s urce # t reCuired as this is set usi#; data retrie:ed !r " the 0ate<ay Ser:er 1O0 TRANSACTIONS TO A Siebel Re" te related +ara"eter this should be set to FI1E FAL1E in the header section of all IF$ files "hilst 1iebel Remote is not being used% All Siebel Re" te related +ara"eters <ill #eed t be re:ie<ed i! Siebel Re" te is t be de+l yed8 ?SERNAME This +ara"eter sh uld # t be set8 Rather this i#! r"ati # is retrie:ed !r " the 0ate<ay8 PASS2ORD This +ara"eter sh uld # t be set8 Rather this i#! r"ati # is retrie:ed !r " the 0ate<ay8 TAB1EO2NER This +ara"eter sh uld # t be set8 Rather this i#! r"ati # is retrie:ed !r " the 0ate<ay8 PROCESS This de!i#es the +r cess t i#: /e <he# EIM is started <ith the IFB !ile8 'his should not be used as it is +re!erred that the +r cess #a"e be s+eci!ied e)ter#ally i8e8 s+eci!ied # the c ""a#d li#e <he# i#: /i#; EIM8 &eneric Process Parameters BATC& The batch r batch ra#;e t be +r cessed8 INC1?DE ?sed i# Shell Pr cesses t aut "ate the ru##i#; ! a #u"ber ! +r cesses aut "atically8 Shell Pr cesses sh uld ;e#erally # t be usedA as scheduli#; ! @ bs is ;e#erally "astered ut> <ith the EIM e#:ir #"e#t a#d s+eci!ic EIM +r cesses t i#: /e are s+eci!ied # the Ser:er Ma#a;er c ""a#d li#e8 De ault.i (A <hich ca# be ! u#d <ithi# the A++licati # Ser:erDs ad"i# direct ryA c #tai#s e)a"+les ! calli#; Shell Pr cesses8 1O0 TRANSACTIONS A Siebel Re" te related +ara"eter8 This +ara"eter is #ly rele:a#t i! Dockin!/"ransaction %o!!in! Syste" Pre!ere#ce is set t True8 N t rele:a#t t either O#eSiebel r184 r O#eBie< r=848 SFIP B?JID DEFA?1T The de!ault :alue ! False sh uld be retai#ed t s+eci!y that EIM uses the de!ault :alue de!i#ed <ithi# the Re+ sit ry <he# setti#; :irtual #ull /eys8 TAB1E This +ara"eter s+eci!ies <hich EIM table is t be +r cessed by the EIM @ b8 TRANSACTION S,1 The +ti #al +ara"eter ca# be used %si"ilarly t SESSION S,1( t ru# ad>h c S,1 a!ter each EIM S,1 state"e#t is c ""itted8 T date # s+eci!ic use has bee# ! u#d ! r this +ara"eter8 T.PE S+eci!ies <hether the EIM +r cess is ! ty+e IMPORT %i#cludes BT ?+date a#d I"+ rt +r cesses(A EKPORTA DE1ETEA MER0E r S&E118 Shell +r cesses sh uld # t ;e#erally be used8 De ault.i (A <hich ca# be ! u#d <ithi# the A++licati # Ser:erDs ad"i# direct ryA c #tai#s e)a"+les ! calli#; Shell Pr cesses8 Pa;e *' ! 7= IN CONFIDENCE

Siebel EIM Best Practice IN CONFIDENCE Parameter Description Header 1ection !nly Parameters +ther parameters can !enerall. (e speci ied in either the header or the process section. ?PDATE STATISTICS As this +ara"eter #ly a++lies t DB*A it should not be used <ithi# BT %BT uses Oracle(8 Import Process Parameters COMMIT OPERATIONS A Siebel Re" te related +ara"eter8 N t rele:a#t t either O#eSiebel r184 r O#eBie< r=848 FI1TER ,?ER. This +ti # ca# be used t s+eci!y <hich r <s i# the EIM table are t be +r cessed e8;8 FI1TER ,?ER. L %C?STJSTATJCD L SActi:eT(8 As BT has i#:ested hea:ily i# e)ter#al +re+arati # ! data ! r EIM %Ab I#iti ;ra+hs(A the# there sh uld be # #eed t use a +ara"eter li/e this8 MISC S,1 This +ara"eter sh uld be used ! r i#itial l ads #ly the auth r <ill re:ie< its a++licability ! r O#eSiebel at a later date8 ATTAC&MENT DIRECTOR. This +ara"eter is used t s+eci!y the l cati # ! attach"e#ts t be l aded i#t Siebel8 EIM "ust ha:e b th read a#d <rite access t this direct ry8 %De!ault L SIEBE1J&OMEGINP?T( E,port Process Parameters N te$ The EIM E)+ rt +r cess is # t ;e#erally used <ithi# BT but C1EAR INTERFACE TAB1E The de!ault ! r this +ara"eter is True "ea#i#; that the EIM e)+ rt +r cess deletes all e)isti#; r <s i# the i#ter!ace table +ri r t starti#; the e)+ rt +r cess8 ?se ! this +ara"eter is #ly + ssible <here # ther usersG+r cesses are shari#; the s+eci!ic EIM table t be used8 EKPORT A11 RO2S Sy# #y" us <ith DE1ETE A11 RO2S EKPORT MATC&ES Sy# #y" us <ith DE1ETE MATC&ES Delete Process Parameters CASCADE DE1ETE ON1. ?sed t s+eci!y <hether child rec rds are deleted %TR?E( r their FF is u+dated t N?11 %TR?E(8 The de!ault is False8 DE1ETE SFIP PRIMAR. This +ara"eter s+eci!ies <hether EIM sh uld +er! r" a cascade u+date t the +ri"ary child c lu"#8 The De!ault :alue is TR?E8 Merge Process Parameters SET BASED 1O00IN0 A Siebel Re" te related +ara"eter8 This +ara"eter is #ly rele:a#t i! Dockin!/"ransaction %o!!in! Syste" Pre!ere#ce is set t True8 N t rele:a#t t either O#eSiebel r184 r O#eBie< r=848

Pa;e *- ! 7= IN CONFIDENCE

Siebel EIM Best Practice IN CONFIDENCE

( Running EIM
3.1 Data
Re:ie<s ! se:eral Siebel i"+le"e#tati #s at BT has re:ealed that there hasA # ccasi #A bee# s "e c #!usi # as t <hat data #eeds t be + +ulated ! r each di!!ere#t @ b ty+e8 This secti # clari!ies <hat data is reCuired ! r each ty+e ! EIM @ b8 Import =Insert> T i#sert a #e< rec rd %i#t Pri"ary r Child Base Tables(A ?ser FeysA ReCuired FieldsA Data c lu"#s a#d Syste" C lu"#s #eed t be + +ulated8 Import =.pdate> T u+date a# e)isti#; rec rd @ust the ?ser FeysA s+eci!ic Data C lu"#s t u+date a#d Syste" C lu"#s #eed t be + +ulated8 Delete T Delete a rec rd usi#; Delete E)actA @ust the ?ser Feys a#d Syste" C lu"#s #eed t be + +ulated8 Merge T Mer;e rec rdsA @ust ?ser Feys a#d Syste" !ields #eed be + +ulated8

The ! ll <i#; de!i#iti #s a++ly t the ab :e$ .ser ?eys the c "bi#ati # ! c lu"#s used by EIM t u#iCuely ide#ti!y a r < %sy# #y" us <ith EC "+ u#d FeysD(8 Re@uired Fields th se !ields that "ust be + +ulated <he# creati#; a #e< rec rd8 Data )olumns re;ular c lu"#s i# the EIM table that "a+ t c lu"#s i# the Base Tables8 1ystem )olumns CREATEDA CREATEDJB.A 1ASTJ?PDA 1ASTJ?PDJB.A RO2JIDA IFJRO2JBATC&JN?MA IFJRO2JMER0EJIDA IFJRO2JSTATA MODIFICATIONJN?MA CONF1ICTJID8

The ! ll <i#; are "a#dat ry Syste" C lu"#s$ IF:R!<:$A')H:7.M c #tr ls the Batch t <hich the rec rd bel #;s R!<:ID used t u#iCuely ide#ti!y rec rds <ithi# a batch8 The c "bi#ati # ! IFJRO2JBATC&JN?M a#d RO2JID "ust be u#iCue8 IF:R!<:MER&E:ID %! r EIM Mer;e +r cesses #ly( set this t the RO2JID <ithi# the batch that the rec rd <ill be "er;ed i#t 8 IF:R!<:1'A' set <ith a "ea#i#;!ul statusA such as EFORJIMPORTDA EFORJ?PDATED8 This <ill be set t EINJPRO0RESSD by EIM duri#; +r cessi#;8

The ther Syste" C lu"#s are de!aulted # a table le:el %Oracle de!ault :alue( i# s "e i#sta#ces # t all i#sta#ces8 It isA h <e:erA ; d +ractice t set these !ields e)+licitly s that rec rds i# EIM tables ca# be easily ide#ti!ied8 N te that the :alues + +ulated i# these ther Syste" C lu"#s <ill # t i#!lue#ce the c rres+ #di#; Base table rec rds these c lu"#s are set by the EIM +r cess8

O#ly l ad the data reCuired by the i#ter!ace8

Pa;e *= ! 7= IN CONFIDENCE

Siebel EIM Best Practice IN CONFIDENCE

3.2 I#- #ile and EIM Processes


This secti # details ;uideli#es that sh uld be ad +ted by BT Siebel i#ter!ace de:el +"e#t tea"s relati#; t h < t l ;ically c llate "ulti+le EIM +r cesses i#t i!b !iles8 I# ;e#eralA there sh uld be a si#;le i!b !ile +er e#tityA <ith "ulti+le +r cesses bei#; "ai#tai#ed <ithi# that i!b !ile e8;8$ A !ile called onesie(el0customer.i ( sh uld c #tai# all Cust "er related EIM +r cesses ! r O#eSiebelA such as$ I"+ rtJO#eSiebelJCust "er ?+dateJO#eSiebelJCust "er CeaseJO#eSiebelJCust "er DeleteJO#eSiebelJCust "er Mer;eJO#eSiebelJCust "er

N te$ the Cease +r cess is a l ;ical delete ! the Cust "er <hilst the Delete +r cess is a +hysical delete8 By i#cludi#; "ulti+le EIM +r cesses ! r a ;i:e# e#tity <ithi# a si#;le IFB !ileA the# c #!i;urati # "a#a;e"e#t ca# be c #tr lled at a# a++r +riate le:el8 The calli#; +r cess %Ab I#iti ada+terA scri+t etc( sh uld s+eci!y the +r cess t ru# %a#d ther +ara"eters( :ia the c ""a#d li#e e8;8$ *SIE'EL#ROOTFbinF9rDr&gr @g *GATE.AYA*G(ORT @e *ENTER(RISE @9 *A((SER$ER @% *SUSER @! *S(ASS @c 8r%n ta9G for co&!onent ei& ;it+ config=8*IF'8, traceflag9=H, errorflag9=2, S5LFlag9=I, EJten"e"(ara&9=8'atc+Range=*'ATC)8, (roce99=8*(NAME88 1&2I3 escape se4uences have (een omitted or clarit.5. RSIEBE1JROOT the +ath <here the Siebel A++licati # Ser:er is i#stalled R0ATE2A. the IP address % r h st#a"e( ! the Siebel 0ate<ay Ser:er R0PORT the + rt # <hich the 0ate<ay Ser:er is ru##i#; RENTERPRISE the #a"e ! the %1 ;ical( Siebel E#ter+rise Ser:er RAPPSERBER the #a"e ! the %1 ;ical( Siebel A++licati # Ser:er RS?SER > the Siebel a++licati # user t be used t ru# the EIM +r cess u#der RSPASS the +ass< rd ! the a! re"e#ti #ed user RIFB the !ull +ath a#d !ile #a"e ! the IFB !ile t be used RBATC& the batch r batch ra#;e t be +r cessed by this EIM +r cess %the use ! batch ra#;es "i#i"iMes the :erheads ass ciated <ith starti#; a# EIM +r cess( RPNAME the #a"e ! the EIM Pr cess <ithi# the IFB !ile t i#: /e

As EIM #a"e>:alue +ara"eter +airs a#d e)te#ded +ara"eters "ay #eed t be c #tr lled # a# ad>h c basisA the# the calli#; +r cess sh uld be c #!i;urable i# the <ay i# <hich the c ""a#d li#e is structured a#d se#t t Ser:er Ma#a;er8 By i#cludi#; "ulti+le EIM +r cesses ! r a ;i:e# e#tity <ithi# a si#;le IFB !ileA the# c #!i;urati # "a#a;e"e#t ca# be c #tr lled at a# a++r +riate le:el8 This ;uideli#e sh uld be ad +ted8

Pa;e *5 ! 7= IN CONFIDENCE

Siebel EIM Best Practice IN CONFIDENCE

3.3 )chedulin!
This secti # discusses s "e !act rs e!!ecti#; the ru##i#; ! EIM as ! ll <s$ As EIM +erates # batches ! dataA the# use ! database ser:er res urces %i# ter"s ! red A archi:i#; a#d u#d ( ca# be +ti"ised by use ! di!!ere#t batch siMes8 EIM ru#s as a si#;le threaded +r cess a;ai#st the database ser:erA there! re it ca# be ru# i# +arallel t i#crease thr u;h+ut8 Di!!ere#t scheduli#; +ti #s8

(%(% Degree of ParallelismALayout


The 0S1 su++ rts the ru##i#; ! EIM batches i# +arallel8 This "ea#s that c #curre#t EIM strea"s are !ired !! at the sa"e ti"eA < r/i#; a;ai#st di!!ere#t ra#;es ! batches si"ulta#e usly8 Duri#; B lu"e ' Per! r"a#ce Testi#; deter"i#ed the +ti"al de;ree ! +arallelis" by :aryi#; the #u"ber ! c #curre#t strea"s a#d a#alysi#; thr u;h+ut8 A li"it c "es ab ut <he# the competition ! r IGO results i# a de;raded thr u;h ut as " re a#d " re strea"s are added8 I! a @ b is " re CP? b u#d tha# IO b u#dA the# +arallelis" ca# ;e#erally be i#creased %e8;8 t 3> <ay(8 By deter"i#i#; the +ti"al de;ree ! +arallelis"A the# better EIM thr u;h+ut ca# be achie:ed alth u;h c "+eti#; +r cesses als #eed t be catered ! r8

Per! r"a#ce Testi#; sh uld :ary the 1ay ut t achie:e the hi;hest thr u;h+ut8

(%(%# $atchsiBe
Si"ilarly t 1ay utA EIMDs best thr u;h+ut ca# be ! u#d by :aryi#; batchsiMe duri#; testi#;8 F r B lu"eA the +ti"al batchsiMe is ;e#erally 14A444 rec rds +er batch alth u;h s"aller batches "ust be used ! r Mer;e a#d Delete @ bs t reduce the ris/ ! causi#; c "+eti#; +r cess t !ail %e8;8 SS#a+sh t t ldT err rs(8 Such a lar;e +ti"al batchsiMe is + ssible <he# the bu!!er cache is lar;e8

Per! r"a#ce Testi#; sh uld :ary the BatchsiMe t achie:e the hi;hest thr u;h+ut8

(%(%( EIM 1treaming )onfigurations


2hilst C #:er;e deter"i#ed that <he# s+a<#i#; " re tha# - EIM strea" this #eeded t be s+read acr ss "ulti+le Siebel A++licati # Ser:ersA such beha:i ur has yet t be e)hibited # B lu"e8 B lu"e ru#s all EIM +r cesses !r " a si#;le Siebel A++licati # Ser:er8

Pa;e *3 ! 7= IN CONFIDENCE

Siebel EIM Best Practice IN CONFIDENCE

B lu"e 'Ds Per! r"a#ce <as c #siste#t i# deter"i#i#; that IGO # the database ser:er <as the si#;le bi;;est b ttle#ec/ %! r l ads i#t tables # t + +ulatedA the si#;le bi;;est b ttle#ec/ is ;e#erally <ith CP?(8 This there! re did # t ;i:e rise t +er! r"i#; testi#; the s+a<#i#; ! EIM !r " "ulti+le Siebel A++licati # Ser:ers8 Per! r"a#ce Testi#; sh uld deter"i#e <hether i#creasi#; the #u"ber ! +hysical Siebel A++licati # Ser:ers e#ha#ces EIM +er! r"a#ce8 I! this beha:i ur is # t :eri!iedA the "ulti>ser:er EIM strea"i#; c #!i;urati # sh uld # t be ad +ted8

(%(%* )ontrolCM 1cheduling


I# B lu"eA the curre#t schedule %"ai#tai#ed a#d +erated !r " C #tr l>M( is based u+ # 7 strea"s a#d l /s s "ethi#; li/e this$

Strea" 1 Pr cess 1

Strea" * Pr cess '

Strea" 7

Pr cess *

Pr cess -

?#l ad E#tity 1 ?#l ad E#tity * ?#l ad E#tity #

Pr cess 7

Pr cess #

Pr cess #

I# this scheduleA EIM table l ads ha++e# at the tail e#d ! the #>li#e day %3+" #<ards( this has # t yet caused a#y +er! r"a#ce i"+act u+ # the #>li#e +r cesses8 I# the e)cer+t ab :eA "ulti+le strea"s ru# i# the schedule$ Strea" 1 ru#s EIM +r cesses %each ! <hich is ru##i#; i# +arallel(8 There are de+e#de#cies bet<ee# the +r cesses bet<ee# the strea"s Strea" * ru#s EIM +r cesses as i# Strea" 18 Strea" 7 ru#s u#l ads <ith de+e#de#cies <ith the EIM +r cesses i# the ther t< strea"s a table is#Dt u#l aded u#til all EIM +r cessi#; a;ai#st that table is c "+lete %SJOR0JEKT isA there! reA u#l aded last i# Strea" 7 as it is accessed by most EIM @ bs(8

I# this a++r ach @ bs that +r cess di!!ere#t EIM a#d Base tables ca# be ru# si"ulta#e uslyA <ith u#l ads ; i#; # i# +arallel de+e#de#cies bet<ee# EIM strea"s a#d u#l ads "ust be e#! rced8 This "a)i"ises use ! the batch <i#d <A as l #; as the database ser:er ca# c +e <ith the de"a#ds +ut u+ # it8

Pa;e *6 ! 7= IN CONFIDENCE

Siebel EIM Best Practice IN CONFIDENCE

2here a +r cess is deter"i#ed t ru# i# is lati # <ith "a)i"u" thr u;h+ut at a certai# de;ree ! +arallelis" duri#; testi#;A this "ay #eed t be reduced i! ru# i# +arallel <ith ther EIM +r cesses a#d al #;side u#l ads8

Pa;e 74 ! 7= IN CONFIDENCE

Siebel EIM Best Practice IN CONFIDENCE

3.6 EIM "a(le Maintenance


Due t the #u"ber ! +erati #s e)ecuted a;ai#st the EIM tablesA the# the ! ll <i#; table "ai#te#a#ce is rec ""e#ded$ 'runcate EIM tables re;ularly This has "a#y ad:a#ta;es i#cludi#; releasi#; u#used st ra;e a#d re" :i#; !ra;"e#tati #8 Analyse EIM tables a!ter each table l ad As EIM e)ecutes "a#y S,1 state"e#ts i# C st Based M deA the# table %and inde*( statistics assist Oracle i# ch si#; a# a++r +riate e)ecuti # +ath8 Re;ular statistics ;atheri#; # the base tables sh uld als be d #e ! r the sa"e reas #s8 1torage Parameters ;i:e# that R < Chai#i#; ca# ccur # s "e EIM tablesA the# the +r @ect DBAs sh uld c #sider setti#; a hi;h PCTFree :alue8

Pa;e 71 ! 7= IN CONFIDENCE

Siebel EIM Best Practice IN CONFIDENCE

* Dolume * Lessons Learned


This secti # utli#es s "e ! the /ey less #s lear#ed duri#; Per! r"a#ce Test a#d De+l y"e#t ! B lu"e ' as <ell as !r " a# e)ercise t :erhaul all EIM i#ter!aces i# Pr ducti # # B lu"e8 Cha#;es "ade t the ODM Ab I#iti c "+ #e#ts ha:e already bee# i#te;rated i#t B lu"e = a#d AMP &i;h 1e:el Desi;#s a#d s are #ly "e#ti #ed here ! r c "+lete#ess8

6.1 Enhancements to EIM Processes


This secti # c #tai#s s "e ! the e#ha#ce"e#ts "ade t EIM Pr cesses that sh uld be ad +ted by subseCue#t BT Siebel +r @ects8

*% % Redundant .pdates
Re" :al ! redu#da#t +erati #s <ithi# the EIM +r cess is the si#;le bi;;est ++ rtu#ity %<ith + ssible e)ce+ti # ! e#suri#; best + ssible s rti#;( t "a/e +er! r"a#ce i"+r :e"e#ts This is best illustrated by e)a"+le8 The ri;i#al ?+dateJB lu"eJC #su"er +r cess l /ed si"ilar t that sh <# bel <$

[U!"ate#$ol%&e#Con9%&er] TY(E = IM(ORT 'ATC) = *'atc+Range TA'LE = EIM#ACCOUNT ONLY 'ASE TA'LES = S#(ARTY, S#ORG#E-T, S#ORG#'U, S#ACCNT#(OSTN INSERT RO.S = S#(ARTY, FALSE INSERT RO.S = S#ORG#E-T, FALSE INSERT RO.S = S#ORG#'U, FALSE INSERT RO.S = S#ACCNT#(OSTN, FALSE NET C)ANGE = FALSE ONLY 'ASE COLUMNS = S#(ARTY0(ARTY#UI/, 1 S#(ARTY0(ARTY#TY(E#C/, 1 S#(ARTY0NAME, 1 S#(ARTY0ROOT#(ARTY#FLG, 1 S#(ARTY0GROU(#TY(E#C/, 1 S#ORG#E-T0NAME, 1 S#ORG#E-T0LOC, 1 S#ORG#E-T0'U#I/, 1 S#ORG#E-T0(AR#RO.#I/, 1 S#ORG#E-T0ACCNT#FLG, 1 S#ORG#E-T0ACTI$E#FLG, 1 000 S#ORG#E-T0ACCNT#TY(E#C/, 1 S#ORG#E-T0OU#TY(E#C/, 1 S#ORG#E-T0INTEGRATION#I/, 1 S#ORG#E-T0OU#NUM, 1 S#ORG#E-T0MASTER#OU#I/, 1 S#ORG#E-T0'ASE#CURCY#C/, 1 S#ORG#'U0ORG#I/, 1 S#ORG#'U0ORG#LOC, 1 S#ORG#'U0ORG#NAME, 1 S#ORG#'U0'U#I/, 1 S#ORG#E-T0CURR#(RI#LST#I/, 1 S#ACCNT#(OSTN0OU#E-T#I/, 1 S#ACCNT#(OSTN0(OSITION#I/, 1 S#ACCNT#(OSTN0RO.#STATUS, 1 S#ORG#E-T0(R#(OSTN#I/, 1 S#ORG#E-T0-#'T#FOLLO.#U(#RATING, 1

Pa;e 7* ! 7= IN CONFIDENCE

Siebel EIM Best Practice IN CONFIDENCE

The ! ll <i#; redu#da#t u+dates <ere re" :ed$ SJOR0JB?A SJACCNTJPOSTN these tables are + +ulated <ith de!ault relati #shi+s by the ada+ter8 As this data #e:er cha#;esA the# redu#da#t u+dates t these lar;e tables <ere re" :edA a lar;e i"+r :e"e#t <as "ade8 This <as d #e by re" :i#; the a! re"e#ti #ed tables !r " the ON1. BASE TAB1ES +ara"eter8 SJOR0JEKT8MASTERJO?JID as the C #su"erDs Master acc u#t is itsel! %a sel!>re!ere#ce(A a#d this ca# #e:er cha#;eA the# the MasterJOuJId c lu"# <as able t be re" :ed8 Bei#; a F rei;# Fey t SJOR0JEKTA ;ai#s i# the e#d t e#d EIM ru#ti"e ! r ?+dateJB lu"eJC #su"er <ill be able t be al" st hal:ed %this cha#;e is yet t be i"+le"e#ted i# Pr ducti #(8 This <as d #e by re" :i#; SJOR0JEKT8MASTERJO?JID !r " the ON1. BASE CO1?MNS +ara"eter8

The ! ll <i#; i#de) cha#;es <ere "ade$ Siebel <as ! u#d t hi#t the i#a++r +riate SJOR0JEKTJ?* i#de) %led by B?JID <hich is # #>selecti:e i# B lu"e( t res l:e relati #shi+s t SJOR0JEKT8 By re" :i#; i#de) hi#ts %a#d i#: /i#; CBO(A Oracle <as able t start usi#; the " re a++r +riate SJOR0JEKTJ?1 i#de) %led by #a"e(8 A "ar;i#al ;ai# <as realised thr u;h this cha#;e8 A #e< i#de) has bee# desi;#ed that <ill !urther i"+r :e this i#ter!ace8 T< i#de)es <ere als ide#ti!ied t reduce IO8 Detailed Oracle c #!i;urati # is # t i#cluded <ithi# this d cu"e#t8

O#ce all these cha#;es ha:e bee# i"+le"e#ted i# li:eA the# this +r cessDs ty+ical thr u;h+ut ! r bul/ u+date +erati #s <ill ha:e i#creased se:eral ! lds8 E#sure that u+date +r cesses d # t u+date tables r res l:e ! rei;# /eys u##ecessarily8

*% %# 1plitting .pdates
A# ther <ay t yield bi; ;ai#s i# t s+lit u+date rec rds ut i#t "ulti+le u+date +r cesses as ++ sed t a si#;le u+date +r cess8 All %+re:i us( ODM EIM u+date @ bs si"+ly detected that s "ethi#; <ithi# the rec rd had cha#;ed a#d +r :ided a !ull u+date rec rd t EIM8 It is i#e!!icie#t t res l:e ! rei;# /eys c lu"#s that ha:e # t cha#;ed redu#da#t u+dates8 Cha#;es t the ?+dateJB lu"eJAsset are curre#tly i# +r ;ress # B lu"e a#d this ser:es as a ; d <ay t illustrate the c #ce+t ! s+lit u+dates$ The ri;i#al Asset ?+date +r cess l /ed li/e this$

[U!"ate#$ol%&e#A99et] TY(E = IM(ORT TA'LE = EIM#ASSET 'ATC) = *'atc+Range USE IN/E- )INTS = FALSE USE ESSENTIAL IN/E- )INTS = FALSE SESSION S5L = 8Alter 9e99ion 9et o!ti&i:er#&o"e = all#ro;98 ONLY 'ASE TA'LES = S#ASSET INSERT RO.S = S#ASSET, FALSE NET C)ANGE = FALSE ONLY 'ASE COLUMNS = S#ASSET0ASSET#NUM, 1 Pa;e 77 ! 7= IN CONFIDENCE

Siebel EIM Best Practice IN CONFIDENCE

S#ASSET0(RO/#I/, 1 S#ASSET0'U#I/, 1 S#ASSET0AUTO#SR#FLG, 1 S#ASSET0COM(#UN/#.RNTY#FLG, 1 S#ASSET0(ER#A//R#I/, 1 S#ASSET0(AR#ASSET#I/, 1 S#ASSET0ROOT#ASSET#I/, 1 S#ASSET0'ILL#ACCNT#I/, 1 S#ASSET0-#'T#)IERARC)Y#C/ E)a"i#ati # re:eals that there are ' FFs t hea:ily + +ulated tables that are res l:ed by this u+date +r cessA #a"elyA PERJADDRJID %FF t SJADDRJPER(A PARJASSETJID a#d ROOTJASSETJID %FFs t SJASSET( a#d BI11JACCNTJID %FF t SJOR0JEKT(8 Fu#cti #al a#alysis deter"i#ed that the Asset &ierarchy < uld never cha#;e a#d that Bill Acc u#t a#d Address < uld #ly e:er rarel. cha#;e8 By drilli#; d <# i# the ODM TrueDi!! +r cess t +r duce se+arate i#di:idual ut+uts i# the e:e#t that #e ! these FFs actually d es cha#;eA the# +er! r"a#ce ca# be dra"atically i"+r :ed8 The result is that a si#;le Asset ?+date +r cess is bei#; s+lit t +r duce the ! ll <i#; EIM u+date +r cesses$ E.pdate:Dolume:AssetF this +r cess is used <here #ly attributes ha:e cha#;ed8 The ' FFs <ere re" :ed !r " this +r cess8 E.pdate:Dolume:Asset:AddressF s+eci!ically desi;#ed ! r B lu"e - <here a lar;e #u"ber ! AssetDs addresses < uld cha#;eA this +r cess #ly u+dates Asset Attributes a#d Asset Address8 The Asset &ierarchy a#d Bill Acc u#t FFs ha:e bee# re" :ed !r " this +r cess8 E.pdate:Dolume:Asset:AllF as it <as deter"i#ed that Asset &ierarchy sh uld #e:er cha#;e a#d that Bill Acc u#t < uld #ly cha#;e rarelyA these cha#;es <ere ;r u+ed t ;ether8 Due t the <ay i# <hich the TrueDi!! c de < r/s <ith s+lit u+datesA the# this +r cess < uld u+date Asset Attributes a#d all FFs %read #(8

The Ab I#iti c de ! r the TrueDi!! ?+date s+lits < r/s as ! ll <s$ 18 I! there are cha#;es t Asset AttributesA the# the record is !la;;ed ! r U?+dateJB lu"eJAssetV8 *8 I! there is a cha#;e t the Asset AddressA the# the record is !la;;ed ! r U?+dateJB lu"eJAssetJAddressV8 By !la;;i#; the rec rd ! r this +r cessA a#y !la; that <as set +re:i usly is :er><ritte#8 78 I! there is a cha#;e i# the Asset &ierarchy r the Bill Acc u#tA the# the record is !la;;ed ! r the U?+dateJB lu"eJAssetJAllV +r cess8 A#y +re:i us !la; that <as set is :er><ritte#8 I# this case all attributes a#d FFs are u+dated8 I# ;e#eral EIM u+date +er! r"a#ce ca# be i"+r :ed by #ly u+dati#; FFs that actually #eed t be u+dated8 This is +articularly +re:ale#t <he# the ! rei;# /ey is t a hea:ily l aded Base table such as SJOR0JEKT8

S+lit ut u+date +r cesses i#t se+arate +r cesses s that ! rei;# /eys are # t res l:ed u##ecessarily8

Pa;e 7' ! 7= IN CONFIDENCE

Siebel EIM Best Practice IN CONFIDENCE

*% %( Hints and !ptimiser Mode


1 ads i#t the already hea:ily + +ulated SJOR0JEKT table <ere i"+r :ed by tur#i#; ! i#de) hi#ts a#d alteri#; the sessi #Ds database +ti"iser t C st Based " de %!r " the de!ault ! Rules based($ USE IN/E- )INTS = FALSE USE ESSENTIAL IN/E- )INTS = FALSE SESSION S5L = 8Alter 9e99ion 9et o!ti&i:er#&o"e = all#ro;98 This tech#iCue <as used s that Oracle < uld +ic/ a cust " i#de) %described i# the #e)t secti #( usi#; RBO t i"+r :e thr u;h+ut8 As O#eSiebel has +ted t use Siebel 585A the# e)ecuti # +ath +ti"isati # <ill < r/ di!!ere#tly8 The Per! r"a#ce Test tea" <ill i#:esti;ate this8

A#alyse EIM +er! r"a#ce usi#; di!!ere#t Oracle +ti"iMer " des <ith a#d <ith ut Siebel ;e#erated i#de) hi#ts8

6.2 Data(ase Con i!uration


*%#% Inde,es
M st ! the ?ser Fey i#de)es i# Siebel d # t c #tai# the RO2JID c lu"# "ea#i#; that * IGOs are !te# reCuired <here @ust #e c uld su!!ice8 As IGO <as c #siste#tly the "a@ r c #tributi#; !act r t less>tha#>desired thr u;h+utA the# this tech#iCue ca# be used t reduce IGO i#te#si:e +erati #s8 e8;8 U(/ATE 9iebel0EIM#ACCOUNT IT SET 000 T#ORG#E-T#MASTER = >SELECT MIN>'T0RO.#I/? FROM 9iebel0S#ORG#E-T 'T .)ERE >'T0'U#I/ = IT0MASTER#ACCNT#'I AN/ 'T0NAME = IT0MASTER#ACCNT#NAME AN/ 'T0LOC = IT0MASTER#ACCNT#LOC??, 000 .)ERE >IF#RO.#'ATC)#NUM = 32 AN/ IF#RO.#STAT#NUM = < AN/ T#ORG#E-T##STA = <?K I# the S,1 ab :eA the SJOR0JEKTJ?1 i#de) %S#ORG#E-T>NAME, LOC, 'U#I/, CONFLICT#I/?( is ch se# by the C st Based O+ti"iMer8 This is a# i"+r :e"e#t # the de!ault Siebel e)ecuti # +ath <here the SJOR0JEKTJ?* i#de) %S#ORG#E-T>'U#I/, NAME, LOC, 'U#I/, CONFLICT#I/?( is hi#tedA but still # t ; d e# u;h8 I# this situati # a# IO "ust be "ade t the SJOR0JEKTJ?1 i#de) i# rder t l cate the "atchi#; r < i# Oracle8 The i#de) retur#s the i#ter#al Oracle R <Id ! the r < i# Cuesti #A but Pa;e 7- ! 7= IN CONFIDENCE

Siebel EIM Best Practice IN CONFIDENCE

the# has t d a# ther IO # the SJOR0JEKT table t ;et the Siebel R <Jid reCuired by the state"e#t8 By creati#; a cust " i#de) # S#ORG#E-T>NAME, LOC, 'U#I/, RO.#I/?A the# the Cuery is +ti"ised as it #eed #ly "a/e a si#;le IO # SJOR0JEKT a#d its i#de)es8 Als A as this #e< i#de) si"+ly :erlays a# e)isti#; Siebel ?* i#de) %but re+laces the u##ecessary CONF1ICTJID <ith the " re use!ul RO2JID( the ris/ t ab# r"al 0?I +erati # is :ery l <8

A#alyse l #; ru##i#; EIM S,1 state"e#ts a#d addGre" :e i#de)es as a++r +riate8 E)+erie#ced DBAs are reCuired t assist <ith this tas/8 Full re;ressi # testi#; is reCuired ! r a#y cha#;es t i#de)es8 I#de)es sh uld #ly be altered a!ter :eri!icati # by e)+erie#ced DBAs8

6.3 +DM $dapter #unctionalit.


ODM e#ha#ce"e#ts ha:e already bee# i#c r+ rated i#t O#eBie< = a#d O#eSiebel ODM &i;h 1e:el Desi;#s a#d are # t discussed i# detail here8 Se:eral +er! r"a#ce e#ha#ce"e#ts <ere "ade t the ODM Ab I#iti ada+ters duri#; Release ' de+l y"e#t8 These cha#;es <ere as ! ll <s$ 1orting by s rti#; the data at the i#+ut ! the ada+ter %be! re a#y thr ttle !u#cti #ality( as <ell as <ithi# the 0S1 itsel!A the# the data <as able t be s rted as best as + ssible8 Data s rti#; <as the si#;le bi;;est <i# %<ith the e)ce+ti # ! re" :i#; redu#da#t u+dates( duri#; Release '8 1erial $atch )alculation alth u;h creati#; batches <ithi# the ada+ter serially as ++ sed t i# +arallel i#creases +r cessi#; ti"e i# the Ab I#iti e#:ir #"e#tA the ;ai#s bser:ed <ithi# EIM are < rth the sacri!ice8 These ;ai#s are a result the i"+r :ed s rti#; that serial calculati # +r :ides8 Filter before throttling the !ilter !u#cti #ality <as used s that thr ttled data #e #i;ht < uld continue where it le t o the +re:i us #i;ht8 This "ea#t that the data <as s rted " re ti;htly8 W E#d ! D cu"e#t X

Pa;e 7= ! 7= IN CONFIDENCE