Adaptatywne serwery WWW

Marek Wojciechowski, Maciej Zakrzewicz

3ROLWHFKQLND3R]QDVND,QVW\WXW,QIRUPDW\NL

ul.

3LRWURZRD3R]QD

e-mail: {marek,mzakrz}@cs.put.poznan.pl

Abstrakt.

$GDSWDW\ZQH VHUZHU\ ::: Z\NRU]\VWXM DQDOL]  SOLNyZ ORJX Z FHOX DXWRPDW\F]QHM

WUDQVIRUPDFML]DZDUWRFLLVWUXNWXU\XGRVW SQLDQ\FKGRNXPHQWyZ:UH]XOWDFLHVHUZHU:::VDPRG]LHOQLH

ÄGRSDVRZXMH´VL GRRF]HNLZDX*\WNRZQLNDÄRGJDGXMF´MHJRLQWHQFMH:DUW\NXOHSU]HGVWDZLRQRGRVW SQH
metody zautomatyzowanej analizy plików logu oraz stosowania znalezionych trendów i korelacji w
dynamicznej transformacji dokumentów WWW..

1. Wprowadzenie

3URMHNWRZDQLH VWUXNWXU\ ]DZDUWRFL VHUZHUD ::: MHVW Z RJyOQRFL SUREOHPHP ]áR*RQ\P L

WUXGQ\P 3URMHNWDQFL SRGHMPXM ]DGDQLH WDNLHJR RSUDFRZDQLD Z\JOGX L SRZL]D SRPL G]\

GRNXPHQWDPL:::DE\E\á\RQHF]\WHOQHLáDWZHZQDZLJDFMLGODX*\WNRZQLNyZ1DSUREOHP

SURMHNWRZDQLDVWUXNWXU\]DZDUWRFLVHUZHUD:::PDMZSá\ZQDVW SXMFHF]\QQLNL

1.

5y*QL X*\WNRZQLF\PRJNRU]\VWDü]VHUZHUD:::FHOX]QDOH]LHQLDinnych informacji.
3U]\NáDGRZR MH*HOL GR VNOHSX LQWHUQHWRZHJR SU]\áF]D VL  X*\WNRZQLN PáRG\ WR GREU]H

E\áRE\DE\VWURQDJáyZQD]DZLHUDáDLQIRUPDFMHRQDMQRZV]\FKJUDFKNRPSXWHURZ\FK.LHG\

MHGQDN SU]HVáDQLD VWURQ\ JáyZQHM ]D*GD X*\WNRZQLN VWDUV]\ ZWHG\ ZVND]DQH E\áRE\

XPLHV]F]HQLH QD QLHM LQIRUPDFML R QDMQRZV]\FK SR]\FMDFK NVL*NRZ\FK L Sá\WDFK ] PX]\N

NODV\F]Q

2. W

Uy*Q\FK momentach czasowychMHGHQX*\WNRZQLNPR*HSRV]XNLZDüinnych informacji.

3U]\NáDGRZR X*\WNRZQLN LQWHUQHWRZHJR ELXUD SRGUy*\ E G]LH Z RNUHVLH ]LPRZ\P

]DLQWHUHVRZDQ\ GRNXPHQWDPL ::: ]DZLHUDMF\PL LQIRUPDFMH R NXURUWDFK QDUFLDUVNLFK Z

$OSDFK QDWRPLDVW Z RNUHVLH OHWQLP WHQ VDP X*\WNRZQLN *\F]\áE\ VRELH SUH]HQWDFML

GRNXPHQWyZ:::RSLVXMF\FKZF]DV\ZEDVHQLH0RU]DUyG]LHPQHJR

3.

=DZDUWRüVHUZHUD:::rozrasta VL ZUD]]XSá\ZHPF]DVX
*G\ GR LVWQLHMFHJR V\VWHPX GRGDZDQH V QRZH GRNXPHQW\ ::: SURMHNWDQW PXVL SRGMü

RGSRZLHG]LDOQGHF\]M RW\PZNWyU\FK]GRW\FKF]DVRZ\FKGRNXPHQWyZXPLHFLüáF]QLNL

GRQRZHMF] FLV\VWHPX

5R]ZL]DQLHP WHFKQLF]Q\P NWyUH DGUHVXMH SU]HGVWDZLRQH SUREOHP\ MHVW personalizacja

]DZDUWRFL VHUZHUyZ ::: 3HUVRQDOL]DFMD SROHJD QD Z\NRU]\VW\ZDQLX ZLHG]\ R SUHIHUHQFMDFK

X*\WNRZQLNyZGRG\QDPLF]QHJRGRVWRVRZ\ZDQLDZ\JOGXLVWUXNWXU\SU]HV\áDQ\FKGRNXPHQWyZ

']L NLWHPXND*G\X*\WNRZQLNPR*HRWU]\P\ZDüLQQ\REUD]]DZDUWRFLLVWUXNWXU\GRNXPHQWyZ

WHJR VDPHJR VHUZHUD :LHG]D R SUHIHUHQFMDFK X*\WNRZQLNyZ PR*H E\ü SR]\VNLZDQD jawnie,

SRSU]H] GRVWDUF]HQLH X*\WNRZQLNRP IRUPXODU]\ L QDU] G]L R FKDUDNWHU]H NRQILJXUDF\MQ\P EG(
niejawnie – w wyniku obserwacji stylu ich dotychczasowej interakcji z serwerem. Wiele

VWRVRZDQ\FKG]LUR]ZL]DZ]DNUHVLHSHUVRQDOL]DFML]DZDUWRFLVHUZHUyZ:::VLOQLHED]XMHQD

LQIRUPDFMDFK X]\VNDQ\FK RG X*\WNRZQLND Z VSRVyE MDZQ\ 7DND IRUPD SR]\VNLZDQLD ZLHG]\

FHFKXMH VL  GX* VXELHNW\ZQRFL L VSRW\ND VL  ] QLHFK FL X*\WNRZQLNyZ Ä]PXV]DQ\FK´ GR

Z\SHáQLDQLD GRGDWNRZ\FK IRUPXODU]\ L DQNLHW 3RQDGWR WDN EXGRZDQH SURILOH X*\WNRZQLNyZ

SRVLDGDMFKDUDNWHUVWDW\F]Q\L]XSá\ZHPF]DVXXOHJDMGHJUDGDFML

:RVWDWQLFKODWDFKFRUD]ZL NV]XZDJ SU]\FLJDMPHWRG\SHUVRQDOL]DFML]DZDUWRFLVHUZHUyZ

:::SRSU]H]QLHMDZQHREVHUZRZDQLHWUHQGyZZ]DFKRZDQLDFKX*\WNRZQLNyZ::::SUDF\

30

Marek Wojciechowski, Maciej Zakrzewicz

[PE97], zaproponowano termin adaptatywne serwery WWW (adaptive web

VLWHVRSLVXMF\serwery

::: NWyUH DXWRPDW\F]QLH XOHSV]DM VZRM ]DZDUWRü L RUJDQL]DFM  QD SRGVWDZLH REVHUZDFML

FLH*HN GRVW SyZ X*\WNRZQLNyZ. Idea adaptatywnych serwerów polega na analizie plików logu

VHUZHUD Z\áDZLDQLX ] QLFK VWDW\VW\F]Q\FK NRUHODFML SRPL G]\ SRELHUDQ\PL GRNXPHQWDPL OXE

SUDFXMF\PL X*\WNRZQLNDPL D QDVW SQLH Z\NRU]\VW\ZDQLX ]QDOH]LRQ\FK NRUHODFML GR EXGRZ\

VWUXNWXU\GRNXPHQWyZ:::Z\V\áDQ\FKX*\WNRZQLNRP:W\PDUW\NXOHRSLVXMHP\VWDQQDXNLL
technologii w zakresie metod konstrukcji adaptatywnych serwerów WWW.

2. Automatyczna adaptacja serwera WWW

Proces adaptacji serwera WWW przebiega w dwóch fazach:

1.

2IIOLQHZ\NRU]\VWDQLHSOLNXORJXVHUZHUDGR]QDOH]LHQLDLSRJUXSRZDQLDQDMF] VWV]\FKFLH*HN

QDZLJDF\MQ\FK X*\WNRZQLNyZ )D]D WD UHDOL]RZDQD MHVW DV\QFKURQLF]QLH Z]JO GHP SRáF]H

X*\WNRZQLNyZQSZRGVW SDFKW\JRGQLRZ\FKOXEPLHVL F]Q\FK

2.

2QOLQHZ\NRU]\VW\ZDQLH]QDOH]LRQ\FKJUXSFLH*HNQDZLJDF\MQ\FKGRWZRU]HQLDdynamicznych
rekomendacji

 GOD X*\WNRZQLNyZ F]\OL ]ELRUX áF]QLNyZ GR GRNXPHQWyZ NWyU\PL FL

X*\WNRZQLF\E GQDMSUDZGRSRGREQLHMVWDW\VW\F]QLH]DLQWHUHVRZDQL)D]DWDMHVWUHDOL]RZDQD

SRGF]DVREVáXJLND*GHJR*GDQLDX*\WNRZQLND
3U]HGVWDZP\ SU]\NáDG SURVWHM DGDSWDFML VHUZHUD ::: ]LOXVWURZDQ\ QD U\VXQNX  6HUZHU

:::]RVWDáRGZLHG]RQ\SU]H]SL FLXX*\WNRZQLNyZNWyU\FKSHáQHFLH*NLQDZLJDF\MQH]RVWDá\
zapisane w pliku logu. W pierwszej fazie adaptacji (offline) wykonywana jest analiza pliku logu i

]QDOH]LRQH ]RVWDM QDVW SXMFH FLH*NL F] VWH

books.html -> albums.html, books.html ->

ord.html, car.html -> radio.html

.D*GD]W\FKFLH*HNSRMDZLáDVL ZRGZLHG]LQRSLVDQ\FK

Z SOLNX ORJX L Z ]ZL]NX ] W\P E G RQH WUDNWRZDQH SU]H] QDV MDNR SUHIHUHQFMH GOD LQQ\FK

X*\WNRZQLNyZ:GUXJLHMID]LHRQOLQHQRZ\X*\WNRZQLNZ\V\áDGRVHUZHUD*GDQLHSU]HVáDQLD
dokumentu WWW (

books.html

 6HUZHU SRELHUD GRNXPHQW ] G\VNX L SU]HJOGD ]QDOH]LRQ\

ZF]HQLHM ]ELyU FLH*HN F] VW\FK ± Z\QLND ] QLHJR *H X*\WNRZQLF\ NWyU]\ SRELHUDOL GRNXPHQW

books.html

E\OLSy(QLHM]DLQWHUHVRZDQLGRNXPHQWDPL

albums.html

i

ord.html

:]ZL]NX]W\PZ

FHOX XáDWZLHQLD QDZLJDFML GR GRNXPHQWX

books.html

 G\QDPLF]QLH GRGDZDQH V áF]QLNL GR

SRZ\*V]\FKGRNXPHQWyZ7DN]PRG\ILNRZDQ\GRNXPHQWWUDILDGRX*\WNRZQLNDU\VXQHN

homepage.html -> books.html -> products.html -> albums.html -> ord.html
car.html -> books.html -> radio.html -> cdisks.html
books.html -> download.html -> ord.html -> albums.html -> contact.html
homepage.html -> car.html -> demo.html -> depts.html
car.html -> download.html -> radio.html
...

3OLNORJXVHUZHUD

=QDMGRZDQLHFLH*HN

F] VW\FK

books.html -> albums.html
books.html -> ord.html
car.html -> radio.html

3UHIHUHQFMH

books.html -> albums.html
books.html -> ord.html
car.html -> radio.html

3UHIHUHQFMH

)GDQLHX*\WNRZQLND

http://x.com/books.html

7UDQVIRUPDFMD

GRNXPHQWX

books.html

'RNXPHQW

books.html

áF]QLNGRalbums.html

áF]QLNGRord.html

=PRG\ILNRZDQ\GRNXPHQW

5\V3U]\NáDGRZ\SURFHVDGDSWDFMLVHUZHUD:::

Adaptatywne serwery WWW

31

'\QDPLF]QLHGRGDQHáF]QLNLGR

GRNXPHQWyZNWyU\PL

QDMSUDZGRSRGREQLHMMHVW

]DLQWHUHVRZDQ\X*\WNRZQLN

'RW\FKF]DVRZD]DZDUWRüGRNXPHQWX

5\V3U]\NáDG\GRNXPHQWyZ:::Z]ERJDFRQ\FKRG\QDPLF]QHUHNRPHQGDFMH

2.1 Faza Offline algorytmu

Struktura pliku logu

,QIRUPDFMH R GRVW SDFK GR VHUZHUD ::: ]DSLV\ZDQH V Z ORJX 'OD ND*GHJR GRVW SX GR

SRMHG\QF]HJRSOLNX]QDMGXMFHJRVL QDVHUZHU]HZORJXSRMDZLDVL QRZ\]DSLV-HGQDN*HLORü

LQIRUPDFML SDPL WDQD Z ]ZL]NX ] GDQ\P GRVW SHP PR*H E\ü Uy*QD Z SU]\SDGNX Uy*Q\FK

VHUZHUyZ ::: $E\ XPR*OLZLü WZRU]HQLH XQLZHUVDOQ\FK QDU] G]L VáX*F\FK GR DQDOL]\ ORJX

SRMDZLá\VL SUyE\VWDQGDU\]DFMLMHJRIRUPDWX']LVLDMPR*QD]DáR*\ü*HSU]HZD*DMFDZL NV]Rü

VHUZHUyZ ::: JHQHUXMH SOLNL ORJX ]JRGQH ] IRUPDWHP ]QDQ\P SRG QD]Z Common Logfile
Format

>/@1LHMHVWWRMHGQDNZSHáQLRERZL]XMF\VWDQGDUGJG\*QLHNWyUHVHUZHU\SDPL WDM

UyZQLH* SHZQH GRGDWNRZH LQIRUPDFMH VWDQGDUG XLF). Common Logfile Format SU]HZLGXMH *H

]DSLVZORJXSRZLQLHQPLHüQDVW SXMFSRVWDü

remotehost rfc931 authuser [date] ”request” status bytes

: SRZ\*V]\P IRUPDFLH SROH remotehost R]QDF]D QD]Z  OXE DGUHV ,3 NRPSXWHUD ] NWyUHJR

QDVWSLáR RGZRáDQLH 3ROH rfc931 ]DZLHUD QD]Z  X*\WNRZQLND QD GDQ\P NRPSXWHU]H ang.
logname). Pole authuser

 ]DZLHUD LQIRUPDFM  R W\P ]D NRJR X*\WNRZQLN VL  SRGDMH 3ROH >date]

LQIRUPXMHRW\PNLHG\QDVWSLáRRGZRáDQLHGDWDLF]DV3ROH´request´]DZLHUD*GDQLHSU]HVáDQH

GRVHUZHUDZWDNLHMIRUPLHZMDNLHMZ\JHQHURZDáMHNOLHQW2EHMPXMHRQRQDRJyáW\SRSHUDFMLL

QD]Z  SOLNX GR NWyUHJR QDVWSLáR RGZRáDQLH ZUD] ]H FLH*N GRVW SX 3ROH status zawiera

]ZUDFDQ\NOLHQWRZLNRGVWDWXVX]JRGQLH]SURWRNRáHP+773Z\NRU]\VW\ZDQ\PZXVáXG]H:::

'áXJRü ]DZDUWRFL SU]HV\áDQHJR GRNXPHQWX SDPL WDQD MHVW Z SROX bytes 3U]\NáDG ]DZDUWRFL
pliku logu serwera WWW przedstawiono na rysunku 3.

154.11.231.17 - - [13/Jul/2000:20:42:25 +0200] "GET / HTTP/1.1" 200 1673

154.11.231.17 - - [13/Jul/2000:20:42:25 +0200] "GET /apache_pb.gif HTTP/1.1" 200 2326

32

Marek Wojciechowski, Maciej Zakrzewicz

192.168.1.25 - - [13/Jul/2000:20:42:25 +0200] "GET /demo.html HTTP/1.1" 200 520

192.168.1.25 - - [13/Jul/2000:20:42:25 +0200] "GET /books.html HTTP/1.1" 200 3402

160.81.77.20 - - [13/Jul/2000:20:42:25 +0200] "GET / HTTP/1.1" 200 1673

154.11.231.17 - - [13/Jul/2000:20:42:25 +0200] "GET /car.html HTTP/1.1" 200 2580

192.168.1.25 - - [13/Jul/2000:20:42:25 +0200] "GET /cdisk.html HTTP/1.1" 200 3856

10.111.62.101 - - [13/Jul/2000:20:42:25 +0200] "GET /new/demo.html HTTP/1.1" 200 971

5\V3U]\NáDGRZ\SOLNORJXVHUZHUD:::

,GHQW\ILNDFMDFLH*HNQDZLJDF\MQ\FK
=SXQNWXZLG]HQLDDQDOL]\LVWRWQ\PLLQIRUPDFMDPLZORJXVHUZHUD:::VQD]ZDOXEDGUHV

,3 NRPSXWHUD ] NWyUHJR QDVWSLáR RGZRáDQLH QD]ZD X*\WNRZQLND GRNRQXMFHJR RGZRáDQLD

GRNáDGQDGDWDLF]DVRUD]SHáQDQD]ZDSOLNXNWyUHJRGRW\F]\áR*GDQLH$QDOL]DSOLNyZORJXSROHJD

QD ]QDMGRZDQLX F] VWR SRZWDU]DMF\FK VL  VHNZHQFML ZFLH*NDFK GRVW SyZ X*\WNRZQLNyZ GR

VHUZHUD ::: OXE QD JUXSRZDQLX X*\WNRZQLNyZ Z\ND]XMF\FK SRGREQH ]DFKRZDQLH = WHJR

SRZRGXNRQLHF]Q\PHWDSHPZVW SQHMREUyENLGDQ\FK]DZDUW\FKZORJXMHVWJUXSRZDQLH]DSLVyZ

GRW\F]F\FKRGZRáDWHJRVDPHJRX*\WNRZQLND*UXSRZDQLHWRRGE\ZDVL QDSRGVWDZLHDGUHVX,3

OXEQD]Z\NRPSXWHUDRUD]QD]Z\X*\WNRZQLND1LHVWHW\QLH]DZV]HQD]ZDX*\WNRZQLNDMHVW]QDQD

6\WXDFMDWDNDPDPLHMVFHF] VWRZSU]\SDGNXJG\X*\WNRZQLNNRU]\VWD]V\VWHPXRSHUDF\MQHJR

NWyU\ QLH ]DNáDGD ZLHORGRVW SX 1D V]F] FLH IDNW *H ] NRPSXWHUD SUDFXMFHJR SRG NRQWURO

V\VWHPX RSHUDF\MQHJR NWyU\ QLH MHVW ZLHORGRVW SQ\ PR*H Z GDQHM FKZLOL NRU]\VWDü W\ONR MHGHQ

X*\WNRZQLNSR]ZDODWUDNWRZDüRGZRáDQLDSRFKRG]FH]WHJRVDPHJRNRPSXWHUDMDNRRGZRáDQLD

MHGQHJRX*\WNRZQLNDJG\QD]ZDX*\WNRZQLNDQLHMHVW]QDQD2F]\ZLFLHSRZ\*V]H]DáR*HQLHMHVW

SRSUDZQH W\ONR Z SU]\SDGNX RGZRáD NWyU\FK F]DV\ ]DZLHUDM VL  ZRNUHVLH RGSRZLDGDMF\P

PR*OLZHPXF]DVRZLWUZDQLDSRMHG\QF]HMVHVMLX*\WNRZQLND0HFKDQL]PWHQQLHSR]ZDODZL FQD

LGHQW\ILNDFM VHNZHQFMLGRVW SyZZUDPDFKZLHOXVHVMLX*\WNRZQLNDQDSU]HVWU]HQLQSPLHVLFD

JG\*]GDQHJRNRPSXWHUDPR*HZUy*Q\FKJRG]LQDFKNRU]\VWDüZLHOHRVyE

=HZ]JO GXQDIDNW*HX*\WNRZQLNPR*HZLHORNURWQLHNRU]\VWDü]XVáXJGDQHJRVHUZHUD:::

]DND*G\PUD]HPV]XNDMFLQQ\FKLQIRUPDFMLQLHNLHG\ZVND]DQHMHVWUR]ELFLHVHNZHQFMLGRVW SyZ

GDQHJR X*\WNRZQLND QD IUDJPHQW\ RGSRZLDGDMFH SRV]F]HJyOQ\P VHVMRP 1LH MHVW WR MHGQDN

]DGDQLHWU\ZLDOQHJG\*SURWRNyáKWWSQLHSRVáXJXMHVL SRM FLHPVHVML1DMSURVWV]HUR]ZL]DQLHWHJR

SUREOHPXSROHJDQDZ\RGU EQLDQLXVHVMLX*\WNRZQLNyZZRSDUFLXR]DáR*HQLH*HMHOLF]DVPL G]\

NROHMQ\PL RGZRáDQLDPL GR VHUZHUD MHVW ]QDF]QLH GáX*V]\ QL* W\SRZ\ F]DV SU]HJOGDQLD MHGQHM

VWURQ\ WR RGZRáDQLD WH QDVWSLá\ Z UDPDFK GZyFK Uy*Q\FK VHVML $OWHUQDW\ZQ\P UR]ZL]DQLHP

PR*HE\üUR]V]HU]HQLHIXQNFMRQDOQRFLVHUZHUDRREVáXJ LGHQW\ILNDWRUyZVHVMLQDF]DV]ELHUDQLD

LQIRUPDFMLR]DFKRZDQLDFKX*\WNRZQLNyZ><-*@

&HOHPDQDOL]\SOLNyZORJXVHUZHUD:::PR*HE\ü]QDMGRZDQLHF] VW\FKFLH*HNQDZLJDFML

OXE ]QDMGRZDQLH JUXS VWURQ GR NWyU\FK X*\WNRZQLF\ F] VWR RGZRáXM VL  Z UDPDFK VHVML :

SLHUZV]\P SU]\SDGNX LVWRWQH V LQIRUPDFMH R ZV]\VWNLFK VWURQDFK GR NWyU\FK RGZRá\ZDá VL

X*\WNRZQLN]XZ]JO GQLHQLHPNROHMQRFLRGZRáD:SR]RVWDá\FKSU]\SDGNDFKPR*HVL RND]Dü

*H LVWRWQH V RGZRáDQLD W\ONR GR W\FK VWURQ NWyU\FK WUHü ]DLQWHUHVRZDáD X*\WNRZQLND VWURQ\

VáX*FHMHG\QLHMDNRFLH*NDGRVW SXGRV]XNDQHJRGRNXPHQWXQLHVXZ]JO GQLDQH:>&06@

]DSURSRQRZDQR SRG]LDá RGZRáD GR VWURQ QD ]RULHQWRZDQH QD ]DZDUWRü L zorientowane na

QDZLJDFM  1LHNWyUH VWURQ\ ]DZLHUDM JáyZQLH RGQRQLNL GR LQQ\FK VWURQ Z ]ZL]NX ] czym

RGZRáDQLD GR QLFK QD SHZQR E G PLDá\ FKDUDNWHU QDZLJDF\MQ\ -HGQDN*H ZLHOH VWURQ ]DZLHUD

]DUyZQRWUHüMDNLRGQRQLNLGRLQQ\FKVWURQ7DNLHVWURQ\PRJUy*Q\PX*\WNRZQLNRPVáX*\üGR

Uy*Q\FKFHOyZ'ODWHJRUR]VGQ\PNU\WHULXPSRG]LDáXGRVW SyZQD]RULHQWRZDQHQDQDZLJDFM L

]DZDUWRü Z\GDMH VL  F]DV QD MDNL X*\WNRZQLN ]DWU]\PXMH VL  QD GDQHM VWURQLH E\ü PR*H
znormalizowany w

VWRVXQNX GR UR]PLDUX VWURQ\ &]DV SU]HJOGDQLD GDQHM VWURQ\ MHVW REOLF]DQ\

Adaptatywne serwery WWW

33

MDNR Uy*QLFD HW\NLHW F]DVRZ\FK GZyFK NROHMQ\FK ]DSLVyZ Z ORJX RGSRZLDGDMF\FK QDVW SQHM L

ELH*FHM VWURQLH : SU]\SDGNX VWURQ NRF]F\FK VHVM  X*\WNRZQLND SU]\MPXMH VL  *H GRVW S GR

QLFKPLDáPLHMVFH]HZ]JO GXQDLFK]DZDUWRüFKRüRF]\ZLFLHZNRQNUHWQ\PSU]\SDGNXZFDOH

QLHPXVLWRE\üSUDZG

Problemy obróbki plików logu

,QIRUPDFMH]DZDUWHZORJXPRJE\üQLHW\ONRQLHSHáQHDOHUyZQLH*]DIDáV]RZDQH]HZ]JO GX

na wykorzystywanie serwerów

SUR[\LSRGU F]QHMSDPL FLSU]HJOGDUHN>3@6HUZHUSUR[\VáX*\

MDNR ÄRNQR QD ZLDW´ GOD ZLHOX NRPSXWHUyZ SR]ZDODMF X]\VNDü GRVW S GR ,QWHUQHWX

X*\WNRZQLNRP QD QLFK SUDFXMF\P =DSLV\ Z ORJX VHUZHUD ::: RGSRZLDGDMFH RGZRáDQLRP

X*\WNRZQLNyZNRPSXWHUyZÄXNU\W\FK´]DVHUZHUHPSUR[\VRSLVDQHDGUHVHPVHUZHUDproxy. W

]ZL]NX]W\PIDNW*HNLOND]DSLVyZZORJXGRW\F]\MHGQHJRDGUHVX,3QLHPXVLZFDOHR]QDF]Dü

L* ]DSLV\ WH RGSRZLDGDM RGZRáDQLRP ] WHJR VDPHJR NRPSXWHUD : >335@ ]DSURSRQRZDQR

PHWRG  Z\NU\ZDQLD WDNLFK V\WXDFML Z RSDUFLX R]DáR*HQLH *H MHOL GDQH RGZRáDQLH GRW\F]\

GRNXPHQWX GR NWyUHJR QLH PD áF]D ZSRSU]HGQLR *GDQ\P GRNXPHQFLH WR SUDZGRSRGREQLH

*GDQLD V NLHURZDQH SU]H] GZyFK Uy*Q\FK X*\WNRZQLNyZ 0LPR *H GRZLDGF]HQLD SRND]XM

>&3@ L* GRVW S GR NROHMQHJR GRNXPHQWX MHVW QDMF] FLHM Z\QLNLHP Z\EUDQLD GRVW SQHJR Z

GRNXPHQFLHáF]Dang. hyperlink) lub powrotem do poprzedniego dokumentu (operacja „Back”),

ZVSRPQLDQD PHWRGD QLH JZDUDQWXMH *DGQHM SHZQRFL 'ODWHJR GOD FHOyZ LGHQW\ILNDFML

X*\WNRZQLNyZ VWRVXMH VL  W]Z cookies OXE GRGDWNRZ DXWRU\]DFM  Cookie jest identyfikatorem

JHQHURZDQ\PSU]H]VHUZHULSU]HV\áDQ\PGRNOLHQWDSU]HJOGDUNLZFHOXSy(QLHMV]HMLGHQW\ILNDFML

X*\WNRZQLND1LHGRVNRQDáRüWHJRPHFKDQL]PXZ\QLND]IDNWX*HX*\WNRZQLF\PRJZGRZROQHM

FKZLOL XVXQü FRRNLH OXE Z RJyOH ]DEURQLü DNFHSWDFML cookies. Dodatkowa identyfikacja

X*\WNRZQLNyZSRSU]H]*GDQLHZ\SHáQLHQLDIRUPDWNLUHMHVWUDF\MQHMUyZQLH*Z\PDJDGREUHMZROL

X*\WNRZQLNyZJG\*GDQHSU]H]QLFKSRGDZDQHPRJE\üSU]HFLH*IDáV]\ZH

5yZQLH LVWRWQ\P SUREOHPHP MDN LGHQW\ILNDFMD X*\WNRZQLNyZ MHVW LGHQW\ILNDFMD IDNW\F]Q\FK

RGZRáD GR GRNXPHQWyZ =H Z]JO GX QD VWRVRZDQLH SU]H] SU]HJOGDUNL SDPL FL SRGU F]QHM

NROHMQHRGZRáDQLDGDQHJRX*\WNRZQLNDGRWHJRVDPHJRGRNXPHQWXPRJQLHE\üRGQRWRZDQHQD

VHUZHU]H JG\* PRJ E\ü ]UHDOL]RZDQH SU]H] VSURZDG]HQLH GRNXPHQWX ]SDPL FL SRGU F]QHM

SU]HJOGDUNL D QLH ] VHUZHUD : VSRVyE ]QDF]F\ PR*H WR ]DNáyFLü RGNU\WH FLH*NL QDZLJDFML

X*\WNRZQLNyZ -HV]F]H SRZD*QLHMV]\ SUREOHP Z\QLND ]H VWRVRZDQLD SDPL FL SRGU F]QHM SU]H]
serwery

SUR[\-HOLX*\WNRZQLNNRU]\VWDMF\] Internetu poprzez serwer SUR[\RGZRáXMHVL GR

GRNXPHQWX ]QDMGXMFHJR VL  Z SDPL FL SRGU F]QHM SUR[\ VHUZHU ::: PR*H E\ü Z RJyOH

QLHZLDGRP\ *H GDQ\ X*\WNRZQLN RGZRá\ZDá VL  GR GDQHJR GRNXPHQWX $E\ REURQLü VL  SU]HG

ZVSRPQLDQ\PL V\WXDFMDPL VHUZHU\ ::: PRJ VWRVRZDü WHFKQLNL ]DSRELHJDMFH

Z\NRU]\VW\ZDQLXSDPL FLSRGU F]QHMRNUHODQHMDNRcache-bustingSROHJDMFHQSQDSRGDZDQLX

GDW ] SU]HV]áRFL MDNR WHUPLQyZ XSá\QL FLD ZD*QRFL SRV]F]HJyOQ\FK GRNXPHQWyZ 7HJR W\SX

WHFKQLNLPRJE\üXFL*OLZHGODX*\WNRZQLNyZJG\*Z\GáX*DMF]DVRGSRZLHG]L=WHJRZ]JO GX

SRMDZLá\VL SURSR]\FMHDE\]DPLDVWPRQLWRURZDQLDZV]\VWNLFKGRVW SyZGRVHUZHUDRJUDQLF]\ü

VL W\ONRGRSHZQHMSUyENLVWDW\VW\F]QHMLQDMHMED]LHGRNRQ\ZDüDQDOL]

Czyszczenie plików logu

3URFHV ZVW SQHM REUyENL GDQ\FK QLH NRF]\ VL  QD LGHQW\ILNDFML RGZRáD SRV]F]HJyOQ\FK

X*\WNRZQLNyZ=DSLV\ZORJXGRW\F]SRMHG\QF]\FKSOLNyZDQLHGRNXPHQWyZWUDNWRZDQ\FKMDNR

RELHNW\]áR*RQH:SU]\SDGNXGRVW SXGRVWURQ\]DZLHUDMFHMQSREUD]\G(ZL NLOXEILOP\ZORJX

]QDMG]LHVL ]DSLVGRW\F]F\JáyZQHJRGRNXPHQWXQDMF] FLHM]UR]V]HU]HQLHPhtml lub htm), ale

WDN*H]DSLV\]ZL]DQH]HZV]\VWNLPLRELHNWDPL]DJQLH*G*RQ\PLZ stronie (obrazami, filmami, itp.).

1D V]F] FLH FKDUDNWHU SOLNX PR*QD Z GX*\P VWRSQLX Z\ZQLRVNRZDü ] MHJR UR]V]HU]HQLD

3U]\NáDGRZHUR]V]HU]HQLDQD]ZSOLNyZRGSRZLDGDMFHRELHNWRP]DJQLH*G*DQ\PZGRNXPHQWDFK
to jpg, jpeg, gif dla obrazów, au, wav

GODG(ZL NyZavi, movGODILOPyZ$E\GDQH(UyGáRZHGR

DQDOL]]DZLHUDá\W\ONRLQIRUPDFMHRGRVW SDFKGRLVWRWQ\FKGRNXPHQWyZQDOH*\SRGGDüSOLNORJX

VHUZHUD:::SURFHVRZLILOWUDFMLZZ\QLNXNWyUHJRLJQRURZDQHV]DSLV\GRW\F]FHSOLNyZQLH

E GF\FKJáyZQ\PLGRNXPHQWDPLRGSRZLDGDMF\PLW]ZVWURQRP:::ang. Web page).

34

Marek Wojciechowski, Maciej Zakrzewicz

2GNU\ZDQLHSUHIHUHQFMLX*\WNRZQLNyZ
3UHIHUHQFMHX*\WNRZQLNyZVUHSUH]HQWRZDQHSU]H]]ELRU\SRGREQ\FKQDMF] FLHMVWRVRZDQ\FK

FLH*HNQDZLJDF\MQ\FK:FHOX]QDOH]LHQLDSUHIHUHQFMLUHDOL]RZDQ\MHVWGZXID]RZ\DOJRU\WP

1.

3U]HV]XNDM ORJ VHUZHUD ::: Z FHOX ]QDOH]LHQLD ZV]\VWNLFK QDMF] FLHM Z\VW SXMF\FK

FLH*HNQDZLJDF\MQ\FK

2.

3RJUXSXM ]QDOH]LRQH FLH*NL QDZLJDF\MQH NLHUXMF VL  LFK ZVSyáVWRVRZDQLHP SU]H]

X*\WNRZQLNyZ W]Q SRGRELHVWZR GZyFK FLH*HN Z\QLND ] WHJR L* ZLHOX X*\WNRZQLNyZ

NWyU]\SRG*DMMHGQ]QLFKSRG*DUyZQLH*GUXJ

2.2. Faza Online algorytmu

2GFKZLOLSLHUZV]HJRSRGáF]HQLDVL X*\WNRZQLNDGRVHUZHUD:::ZV]\VWNLHRSHUDFMHWHJR

X*\WNRZQLNDVUHMHVWURZDQHZIRUPLHW]Zhistorii sesji=DND*G\PUD]HPNLHG\X*\WNRZQLN*GD

SU]HVáDQLD GRNXPHQWX KLVWRULD MHJR VHVML MHVW GRSDVRZ\ZDQD GR LVWQLHMF\FK JUXS FLH*HN

QDZLJDF\MQ\FK L Z\ELHUDQH V WH JUXS\ NWyUH Z\ND]XM VL  QDMZL NV]\P GRSDVRZDQLHP =ELyU

áF]QLNyZ GR GRNXPHQWyZ RSLVDQ\FK Z FLH*NDFK QDZLJDF\MQ\FK GRSDVRZDQ\FK JUXS VWDMH VL

GRGDWNRZ\P HOHPHQWHP ZL]XDOQ\P NWyU\ G\QDPLF]QLH MHVW GRáF]DQ\ GR *GDQHJR GRNXPHQWX
[YJG+96].

3. Podsumowanie

: DUW\NXOH SU]HGVWDZLRQR DUFKLWHNWXU  V\VWHPX DXWRPDW\F]QHM SHUVRQDOL]DFML ]DZDUWRFL

VHUZHUyZ:::NWyU\XPR*OLZLDWZRU]HQLHURGRZLVN:::GRSDVRZXMF\FKVL GR]DFKRZD

X*\WNRZQLNyZ 2EHFQLH Z ,QVW\WXFLH ,QIRUPDW\NL 3ROLWHFKQLNL 3R]QDVNLHM UR]ZLMDQ\ MHVW PRGXá

UR]V]HU]DMF\IXQNFMRQDOQRü Oracle Application 6HUYHUD R WDN UR]XPLDQ DGDSWDW\ZQRü ']L NL

]DVWRVRZDQLX]DSUH]HQWRZDQHMILOR]RILLF] üRGSRZLHG]LDOQRFL]DZ\JOGLVWUXNWXU ]DZDUWRFL

VHUZHUD:::MHVWSU]HQRV]RQD]SURMHNWDQWyZQDX*\WNRZQLNyZ

Literatura

[CP95]

Catledge L.D., Pitkow J.E., ”Characterizing Browsing Strategies in the World Wide Web”, Proc. of

the 3rd Int’l World Wide Web Conference, 1995.

[CM99]

Cooley, R., Mobaser, B., Srivastava, J., „Data preparation for mining World Wide Web browsing

patterns”, Journal of Knowledge and Information Systems, 1, 1999.

[CMS97]

Cooley R., Mobasher B., Srivastava J., “Grouping Web Page References into Transactions for

Mining World Wide Web Browsing Patterns”, Proc. of the 1997 IEEE Knowledge and Data

Engineering Exchange Workshop (KDEX), Newport Beach, California, November 1997.

[H75]

Hartigan J., Clustering Algorithms, John Wiley, 1975.

[HKM97]

Han, E-H, Karypis, G., Kumar, V., Mobasher, B., “Clustering based on association rule

hypergraphs”, Proc. of SIGMOD’97 Workshop on Research Issues in Data Mining and

Knowledge Discovery (DMKD’97), May, 1997.

[L95]

Luotonen A., “The common log file format”, http://www.w3.org/pub/WWW/, 1995.

[PPR96]

Pirolli P., Pitkow J., Rao R., “Silk From a Sow’s Ear: Extracting Usable Structure from the World

Wide Web”, Conference on Human Factors in Computing Systems (CHI 96), Vancouver, British

Columbia, Canada, 1996.

Adaptatywne serwery WWW

35

[P97]

Pitkow J., “In search of reliable usage data on the www”, Sixth Int’l World Wide Web Conference,

Santa Clara, California, 1997.

[YJG+96]

Yan T.W., Jacobsen M., Garcia-Molina H., Dayal U., “From User Access Patterns to Dynamic

Hypertext Linking”, Proc. of the 5th Int’l World Wide Web Conference, 1996.

[PE97]

Perkowitz, M., Etzioni, O., “Adaptive Web Sites: an AI challenge”, Proc. 15

th

Int. Joint Conf. AI,

1997.