External Schema for Topic Map Database

23
External Schema of Topic Map Databases Keita Nabeta 1 , Takashi Kojima 2 , Yuki Kuribara 1 , Takashi Yamazaki 1 , Masaomi Kimura 2 1 Graduate School of Engineering, Shibaura InsEtute of Technology 2 Faculty of Engineering, Shibaura InsEtute of Technology

description

In order to cope with large-scale topic maps that store a lot of information, it is necessary to utilize topic map databases. Although, database management systems should provide users with external schema functions such as views, topic map databases do not have such functions. In this paper, we propose a method of implementing a view function, by focusing on the fact that the substructure of topic maps can be regarded as a topic map. In order to realize the idea, we developed an access control system based on the view function. Through an experiment to measure the execution time, we confirmed that these functions work correctly and have little effect on the execution time.

Transcript of External Schema for Topic Map Database

Page 1: External Schema for Topic Map Database

External  Schema  of    Topic  Map  Databases

Keita  Nabeta1,  Takashi  Kojima2,    Yuki  Kuribara1,  Takashi  Yamazaki1,  Masaomi  Kimura2  

1Graduate  School  of  Engineering,  Shibaura  InsEtute  of  Technology  2Faculty  of  Engineering,  Shibaura  InsEtute  of  Technology  

Page 2: External Schema for Topic Map Database

Contents

•  IntroducEon  •  Method  

– Method  to  divide  the  topic  map  –  VIEW  

–  Access  control  system    

•  Experiment  •  Result  &  Discussion  •  Conclusion

30  /  9  /    2010 External  Schema  of  Topic  Map  Database 2

Page 3: External Schema for Topic Map Database

Topic  Map  Database

•  A  topic  map  database  should  provide  an  efficient  method  to  process  data  (e.g.  retrieval,  update).  

30  /  9  /    2010 External  Schema  of  Topic  Map  Database

Update

Retrieval

Topic  map

3

Page 4: External Schema for Topic Map Database

External  Schema

•  In  order  to  limit  user  access  to  a  part  of  some  topic  map,  it  is  desirable  that  the  database  has  external  schema.  –  e.g.)  privacy,  violent  content

30  /  9  /    2010 External  Schema  of  Topic  Map  Database

Accessible

4

Page 5: External Schema for Topic Map Database

External  schema  of  relaEonal  databases

•  RelaEonal  databases  (RDB)  provide  us  with  an  external  schema,  VIEW.  

30  /  9  /    2010 External  Schema  of  Topic  Map  Database

Original  relaEon

VIEW

projecEons  and    selecEons

Users  can  access  the  VIEW  as  if  it  is  a  table,  since  the  VIEW  has  the  same  structure  as  the  original  table.  

5

Page 6: External Schema for Topic Map Database

External  schema  of  topic  maps

•  We  can  regard  the  substructure  of  topic  maps  as  a  topic  map.  

•  Therefore,  we  can  expect  that  it  is  possible  to  realize  the  external  schema  of  topic  maps  by  defining  the  substructure.

30  /  9  /    2010 External  Schema  of  Topic  Map  Database

View

6

Page 7: External Schema for Topic Map Database

ObjecEve  of  our  study

30  /  9  /    2010 External  Schema  of  Topic  Map  Database

We  implement  the  VIEW  to  the  topic  map  database.  

–  In  order  to  define  a  VIEW,  we  propose  the  method  to  specify  the  substructure  of  a  topic  map.  

– We  also  propose  the  way  to  realize  the  funcEon  to  access  the  VIEW.

7

Page 8: External Schema for Topic Map Database

The  method  to  divide  the  topic  map

•  In  order  to  divide  the  topic  map  into  substructures,  we  employed  a  network  clustering  technique  as  an  example  to  define  substructure.  – We  regard  topics  and  associaEons  as  nodes  and  edges.  – We  can  specify  a  group  of  topics  connected  to  each  other.  

30  /  9  /    2010 External  Schema  of  Topic  Map  Database 8

Page 9: External Schema for Topic Map Database

Clustering  syntax

•  We  implemented  clustering  syntax.  

–  The  query  in  this  syntax  returns  all  topics  that  belong  to  the  cluster  including  an  input  topic  as  a  parameter.  

–  The  VIEW  is  realized  by  appending  this  syntax  to  predicates  in  query  as  is  done  to  realize  VIEW  in  RDB.  

30  /  9  /    2010 External  Schema  of  Topic  Map  Database

topicA

cluster(topicA)?

9

Page 10: External Schema for Topic Map Database

•  We  realized  VIEW  by  adding  the  cluster  syntax  to  predicates  in  a  given  query.  

     

Views

30  /  9  /    2010 External  Schema  of  Topic  Map  Database

topic-­‐name($TOPIC,  $NAME)?

topic-­‐name($TOPIC,  $NAME)  AND  cluster(topicA)?

User’s  query

Append  ‘AND’  operaEor  and  cluster  syntax

AND

10

Page 11: External Schema for Topic Map Database

Access  control  system

•  We  implemented  the  funcEon  to  access  the  VIEW  as  an  access  control  system.  

•  For  the  access  control  system,  we  use  following  informaEon.  –  User  list  

•  User  Name  •  Password  •  User  ID  •  Group  ID  

–  Authority  list  •  ID  (User  ID  /  Group  ID)  •  ObjecEve  syntax  •  Predicate

30  /  9  /    2010 External  Schema  of  Topic  Map  Database 11

Page 12: External Schema for Topic Map Database

Flow  of  access  control  mechanism

30  /  9  /    2010 External  Schema  of  Topic  Map  Database

User  Name   Password User  ID Group  ID

User  A aaaa 1 100

User  B bbbb 2 200

ID Objec6ve  syntaxes Predicates

1 topic-­‐name cluster(topicA)?

200 topic-­‐name cluster(topic1178)?

User  list

Authority  list

User  Name:  ‘User  A’  Password:  ‘aaaa’

Query:  topic-­‐name($TOPIC,$NAME)?

User  ID:  1  Group  ID:  100

topic-­‐name($TOPIC,$NAME)          AND  cluster(topicA)?

12

Page 13: External Schema for Topic Map Database

DemonstraEon  of  the  VIEW  and  the  access  control  funcEon

•  In  order  to  demonstrate  the  VIEW  and  the  access  control  funcEon.  –  Query:        topic-­‐  name($TOPIC,  $NAME)?  –  User:      a  user  without  access  limitaEon  (User  A)      a  user  with  access  limitaEon  access  (User  B)  

30  /  9  /    2010 External  Schema  of  Topic  Map  Database 13

Page 14: External Schema for Topic Map Database

The  result  returned  to  the  use  without  access  limitaEon  (User  A)

Input  your  user  name  and  password  User  name:  User  A    Password:  aaaa    You  succeeded  to  access  database    Select  Topic  Maps:  queryTM(Poke.db4o.pokmeonTM)    Query:  topic-­‐name($TOPIC,  $NAME)?  Row:  174  

 $TOPIC  =  bulbasaur      $NAME  =  bulbasaur    $TOPIC  =  ivysaur      $NAME  =  ivysaur      $TOPIC  =  venusaur      $NAME  =  venusaur          .            .        .            .        .            .    $TOPIC  =  monster      $NAME  =  monster    $TOPIC  =  pokemon      $NAME  =  pokemon    $TOPIC  =  instance-­‐of    $NAME  =  instance-­‐of  

30  /  9  /    2010 External  Schema  of  Topic  Map  Database

UserA  can  extract    all  topics  and    their  names.

14

Page 15: External Schema for Topic Map Database

The  result  returned  to  the  use  with  access  limitaEon  (User  B)

Input  your  user  name  and  password  User  name:  User  B    Password:  bbbb    You  succeeded  to  access  database    Select  Topic  Maps:  queryTM(Poke.db4o.pokmeonTM)    Query:  topic-­‐name($TOPIC,  $NAME)?  Row:  10  

 $TOPIC  =  raichu      $NAME  =  raichu    $TOPIC  =  picachu      $NAME  =  picachu    $TOPIC  =  magnemite    $NAME  =  magnemite    $TOPIC  =  magneton    $NAME  =  magneton    $TOPIC  =  voltorb      $NAME  =  voltorb    $TOPIC  =    electrode      $NAME  =  electrode    $TOPIC  =    jolteon      $NAME  =  jolteon    $TOPIC  =  electric      $NAME  =  electric    $TOPIC  =  electabuzz    $NAME  =  electabuzz    $TOPIC  =  zapdos        $NAME  =  zapdos  

30  /  9  /    2010 External  Schema  of  Topic  Map  Database

UserB  can  extract    only  topics  and    their  names  in  the  cluster.

15

Page 16: External Schema for Topic Map Database

Experiment

•  We  evaluated  the  increase  of  execuEon  Eme  caused  by  the  addiEon  of  access  control  procedures  –  using  following  two  topic  maps.  

30  /  9  /    2010 External  Schema  of  Topic  Map  Database

Pokemon  topic  map Large-­‐scale  random  topic  map

Topic 174 2,998

Base  name 174 2,998

AssociaEon 432 9,118

Role 864 18,236

Occurrence 172 0

16

Page 17: External Schema for Topic Map Database

Verifying  affect  of  execuEon  Eme

•  (As  an  example)  we  used  the  typical  query:  –  ‘topic-­‐name($TOPIC,  $NAME)?’.    

•  We  calculated  the  average  execuEon  Eme  of  100  trials  under  the  following  condiEons:    –  query  execuEon  without  access  control  –  execuEon  of  queries  submiked  by  user  without  access  limitaEon  

–  execuEon  of  queries  submiked  by  user  with  access  limitaEon

30  /  9  /    2010 External  Schema  of  Topic  Map  Database 17

Page 18: External Schema for Topic Map Database

Average  execuEon  Eme

3,580.00  

1,717.19  

3,579.76  

1,696.60  

3,293.59  

1,488.61  

0   1,000   2,000   3,000   4,000  

Large-­‐scale  random  topic  map  

Pokemon  topic  map  

Without  access  control  

User  without  access  limitaEon  

User  with  access  limitaEon  

ms

30  /  9  /    2010 External  Schema  of  Topic  Map  Database

The  user  authenEcaEon  does  not  affect  the  execuEon  Eme  for  a  topic  map  that  has  up  to  3,000  topics.

18

Page 19: External Schema for Topic Map Database

Conclusion

•  We  proposed  a  method  to  create  VIEW.  

– We  proposed  the  cluster  syntax  to  specify  a  substructure  of  topic  map.  

–  By  appending  the  ‘AND’  operator  and  the  cluster  syntax  to  the  given  query,  we  realized  the  external  schema  (VIEW)  of  topic  maps.  

– We  also  implemented  the  funcEon  to  access  the  VIEW.

30  /  9  /    2010 External  Schema  of  Topic  Map  Database 19

Page 20: External Schema for Topic Map Database

Conclusion

•  We  confirmed  that  there  is  only  small  increase  on  execuEon  Eme  caused  by  the  addiEon  of  the  access  control  mechanism    –  for  topic  maps  that  have  up  to  3,000  topics.

30  /  9  /    2010 External  Schema  of  Topic  Map  Database 20

Page 21: External Schema for Topic Map Database

PerspecEve

•  We  will  study  the  way  to  realize  inserEon  and  deleEon  operaEons  to  the  VIEW.    

•  It  is  necessary  to  discuss  the  way  to  define  the  substructure  of  topic  maps  other  than  method  based  on  clustering  technique.  

30  /  9  /    2010 External  Schema  of  Topic  Map  Database 21

Page 22: External Schema for Topic Map Database

Thank  you  for  your  akenEon!

30  /  9  /    2010 External  Schema  of  Topic  Map  Database 22

Page 23: External Schema for Topic Map Database

References 1.  Yuki  Kurabara,  Takeshi  Hosoya,  Masaomi  Kimura:  TOME:  Topic  Maps  

Database  Extended.  The  4th  South  East  Asian  Technical  University  ConsorEum  (SEATUC)  Symposium.  pp.245—248  (2010)  

2.  Versant  CorporaEon:  db4objects,  hkp://www.db4o.com/  3.  Joerg  Reichardt,  Stefan  Bornhold  :  StaEsEcal  mechanics  of  community  

detecEon,Physical  ReVIEW  E,  vol.  74,  016110,  pp.1-­‐-­‐14  (2006)    4.  Pokemon  Topic  Map,hkp://www.ontopia.net/omnigator/models/

topicmap_complete.jsp?tm=pokemon.ltm  5.  WANDORA,  hkp://www.wandora.org/  6.  Motomu  Naito:  An  IntroducEon  to  Topic  Maps.  Tokyo  Denki  University  

Press(2006)  7.  Ontopia:  tolog  Language  tutorial,  hkp://www.ontopia.net/  8.  ISO/IEC  JTC1/SC34,  Topic  Map  –  Data  Model,hkp://

www.isotopicmaps.org/sam/sam-­‐model/

30  /  9  /    2010 External  Schema  of  Topic  Map  Database 23