cn 5011 kelemahan sistem relasional

12
1 IF-ITB/TW/1 Agustus 2003 IF5011 Limitations of the Relational Systems Page 1 Basis Data Non Relasional Limitations of the Relational Model Tricya Widagdo Departemen Teknik Informatika Institut Teknologi Bandung

Upload: khaerul-azmi

Post on 23-Jun-2015

564 views

Category:

Documents


6 download

TRANSCRIPT

Page 1: Cn 5011 kelemahan sistem relasional

1

IF-ITB/TW/1 Agustus 2003IF5011 – Limitations of the Relational Systems

Page 1

Basis Data Non RelasionalLimitations of the Relational Model

Tricya WidagdoDepartemen Teknik Informatika

Institut Teknologi Bandung

Page 2: Cn 5011 kelemahan sistem relasional

2

IF-ITB/TW/1 Agustus 2003IF5011 – Limitations of the Relational Systems

Page 2

Areas needing advanced database systems

• both in terms of structure of data and of operations to be performed on data– computer aided design (CAD)

– computer aided manufacturing (CAM)– computer aided software engineering (CASE)– office information systems, multimedia systems and digital

publishing– geographic information systems – etc.

Page 3: Cn 5011 kelemahan sistem relasional

3

IF-ITB/TW/1 Agustus 2003IF5011 – Limitations of the Relational Systems

Page 3

Drawbacks of relational databases

w poor representation of real world entities

w semantic overloading

w poor support for integrity constraints

w homogeneous data structure

w limited operations

w difficulty in handling recursive queries

w impedance mismatch

w other problems ...

Page 4: Cn 5011 kelemahan sistem relasional

4

IF-ITB/TW/1 Agustus 2003IF5011 – Limitations of the Relational Systems

Page 4

Poor representation of real world objects

w normalisation -> fragmentationw relations that have no correspondent in the real worldw to infer information -> joins -> very costly

w poor semanticsw everything is a RELATION (see next slide)

Page 5: Cn 5011 kelemahan sistem relasional

5

IF-ITB/TW/1 Agustus 2003IF5011 – Limitations of the Relational Systems

Page 5

Semantic overloading

w no distinction between entity and relationship; no distinction between different types of relations (e.g. “has” “is the parent of”, “has the address”, “is located”, “earns”)

w result: this semantic cannot be expressed (e.g. be build into the operators)

w there is very little meaning in relations (i.e. the systems “knows” very little of the data it contained); what solutions to increase the semantics exist?

w domains - provided by the relational model, but not fully supported by relational DBMSs

w keys - provided and supported

w support for enterprise integrity constraints …?

Page 6: Cn 5011 kelemahan sistem relasional

6

IF-ITB/TW/1 Agustus 2003IF5011 – Limitations of the Relational Systems

Page 6

Poor support for integrity constraintsw integrity ≡ validity and consistency

w entity integrity + referential integrity + domainsw many systems do not fully support -> build them into applications

-> effort + inconsistencies

w integrity is expressed in terms of constraints (rules that the database must comply with)

w the constraint checking mechanism is simplew degree of compliance; categories of integrity constraints ...

w what does it mean must comply with? w what is the logical model? must be deduced from the DB? the DB must be

consistent with them?must be true of the DB?

w the relational model does not support enterprise integrity constraints

w SQL provides support for constraints as part of the DB def.

Page 7: Cn 5011 kelemahan sistem relasional

7

IF-ITB/TW/1 Agustus 2003IF5011 – Limitations of the Relational Systems

Page 7

Homogeneous data structure

w horizontal and vertical homogeneityw too restrictive

w real world objects have a more complex structure

w result: unnatural structures + many joins (inefficient)

w example: composite parts

w on the other hand, this symmetric structure is one of the strengths of the relational model (why?)

w binary large objects (BLOB)

w are allowed

w typically, they are references to files, therefore some advantages provided by DBMSs may be lost (e.g. security)

w their inner structure cannot be accessed

Page 8: Cn 5011 kelemahan sistem relasional

8

IF-ITB/TW/1 Agustus 2003IF5011 – Limitations of the Relational Systems

Page 8

Limited set of operations

w only a fix set operations

w tuple oriented

wset oriented

w not sufficient for many applications (e.g. geographic information systems - distance, area, …)

w no new operations cannot be specified

wsome DBMSs allow for type extensibility

Page 9: Cn 5011 kelemahan sistem relasional

9

IF-ITB/TW/1 Agustus 2003IF5011 – Limitations of the Relational Systems

Page 9

Difficulty in handling recursive queries

Staff_no Manager_staff_no

S5 S4S4 S3S3 S2S2 S1S1 NULL

Staf f_no Manager_staf f_no

S 5 S 4S 5 S 3… …S 5 S 1S 4 S 3… …S 3 S 2S 3 S 1S 2 S 1S 1 N U L L

Transitive closure

Page 10: Cn 5011 kelemahan sistem relasional

10

IF-ITB/TW/1 Agustus 2003IF5011 – Limitations of the Relational Systems

Page 10

Impedance mismatch

w SQL92 lacks computational completenessw solution: embedded SQLw drawback of solution: impedance mismatch

wmixing different paradigmsw SQL: set based operatorsw high level programming languages: record based (or

even less) (example C)w conversion to records needed (inefficient!)

w type mismatchw solution by Date: build set level facilities in high level

programming languages

Page 11: Cn 5011 kelemahan sistem relasional

11

IF-ITB/TW/1 Agustus 2003IF5011 – Limitations of the Relational Systems

Page 11

Other problems

w other problems also existw no support for long lived transactions (not inherent to the relational

model)

w schema amendments - difficult

w poor navigational access

w therefore, other approaches might be needed

Page 12: Cn 5011 kelemahan sistem relasional

12

IF-ITB/TW/1 Agustus 2003IF5011 – Limitations of the Relational Systems

Page 12

Conclusions

• first generation DBMSs• 1060s-1970s• two approaches: hierarchical and network

• main problems: complex programs needed to be written to answer simple queries (navigational databases); minimal data independence; no widely accepted theoretical foundations

• second generation• 1970 - Codd’s paper

• 1970-1980 - many experimental relational DBMSs• now: over 100 commercial DBMSs (some stretch the definition

of a relational DBMS)

• third generation … (“to be continued”)