We present Nyaya , a flexible system for the management of Semantic-Web data which couples a general-purpose storage mechanism with efficient ontology reasoning and querying capabilities. Nyaya processes large Semantic-Web datasets, expressed in a variety of formalisms, by transforming them into a collection of Semantic Data Kiosks. Each kiosk exposes the native meta-data in a uniform fashion using Datalog± , a very general rule-based language for the representation of ontological constraints. The kiosks form a Semantic Data Market where the data in each kiosk can be uniformly accessed using conjunctive queries and where users can specify user-defined constraints over the data. Nyaya is easily extensible and robust to updates of both data and meta-data in the kiosk and can readily adapt to different logical organization of the persistent storage. The approach has been experimented using well-known benchmarks, and compared to state-of-the-art research prototypes and commercial systems.

1. Semantic Data Markets
A Flexible Environment for Knowledge Management
R. De Virgilio, G. Orsi, L. Tanca and R. Torlone
CIKM 2011 – Glasgow (UK)

2. Semantic Data Management:
Overview
Problem:
store,
query, and
reason over
semantically annotated data.

3. Semantic Data Management:
Overview
Problem:
store,
query, and
reason over
semantically annotated data.
AT SCALE

4. Semantic Data Management:
Overview
Common limitations:
language-dependent frameworks,
opaque logical and physical organization,
tractable fragments are often ignored.

5. Semantic Data Management:
Overview
Common limitations:
language-dependent frameworks,
opaque logical and physical organization,
tractable fragments are often ignored.
Nyaya: an environment for semantic data management.
[Cali’ et Al. PODS ‘09]
uniform representation of semantic data with Datalog±, [Cali’ et Al. VLDB ‘10]
flexible and transparent storage policy, [Atzeni et Al. VLDBJ ‘08]
[Gottlob et Al. ICDE ‘11]
efficient reasoning and querying. [Orsi et Al. VLDB ‘11]

6. Nyaya:
The kiosk
ΣO ΣO : ontological constraints
ΣS ΣS : storage constraints (mapping)
D D : database

7. Nyaya:
The kiosk
ΣO ΣO : ontological constraints
RDF ΣS ΣS : storage constraints (mapping)
D D : database

8. Nyaya:
The kiosk
schema ΣO ΣO : ontological constraints
RDF ΣS ΣS : storage constraints (mapping)
data D D : database

9. Nyaya:
The kiosk
schema ΣO ΣO : ontological constraints
RDF ΣS ΣS : storage constraints (mapping)
data D D : database

10. Nyaya:
The kiosk
schema ΣO ΣO : ontological constraints
storage ΣS : storage constraints (mapping)
RDF meta ΣS
model
data D D : database

11. Nyaya:
The kiosk
schema ΣO ΣO : ontological constraints
storage ΣS : storage constraints (mapping)
RDF meta ΣS
model
data D D : database

12. Nyaya:
Example
RDF
database constraints

13. Nyaya:
The semantic data market
ΣO
ΣS
D

14. Nyaya:
The semantic data market
ΣO
ΣS
D

15. Nyaya:
The semantic data market
ΣO ΣO ΣO ΣO
ΣS ΣS ΣS … ΣS
D D D D

16. Nyaya:
The semantic data market
user-defined constraints
ΣO ΣO ΣO ΣO
ΣS ΣS ΣS … ΣS
D D D D

17. Nyaya:
The semantic data market
Union of
front-end Conjunctive Queries
application
user-defined constraints
ΣO ΣO ΣO ΣO
ΣS ΣS ΣS … ΣS
D D D D

18. Query Reformulation
Use of FO-rewritability
Q O

19. Query Reformulation
Use of FO-rewritability
I phase
Q O
compilation
(ΣO)
QO

20. Query Reformulation
Use of FO-rewritability
I phase
Q O
compilation
(ΣO)
QO
II phase
compilation
QS
(ΣS)
S

21. Query Reformulation
Use of FO-rewritability
I phase
Q O
compilation
(ΣO)
QO
II phase
compilation
Q* QS
(ΣS)
SQL S
translation

22. Query Reformulation
Use of FO-rewritability
I phase
Q O
compilation
(ΣO)
QO
II phase
compilation
Q* QS
(ΣS)
SQL S
translation
D evaluation

23. Query Reformulation
Example [Gottlob, Orsi and Pieris ICDE ‘11]
professor(X)  Y teaches(X,Y)
ΣO
teaches(X,Y)  student(Y)
Q q(A)  teaches(A,B), student(B)

24. Query Reformulation
Example [Gottlob, Orsi and Pieris ICDE ‘11]
professor(X)  Y teaches(X,f(X))
ΣO
teaches(X,Y)  student(Y)
Q q(A)  teaches(A,B), student(B)

25. Query Reformulation
Example [Gottlob, Orsi and Pieris ICDE ‘11]
t[1]
professor(X)  Y teaches(X,f(X))
ΣO
teaches(X,Y)  student(Y) p[1]
f s[1]
Q q(A)  teaches(A,B), student(B)
t[2]

26. Query Reformulation
Example [Gottlob, Orsi and Pieris ICDE ‘11]
t[1]
professor(X)  Y teaches(X,f(X))
ΣO
teaches(X,Y)  student(Y) p[1]
f s[1]
Q q(A)  teaches(A,B), student(B)
t[2]

27. Query Reformulation
Example [Gottlob, Orsi and Pieris ICDE ‘11]
t[1]
professor(X)  Y teaches(X,f(X))
ΣO
teaches(X,Y)  student(Y) p[1]
f s[1]
Q q(A)  teaches(A,B)
t[2]

28. Query Reformulation
Example [Gottlob, Orsi and Pieris ICDE ‘11]
t[1]
professor(X)  Y teaches(X,f(X))
ΣO
teaches(X,Y)  student(Y) p[1]
f s[1]
Q q(A)  teaches(A,B)
t[2]

29. Query Reformulation
Example [Gottlob, Orsi and Pieris ICDE ‘11]
t[1]
ΣO professor(X)  Y teaches(X,f(X))
p[1]
f s[1]
Q q(A)  teaches(A,B)
t[2]

30. Query Reformulation
Example [Gottlob, Orsi and Pieris ICDE ‘11]
t[1]
ΣO professor(X)  Y teaches(X,f(X))
p[1]
f s[1]
Q q(A)  teaches(X,Y) { XA, Bf(X) }
t[2]

31. Query Reformulation
Example [Gottlob, Orsi and Pieris ICDE ‘11]
t[1]
ΣO professor(X)  Y teaches(X,f(X))
p[1]
f s[1]
q(A)  teaches(A,B)
Q
q(A)  professor(A) t[2]

32. Query Reformulation
Example [Gottlob, Orsi and Pieris ICDE ‘11]
t[1]
ΣO professor(X)  Y teaches(X,f(X))
p[1]
f s[1]
q(A)  teaches(A,B)
QΣ
q(A)  professor(A) t[2]
professor(X)  i-class(Z0,X,Z1), class(Z1,’professor’)
ΣS
teaches(X,Y)  i-objectproperty(Z0,Z1,Z2,Z3), i-class(Z1,X,Z0),
i-class(Z2,Y,Z7), objectproperty(Z3,’teaches’,Z4,Z5)

33. Query Reformulation
Example [Gottlob, Orsi and Pieris ICDE ‘11]
t[1]
ΣO professor(X)  Y teaches(X,f(X))
p[1]
f s[1]
q(A)  teaches(A,B)
QΣ
q(A)  professor(A) t[2]
professor(X)  i-class(Z0,X,Z1), class(Z1,’professor’)
ΣS
teaches(X,Y)  i-objectproperty(Z0,Z1,Z2,Z3), i-class(Z1,X,Z0),
i-class(Z2,Y,Z7), objectproperty(Z3,’teaches’,Z4,Z5)
q(A)  i-objectproperty(Z0,Z1,Z2,Z3), i-class(Z1,A,Z0),
QS i-class(Z2,B,Z7), objectproperty(Z3,’teaches’,Z4,Z5)
q(A)  i-class(Z0,A,Z1), class(Z1,’professor’)

34. Query Reformulation
Example [Gottlob, Orsi and Pieris ICDE ‘11]
t[1]
ΣO professor(X)  Y teaches(X,f(X))
p[1]
f s[1]
q(A)  teaches(A,B)
QΣ
q(A)  professor(A) t[2]
professor(X)  i-class(Z0,X,Z1), class(Z1,’professor’)
ΣS
teaches(X,Y)  i-objectproperty(Z0,Z1,Z2,Z3), i-class(Z1,X,Z0),
i-class(Z2,Y,Z7), objectproperty(Z3,’teaches’,Z4,Z5)
q(A)  i-objectproperty(Z0,Z1,Z2,#(teaches)), i-class(Z1,A,Z0)
QS
q(A)  i-class(Z0,A,#(professor))

35. Experiments
Querying
UOBM Tbox (Approximated)
Instance of 12.8 million triples

36. Experiments
Querying
UOBM Tbox (Approximated)
Instance of 12.8 million triples

37. Experiments
Loading and Updates

38. Experiments
Loading and Updates
If the language of ΣO is FO-rewritable
fact updates reduce to updates in a DBMS
predicate updates reduce to re-compute the rewriting

39. Conclusion
What should we do?
Identifying tractable classes of ontological constraints is crucial
current commercial systems do not do that
Intensional query reformulation delivers very good query performance
Ontology-based data access (ODBA) seamlessly extends traditional
database technology

40. This is the end
Thank you
The Nyaya Family
http://mais.dia.uniroma3.it/Nyaya