Fall 2000 CS 599 – Spatial and Temporal Databases
Research Issues in Spatio-temporal Database Systems
Timos Sellis
Presented byHandy Patriawan
Fall 2000 CS 599 – Spatial and Temporal Databases
Introduction Common techniques and tools
R-Tree for both Spatial and Temporal Join Spatial and Temporal into one
Proposed name: STDBMSs Spatio-temporal DBMSs
Current application Geographic Information Systems (GIS) Environmental Information System Multimedia
Fall 2000 CS 599 – Spatial and Temporal Databases
Introduction (cont’d) Examples:
Fall 2000 CS 599 – Spatial and Temporal Databases
Introduction (cont’d) CHOROCHRONOS
http://www.chorochronos.com Issues to be addressed:
Ontology, structure and representation of space and time
Data model and query language GUI Query processing, storage and indexing
techniques Architecture
Fall 2000 CS 599 – Spatial and Temporal Databases
Introduction (cont’d) Extension of Spatial or Temporal
DBMSs is not enough Adding Temporal support to SDBMSs Adding Spatial support to TDBMSs
Final frontier: unified Spatio-temporal DBMS Spatio-temporal data structures,
operators and user-interfaces
Fall 2000 CS 599 – Spatial and Temporal Databases
Issues… Ontological Issues
Ontology: the study or concern about what kinds of things exist – what entity there are in the universe (http://www.whatis.com)
Including Structure and Representation Finding a common framework
• E.g. common structure and representation
Fall 2000 CS 599 – Spatial and Temporal Databases
Issues… Ontological Issues (cont’d)
• Ontology of spatial entities: life and motion
Fall 2000 CS 599 – Spatial and Temporal Databases
Region
Issues… Ontological Issues (cont’d)
• Spatial objects are located “somewhere” in space
ObjectPart Location
Fall 2000 CS 599 – Spatial and Temporal Databases
Issues… Models and Languages for STDBMSs
Focus on 3 topics:• Language for spatio-temporal relations• Models and languages for STDBMSs• Design and techniques for STDBMSs
Research is divided into 2 categories:• Tight integration of spatial & temporal• Adding spatial support to temporal
Fall 2000 CS 599 – Spatial and Temporal Databases
Issues… Different classes of applications
Events in space & time, Moving region, etc.
See [4] for more details Different data structures &
operations has been defined as well (see [4]) E.g. mpoint x mpoint -> mreal:distance
Fall 2000 CS 599 – Spatial and Temporal Databases
Issues… Constraint database models
Development in spatio-temporal constraint-based database model
Example: DEDALE• Prototype of a constraint DBMS for spatio-temporal
information• See http://www-rocq.inria.fr/verso/dedale/• See http://sikkim.cnam.fr/dedale.html
• Still based on SQL• Display the parts of roads which is inside the rectangle RECT
drawn on the screen• Select SELECT(r.geometry, RECT)
From r in RoadsWhere Satisfy(SELECT(r.geometry, RECT))
Fall 2000 CS 599 – Spatial and Temporal Databases
Issues… Extending current SQL
STSQL Allowing relations to have attributes that are
spatial or temporal Not intended for movement of continuously
moving or changing objects
t=nt=n+1
t=n+2 X
Y
Fall 2000 CS 599 – Spatial and Temporal Databases
Proposed Indexing for Moving Object “Indexing the Positions of Continuously Moving
Objects” Simonas Saltenis, Christian S. Jensen, Scott T.
Leutenegger and Mario A. Lopez Issues: How can we index a moving object (and or
changing object) during time periods? Proposed solution:
• R*-Tree with time as parameter (boring….)• Function…
Fall 2000 CS 599 – Spatial and Temporal Databases
Function for Indexing Moving Objects
t=n X
Y
Time Functionn Y=X
Y=X
t=n+1
n+1
Y=X-2
Y=X-2
t=n+2
n+2 Y=3
Y=3
Fall 2000 CS 599 – Spatial and Temporal Databases
Storage, Indexing and Querying Storage and Indexing:
Support for motion Benchmarking
Indexing: R-Tree, Quadtree, etc. Benchmark:
Make sure to include everything needed Test access method, data generation, query
processing and result analysis ---> Common platform for benchmarking the
STDBMS as a whole
Fall 2000 CS 599 – Spatial and Temporal Databases
Storage, Indexing and Querying Query processing and optimization:
Efficient access for spatial, temporal and spatio-temporal
Simple cost model for query optimization
Study of spatial and temporal constraint database
Fall 2000 CS 599 – Spatial and Temporal Databases
Conclusion Good progress… still more work is
needed: Devising clean and complete semantics
for data models and operators Efficient implementation Indexing and query optimization Alternatives architectures for building
STDBMSs