Databases Week 1 – Introduction and Overview
Database is a collection of data with:
a certain logical structure
a specific semantics
a specific group of users
Database Management System (DBMS) allows to
create, modify and manipulate a database
query (retrieve) the data using a query language
support persistent storage of large amounts of data
enable durability and recovery from failure
control access to the data by many users in parallel
without unexpected interactions among users (isolation)
actions on the data should never be partial (atomicity)
Why not store data in files?
no query language
weak logical structure (limited to directories)
no efficient access (searching through files take many hours)
no or limited protection from data loss
no access control for parallel manipulation of data
ANSI SPARC architecture
Motivation for DBMS
Data independence
logical view on the data independent of physical storage
user interacts with a simple view of the data
behind scenes are complex storage structures that allow rapid access
and manipulation
Avoidance of duplication
different views on the same database
for different users or different applications
Hiding parts of data for privacy or security
, Achieved by ANSI SPARC architecture:
Physical level: where data is stored
how the data is stored
“disk pages, index structures, byte layout, record order”
Logical level: Simplified view on physical data
describes data stored in database
and relations among the data
View (external): different views of users and applications on logical level
application programs hide details of data types
hide information (“exam grades”) for privacy or security
This ensures logical and physical data independence
Logical data independence = the ability to modify the logical schema
without breaking existing applications
applications access the views, not the logical database
Physical data independence = the ability to modify the physical schema
without changing the logical schema
“a change in workload might cause the need for:
different indexing structures
different database engine
distributing the database on multiple machines”
Relational Model
We work with relational databases. View and logical level represent data
as relations/tables.
“Relational database instance = table with all info on customer (ID,
name, address)
In the Pure relational model, a table is a set of tuples:
has no duplicate tuples (rows)
no order on the tuples
Database schema = structure of the database, that is, relations +
constraints
2
Database is a collection of data with:
a certain logical structure
a specific semantics
a specific group of users
Database Management System (DBMS) allows to
create, modify and manipulate a database
query (retrieve) the data using a query language
support persistent storage of large amounts of data
enable durability and recovery from failure
control access to the data by many users in parallel
without unexpected interactions among users (isolation)
actions on the data should never be partial (atomicity)
Why not store data in files?
no query language
weak logical structure (limited to directories)
no efficient access (searching through files take many hours)
no or limited protection from data loss
no access control for parallel manipulation of data
ANSI SPARC architecture
Motivation for DBMS
Data independence
logical view on the data independent of physical storage
user interacts with a simple view of the data
behind scenes are complex storage structures that allow rapid access
and manipulation
Avoidance of duplication
different views on the same database
for different users or different applications
Hiding parts of data for privacy or security
, Achieved by ANSI SPARC architecture:
Physical level: where data is stored
how the data is stored
“disk pages, index structures, byte layout, record order”
Logical level: Simplified view on physical data
describes data stored in database
and relations among the data
View (external): different views of users and applications on logical level
application programs hide details of data types
hide information (“exam grades”) for privacy or security
This ensures logical and physical data independence
Logical data independence = the ability to modify the logical schema
without breaking existing applications
applications access the views, not the logical database
Physical data independence = the ability to modify the physical schema
without changing the logical schema
“a change in workload might cause the need for:
different indexing structures
different database engine
distributing the database on multiple machines”
Relational Model
We work with relational databases. View and logical level represent data
as relations/tables.
“Relational database instance = table with all info on customer (ID,
name, address)
In the Pure relational model, a table is a set of tuples:
has no duplicate tuples (rows)
no order on the tuples
Database schema = structure of the database, that is, relations +
constraints
2
