ECS 165A: Introduction to Database Systems

1 ECS 165A: Introduction to Database SystemsBertram Lud ascherFall 2013 ECS-165A11. Introduction to Relational DatabasesWhat is a Database system ? In this class we will learn aboutDatabases(DBs) andDatabaseManagement Systems (DBMSs) A Database is a (typically very large) integrated collection ofinterrelated data which are stored in a persistent format (files).Data describe information and activities about one or morerelated organizations (portion of thereal world). For example,a university Database might contain information about entities ( , students, courses, faculties, .. ) relationships among entities ( , Bill is taking ecs 165a ) A Database Management system is a collection of softwarepackages designed to store, access, and manage databases.

2 Itprovides users and applications with an environment that isconvenient, e cient, and IntroductionECS-165A2 File system versus DBMSFile processing approach: prominent approach supported byconventional Operating Systems for data intensive applicationprograms in the 60 s and 70 s (and even sometimes today).In the file processing approach, each application uses its own files:ProductsSuppliersApplicationsAppli cationsCustomersProductsDrawbacks: Data are stored redundantly in many files using di erentformats (=)inconsistencies among data, because changesare not coordinated) No transaction management and concurrency control tosupport multiple users Di cult to access data (no high-levelquery languages) No security mechanisms (OS file permissions are not enough!

3 No recovery mechanisms Program data dependence (=)high maintenance costs) Poor data modeling concepts ..1. IntroductionECS-165A3 Purpose of a Database system Data IntegrationAll data are uniformly managed. E cient Data AccessDatabase languages are provided tostore, access and manage data. Data DictionaryContains all data about objects and structuresof the Database (metadata) User/Application-ViewsDi erent views for di erent users andapplications Integrity Constraintsare enforced by the DBMS. Security Mechanismsto protect data from security threats Transactionscombine sets of operations on data into logicalatomic units Synchronizationof concurrent user transactions Recoveryof data after system crash ad-hocqueries, report generation, interfaces to other databasesystems, interfaces for application programming.

4 1. IntroductionECS-165A4 Trends in Database SystemsSome Database Application Settings EOS NASA s Earth Observation system : Petabyte-sizeddatabases, thousands of customers and applications E-Commerce and Financial Applications: Integration ofheterogeneous information sources ( , catalogs) Health-Care Information Systems and Electronic PatientRecord: Integration of heterogeneous forms of legacy data Digital Libraries & Digital Publishing: Management anddelivery of large bodies of textual and multimedia data Collaborative Work and Design: Integration of heterogeneousinformation sources, concurrency control, workflowsTrends that A ect Database Management Systems Record-based data has been joined by various kinds ofsemistructured and multimedia data Information Superhighway: Web = Constantly growingcollection of heterogeneous information sources High bandwidth, more storage capacity, high performancecomputing.

5 Challenges Data management in sensor networks ( , RFID) Grid computing Integration of multimedia objects and standard data User interfaces development, information visualization Distribution, heterogeneity, complexity of data; security andprivacy; data integration and conversion; data quality, ..1. IntroductionECS-165A5 Why Study Database Systems ? Shift from DBS as a specialized application to centralcomponent of modern computing environments ( Data andInformation Science ) Shift from computation to information at the low end, we have a very messy Web space containinghundreds of thousands of information sources at the high end, we have various needs for scientificapplications Datasets are increasing in diversity, complexity, and volume Digital Libraries, E-, Human Genome Project.

6 The need for (advanced) DBMS is exploding .. !!! Database Systems encompass most of CS topics Operating Systems (file management, process management, .. ) Theory (languages, algorithms, computation, complexity, .. ) Artificial Intelligence (knowledge based Systems , intelligent search, .. ) Software Engineering (application development, GUIs, .. ) Multimedia (video on demand, audio, movies, .. ) Logic (query languages, query optimization, .. )1. IntroductionECS-165A6 Databases are Ubiquitous1. IntroductionECS-165A7Di erent Views of DataA major purpose of a DBMS is to provide users with an abstractview of data, it hides details of how data are stored andmaintained on a (or Conceptual ) Level Physical LevelView 1 View 2 View of abstractionLevels of Abstraction in a DBMS Physical Leveldescribeshowdata records are actually storedand how files and indexes are organized and used Logical Level(sometimes also called Conceptual Level)describeswhatdata are stored in the DBS in terms of entitiesand relationships; emphasis on logical structure of the Database View Leveldescribes how users and applications see the data.

7 Abstraction is achieved by describing each level in terms of adatabase schema, which, in turn, is based on a data model1. IntroductionECS-165A8 Data Models, Schemas, and Instances AData Modelis a collection of concepts for describing data and relationships among data data semantics and data constraints Object-Based logical Models Entity-Relationship (ER) Model Object-Oriented (OO) Model Record-Based logical Models Relational Model Network Model Hierarchical Model Adatabase schemais a description of a particular collectionof data, using a given data a Database schema is the actual content of thedatabase at a particular point in time. Schemas exist at di erent levels of abstraction Conceptual (or Logical) Schema: typically builds the basisfor designing a Database (main focus in this course) View (or External) Schemas: typically determined duringrequirements analysis (often require integration into oneconceptual schema) Physical Schema: storage structures associated withrelations1.

8 IntroductionECS-165A9 Example: University Database Conceptual Schema: can be based on Entity-Relationship,Object-Oriented, or Relational of an Entity-Relationship schema:EnrolledcreditsStudentssidname dobmajoraddressCoursescidcnamegrade Relational schema:Students(sid:integer, name:string, major:string,address:string, dob:date)Courses(cid:integer, cname:string, credits:integer)Enrolled(sid:integer, cid:integer, grade:string) Physical Schema: Relations are stored as unordered files Index on first column of Students External Schemas (Views): CourseInfo(cid:integer, enrollment:integer)1. IntroductionECS-165A10 Data Independence Ability to modify definition of schema at one level withouta ecting a schema definition at a higher level Achieved through the use of three levels of data abstraction(also calledthree level schema architecture) Logical Data Independence: Ability to modify logicalschema without causing application programs to berewritten Physical Data Independence:Abilitytomodifyphysicalsche ma without causing logical schema or applications to berewritten (occasionally necessary to improve performance)1.

9 IntroductionECS-165A11 Database LanguagesA Database Management system o ers two di erent types oflanguages Data Definition Language (DDL) Specification language (notation) for defining a databaseschema; includes syntax and semantics DDL compiler generates set of tables stored in the DBMS sdata dictionary(containsmetadata, data about data ) Data storage and definition language special type of DDLin which storage structures and access methods used bythe DBS are specified Data Definition in SQL (Structured Query Language):create table students(sidnumber,namechar(70),majorcha r(50),dobdate,address char(100)); Data Manipulation Language (DML) Language for accessing and manipulating the data that isorganized according to underlying data model Two classes of languagesProcedural user specifies howrequired data is retrievedDeclarative user specifies whatdata is required withoutspecifying how to get those data(declarative =non-procedural)1.

10 IntroductionECS-165A12 Transaction Management Atransactionis an atomic sequence of Database operationsthat perform a logical function in a Database application. Transaction-management component of DBMS ensures thatthe DB remains in a consistent (correct) state despite systemfailures ( , system crash, power failure) and transactionfailures. Concurrency control manager controls interaction amongconcurrent user transactions to ensure the (read) consistencyof the Management The storage manager is a DBMS software component thatimplements interface between low-level data stored in adatabase (files) and the application programs and queriessubmitted to the system . The storage manager is responsible for the following tasks: interaction with the operating system s file manager e cient storage, retrieval, and update of the data1.