Download to read offline
![DML Compiler import java.sql.*; public class JDBCExample1 { public static void main(String[] args) { try { System.out.println("Loading JDBC driver..."); Class.forName("com.mysql.jdbc.Driver"); System.out.println("JDBC driver loaded!"); } catch (ClassNotFoundException e) { throw new RuntimeException(e); } String url = "jdbc:mysql://localhost:3306/employeeschema"; String username = "root"; String password = "mypassword123"; String query = "select E.Name, D.DName" + "from employee E, department D" + "where E.DNR=D.DNR;"; Connection connection = null; Statement stmt=null; 9 try { System.out.println("Connecting to database"); connection = DriverManager.getConnection(url, username, password); System.out.println("MySQL Database connected!"); stmt = connection.createStatement(); ResultSet rs = stmt.executeQuery(query); while (rs.next()) { System.out.print(rs.getString(1)); System.out.print(" "); System.out.println(rs.getString(2)); } stmt.close(); } catch (SQLException e) { System.out.println(e.toString()); } finally { System.out.println("Closing the connection."); if (connection != null) { try { connection.close(); } catch (SQLException ignore) {}}}}](https://image.slidesharecdn.com/chapter2architectureandclassificationofdbms-250312025210-e74c8508/75/Chapter-2-Architecture-and-Classification-of-DBMS-ppt-9-2048.jpg)
Chapter 2 Architecture and Classification of DBMS.ppt
- 1. Architecture and Classificationof DBMSs www.pdbmbook.com
- 2. Introduction • Architecture ofa DBMS • Categorization of DBMSs 2
- 3. Architecture of aDBMS 3
- 4. Architecture of aDBMS • Connection and Security Manager • DDL Compiler • Query Processor • Storage Manager • DBMS Utilities • DBMS Interfaces 4
- 5. Connection and SecurityManager • Connection manager provides facilities to setup a database connection (locally or through a network) – verifies logon credentials and returns a connection handle – database connection can either run as single process or as thread within a process • Security manager verifies whether a user has the right privileges – read versus write access 5
- 6. DDL Compiler • Compilesthe data definitions specified in DDL • Ideally 3 DDLs (internal/logical/external data model) • DDL compiler first parses the DDL definitions and checks their syntactical correctness • DDL compiler then translates the data definitions to an internal format and generates errors if required • Upon successful compilation, DDL compiler registers the data definitions in the catalog 6
- 7. Query processor • Queryprocessor assists in the execution of database queries such as retrieval, insertion, update or removal of data • Key components: – DML compiler – Query parser – Query rewriter – Query optimizer – Query executor 7
- 8. DML Compiler • DMLcompiler compiles the DML statements • Procedural DML – DML explicitly specifies how to navigate in the database – record-at-a-time DML – no query processor • Declarative DML – DML specifies what data should be retrieved or what changes should be made – set-at-a-time DML – query processor 8
- 9. DML Compiler import java.sql.*; publicclass JDBCExample1 { public static void main(String[] args) { try { System.out.println("Loading JDBC driver..."); Class.forName("com.mysql.jdbc.Driver"); System.out.println("JDBC driver loaded!"); } catch (ClassNotFoundException e) { throw new RuntimeException(e); } String url = "jdbc:mysql://localhost:3306/employeeschema"; String username = "root"; String password = "mypassword123"; String query = "select E.Name, D.DName" + "from employee E, department D" + "where E.DNR=D.DNR;"; Connection connection = null; Statement stmt=null; 9 try { System.out.println("Connecting to database"); connection = DriverManager.getConnection(url, username, password); System.out.println("MySQL Database connected!"); stmt = connection.createStatement(); ResultSet rs = stmt.executeQuery(query); while (rs.next()) { System.out.print(rs.getString(1)); System.out.print(" "); System.out.println(rs.getString(2)); } stmt.close(); } catch (SQLException e) { System.out.println(e.toString()); } finally { System.out.println("Closing the connection."); if (connection != null) { try { connection.close(); } catch (SQLException ignore) {}}}}
- 10. DML Compiler • Impedancemismatch problem –mapping between OO (e.g. Java) and relational (e.g. SQL) concepts • Impedance mismatch solutions –host language and DBMS with comparable data structures (e.g., Java and OODBMS) –middleware to map the data structures from the DBMS to the host language and vice versa 10
- 11. DML Compiler 11 public classEmployee { private int EmployeeID; private String Name; private String Gender; private int DNR; public int getEmployeeID() { return EmployeeID; } public void setEmployeeID( int id ) { this.EmployeeID = id; } public String getName() { return Name; } public void setName( String name ) { this.Name = name; } …} CREATE TABLE Employee ( 'EmployeeID' INT NOT NULL, 'Name' VARCHAR(45) NULL, 'Gender' VARCHAR(45) NULL, 'DNR' INT NULL) Java SQL EmployeeID Name Gender DNR 100 Bart Baesens Male 2 110 Wilfried Lemahieu Male 4 120 Seppe vanden Broucke Male 6 …
- 12. DML Compiler • DMLcompiler starts by extracting the DML statements from the host language. • DML compiler then collaborates with the query parser, query rewriter, query optimizer, and query executor for executing the DML statements • Errors are generated and reported if necessary 12
- 13. Query Parser andQuery Rewriter • Query parser parses the query into an internal representation format • Query parser checks the query for syntactical and semantical correctness • Query rewriter optimizes the query, independently of the current database state 13
- 14. Query Optimizer • Queryoptimizer optimizes the query based upon the current database state (based upon e.g. predefined indexes) • Query optimizer comes up with various query execution plans and evaluates their cost in terms of estimated – number of I/O operations – CPU processing cost – execution time • Estimates based on catalog information combined with statistical inference • Query optimizer is a key competitive asset of a DBMS 14
- 15. Query executor • Resultof the query optimization is a final execution plan • Query executor takes care of the actual execution by calling on the storage manager to retrieve the data requested 15
- 16. Storage manager • Storagemanager governs physical file access and supervises the correct and efficient storage of data • Storage manager consists of – transaction manager – buffer manager – lock manager – recovery manager 16
- 17. Transaction manager • Transactionmanager supervises execution of database transactions – a database transaction is a sequence of read/write operations considered to be an atomic unit • Transaction manager creates a schedule with interleaved read/write operations • Transaction manager guarantees ACID properties • COMMIT a transaction upon successful execution and ROLLBACK a transaction upon unsuccessful execution 17
- 18. Buffer Manager • Buffermanager manages buffer memory of the DBMS • Buffer manager intelligently caches data in the buffer • Example strategies: – Data locality: data recently retrieved is likely to be retrieved again – 20/80 law: 80% of the transactions read or write only 20% of the data • Buffer manager needs to adopt smart replacement strategy in case buffer is full • Buffer manager needs to interact with lock manager 18
- 19. Lock Manager • Lockmanager provides concurrency control which ensures data integrity at all times • Two types of locks: read and write locks • Lock manager is responsible for assigning, releasing, and recording locks in the catalog • Lock manager makes use of a locking protocol which describes the locking rules, and a lock table with the lock information 19
- 20. Recovery Manager • Recoverymanager supervises the correct execution of database transactions • Recovery manager keeps track of all database operations in a log file • Recovery manager will be called upon to undo actions of aborted transactions or during crash recovery 20
- 21. DBMS Utilities • Loadingutility • Reorganization utility • Performance monitoring utilities • User management utilities • Backup and recovery utility 21
- 22. DBMS Interfaces • Web-basedinterface • Stand-alone query language interface • Command line interface • Forms-based interface • Graphical user interface • Natural language interface • Admin interface • Network interface • … 22
- 23. DBMS Interfaces 23 Query window Results window Logwindow Navigator window
- 24. Categorization of DBMSs •Categorization based on data model • Categorization based on degree of simultaneous access • Categorization based on architecture • Categorization based on usage 24
- 25. Categorization based ondata model • Hierarchical DBMSs – adopt a tree like data model – DML is procedural and record oriented – no query processor (logical and internal data model intertwined) – E.g., IMS (IBM) • Network DBMSs – use a network data model – CODASYL DBMSs – DML is procedural and record oriented – no query processor (logical and internal data model intertwined) – CA-IDMS (Computer Associates) 25
- 26. Categorization based ondata model • Relational DBMSs – use the relational data model – currently the most popular in industry – SQL (declarative and set oriented) – query processor – strict separation between the logical and internal data model – E.g., MySQL (open source, Oracle), Oracle DBMS (Oracle), DB2 (IBM), Microsoft SQL (Microsoft) 26
- 27. Categorization based ondata model • Object-Oriented DBMSs (OODBMS) – based upon the OO data model – No impedance mismatch in combination with OO host language – E.g., db4o (open source, Versant), Caché (Intersystems) GemStone/S (GemTalk Systems) – only successfull in niche markets, due to their complexity 27
- 28. Categorization based ondata model • Object-Relational DBMSs (ORDBMSs) – also referred to as extended relational DBMSs (ERDBMSs) – use a relational model extended with OO concepts – DML is SQL (declarative and set oriented) – E.g., Oracle DBMS (Oracle), DB2 (IBM), Microsoft SQL (Microsoft) 28
- 29. Categorization based ondata model • XML DBMSs – use the XML data model to store data – Native XML DBMSs (e.g., BaseX, eXist) map the tree structure of an XML document to a physical storage structure – XML-enabled DBMSs (e.g., Oracle, IBM DB2) are existing DBMSs that are extended with facilities to store XML data 29
- 30. Categorization based ondata model • NoSQL DBMSs – targeted at storing big and unstructured data – can be classified into key-value stores, column- oriented databases and graph databases – focus on scalability and the ability to cope with irregular or highly volatile data structures – E.g., Apache Hadoop, MongoDB, Neo4j 30
- 31. Categorization based upondegree of simultaneous access • Single user versus multi user systems 31 database server instance 1 dispatcher incoming connections server instance 3 server instance 2
- 32. Categorization based onarchitecture • Centralized DBMS architecture – data is maintained on a centralized server • Client server DBMS architecture – active clients request services from passive servers – fat server versus fat client variant • n-tier DBMS architecture – client with GUI functionality, application server with applications, database server with DBMS and database, and web server for web based access 32
- 33. Categorization based onarchitecture • Cloud DBMS architecture – DBMS and database are hosted by a third-party cloud provider – E.g., Apache Cassandra project and Google’s BigTable • Federated DBMS – provides a uniform interface to multiple underlying data sources – hides the underlying storage details to facilitate data access 33
- 34. Categorization based onarchitecture • in-memory DBMS – stores all data in internal memory instead of slower external storage (e.g., disk) – often used for real-time purposes – E.g., HANA (SAP) 34
- 35. Categorization based onusage • On-line transaction processing (OLTP) – focus on managing operational or transactional data – database server must be able to process lots of simple transactions per unit of time – DBMS must have good support for processing a high volume of short, simple queries • On-line analytical processing (OLAP) – focus on using operational data for tactical or strategical decision making – limited number of users formulates complex queries – DBMS should support efficient processing of complex queries which often come in smaller volumes 35
- 36. Categorization based onusage • Big Data & Analytics – NoSQL databases – focus on more flexible, or even schema-less, database structures – store unstructured information such as emails, text documents, Twitter tweets, Facebook posts, etc. • Multimedia – Multimedia DBMSs provide storage of multimedia data such as text, images, audio, video, 3D games, etc. – should also provide content-based query facilities 36
- 37. Categorization Based onUsage • Spatial applications – spatial DBMSs support storage and querying of spatial data (both 2D and 3D) – Geographical Information Systems (GIS) • Sensoring – sensor DBMSs manage sensor data such as biometric data from wearables, or telematics data 37
- 38. Categorization based onusage • Mobile – Mobile DBMSs run on smartphones, tablets or other mobile devices. – should always be online, have a small footprint and be able to deal with limited processing power, storage and battery life • Open source – code of open source DBMSs is publicly available and can be extended by anyone – See www.sourceforge.net – E.g., MySQL (Oracle) 38
- 39. Conclusions • Architecture ofa DBMS • Categorization of DBMSs 39
- 40.