Relational Algebra & Client-Server Systems, CS263 Lectures 11 and 12

Relational Algebra v Relational algebra operations work on one or more relations to define another relation leaving the original intact. v Both operands and results are relations, so output from one operation can become input to another operation. v Allows expressions to be nested, just as in arithmetic. This property is called closure. v 5 basic operations in relational algebra: Selection, Projection, Cartesian product, Union, and Set Difference. v These perform most of the data retrieval operations needed. v Also have Join, Intersection, and Division operations, which can be expressed in terms of 5 basic operations.

Relational Algebra Operations

Selection (Restriction)  predicate (R) Works on a single relation R and defines a relation that contains only those tuples of R that satisfy the specified condition (predicate). Example: List all staff with a salary greater than £10,000.  salary > (Staff) 

Projection  col1,..., coln (R) Works on a single relation R and defines a relation that contains a vertical subset of R, extracting the values of specified attributes and eliminating duplicates. Example: Produce a list of salaries for all staff, showing only their staffNo, fName, lName, and salary details.  staffNo, fName, lName, salary (Staff) 

Union R  S Union of two relations R and S defines a relation that contains all the tuples of R, or S, or both R and S, duplicate tuples being eliminated. R and S must be union-compatible (i.e. same attributes).  Example: Produce a list of all staff that work in either of two departments (each department has a separate database), showing only their staffNo, and date of birth.  staffNo, dob (Staff_DepA)  staffNo, dob (Staff_DepB)  staffNodob SL SA DS staffNodob CC SA staffNodob SL SA DS CC  Staff_DepA Staff_DepB

Intersect R  S Defines a relation consisting of the set of all tuples that are in both R and S. R and S must be union-compatible.  Example: Produce a list of staff that work in both department A and department B, showing only their staffNo, and date of birth. (  staffNo, dob (Staff_DepA))  (  staffNo, dob (Staff_DepB)) staffNodob SL SA DS staffNodob CC SA staffNodob SA Staff_DepA Staff_DepB 

Set Difference R – S Defines a relation consisting of the tuples that are in relation R, but not in S. R and S must be union-compatible. staffNodob SL SA DS staffNodob CC SA Example: Produce a list of all staff that only work in department A (each department has a separate database), showing only their staffNo, and date of birth.  staffNo, dob (Staff_DepA)  staffNo, dob (Staff_DepB) staffNodob SL DS Staff_DepA Staff_DepB

Cartesian product R X S Defines a relation that is the concatenation of every tuple of relation R with every tuple of relation S. X Example: Combine details of staff and the departments they work in.  staffNo, job, dept (Staff)  dept, name (Dept) X staffNo job dept SL10 Salesman 10 SA51 Manager 20 DS40 Clerk 20 dept name 10 Stratford 20 Barking X staffNo job dept dept name SL10 Salesman Stratford SA51 Manager Stratford DS40 Clerk Stratford SL10 Salesman Barking SA51 Manager Barking DS40 Clerk Barking Staff Dept

Relational Algebra Operations

Join R S Defines a relation that results from a selection operation (with a join predicate) over the Cartesian product of relation R and relation S. Example: Produce a list of staff and the departments they work in. (  staffNo, job, dept (Staff)) (  dept, name (Dept)) staffNo job dept SL10 Salesman 10 SA51 Manager 20 DS40 Clerk 20 dept name 10 Stratford 20 Barking staffNo job dept dept name SL10 Salesman Stratford SA51 Manager Barking DS40 Clerk Barking Staff Dept Staff.dept = Dept.dept Because the predicate operator is an ‘=‘ this is known as an Equijoin

Natural Join R S This performs an Equijoin of the two relations R and S over all common attributes. One occurrence of each common attribute is eliminated from the result. Example: Produce a list of staff and the departments they work in. (  staffNo, job, dept (Staff)) (  dept, name (Dept)) staffNo job dept SL10 Salesman 10 SA51 Manager 20 DS40 Clerk 20 dept name 10 Stratford 20 Barking staffNo job dept name SL10 Salesman 10 Stratford SA51 Manager 20 Barking DS40 Clerk 20 Barking Staff Dept

Left Outer Join R S Left outer join is a join in which tuples from R that do not have matching values in common columns of S are also included in the resulting relation. (  dept, name (Dept)) (  staffNo, job, dept (Staff)) dept name staffNo job 10 Stratford SL10 Salesman 20 Barking SA51 Manager 20 Barking DS40 Clerk 30 Watford Example: Produce a list of all departments and associated staff that work in them. staffNo job dept SL10 Salesman 10 SA51 Manager 20 DS40 Clerk 20 Staff dept name 10 Stratford 20 Barking 30 Watford Dept

Intersect R  S Defines a relation consisting of the set of all tuples that are in both R and S. R and S must be union-compatible.  Example: Produce a list of staff that work in both department A and department B, showing only their staffNo, and date of birth. (  staffNo, dob (Staff_DepA))  (  staffNo, dob (Staff_DepB)) staffNodob SL SA DS staffNodob CC SA staffNodob SA Staff_DepA Staff_DepB 

Division R S Defines a relation over common attributes C that consists of set of tuples from R that match a combination of every tuple in S.   Example: Show all staff that use all the company’s programming languages. Staff_Prog Prog  language COBOL BASIC staffNolanguage SL10COBOL SA51BASIC SA51COBOL SE14BASIC SE18BASIC staffNo SA51 Staff_Prog Prog 

CS263 Lec. 12: Client/Server systems Operate in a networked environment Processing of an application distributed between front-end clients and back-end servers Generally the client process requires some resource, which the server provides to the client Clients and servers can reside in the same computer, or they can be on different computers that are networked together, usually: Client – Workstation (usually a PC) that requests and uses a service Server – Computer (PC/mini/mainframe) that provides a service. For DBMS, server is a database server

Three components of application logic  1. Input – output or presentation logic component – responsible for formatting and presenting data on the user’s screen (or other output device) and managing user input from keyboard (or other input device)  2. Processing component logic – handles data processing logic (validation and identification of processing errors), business rules logic, and data management logic (identifies the data necessary for processing the transaction or query)  3. Storage component logic – responsible for data storage and retrieval from the physical storage devices – DBMS activities occur here

Client/Server architectures  File Server Architecture  Database Server Architecture  Three-tier Architecture Client does extensive processing Client does little processing

File server architecture  The first client/server architectures developed  All processing is done at the PC that requested the data, I.e. the client handles the presentation logic, the processing logic and much of the storage logic  A file server is a device that manages file operations and is shared by each of the client PCs attached to the LAN  Each file server acts as an additional hard disk for each of the client PCs  Each PC may be called a FAT CLIENT (most processing occurs on the client)  Entire files are transferred from the server to the client for processing.

Three problems with file server architecture  1. Huge amount of data transfer on network - when client wants to access data whole table(s) transferred to PC – so server is doing very little work  2. Each client authorised to use DBMS when DB application program runs on that PC - one database but many concurrently running copies of DBMS (one on each PC) – heavy resource demand on clients  3. DBMS copy in each client must manage shared database integrity - must recognize shared locks, integrity checks, etc - programmers must recognise various conditions that can arise in this environment and understand concurrency, recovery and security controls

File Server Architecture FAT CLIENT

Database server (2-tier) architectures  Client responsible for managing user interface, I/O processing logic, data processing logic and some business rules logic (front-end programs)  Database server performs data storage and access processing (back-end functions) – DBMS only on server  Clients do not have to be as powerful, and server can be tuned to optimise data processing performance  Greatly reduces data traffic on the network, as only records (rather than tables) that match request transmitted to client  Improved data integrity as all processed centrally

Stored procedures Modules of code implementing application logic – included on the database server. Advantages:  Performance improves for compiled SQL statements  Reduced network traffic as processing moves from the client to the server  Improved security if stored procedure is accessed rather than data and code being moved to server  Improved data integrity - multiple applications access same stored procedure  Thinner clients (and a fatter database server) Disadvantages:  Writing stored procedures takes more time than using e.g. VB + proprietary nature reduces portability + performance degrades as number of on-line users increases 

Database server architecture Thinner clients DBMS only on server

3-tier architectures  In general, these include another server layer in addition to the client and database server  This additional server may be used for different purposes  Often application programs reside on the additional server (the application server)  Or additional server may hold a local database whilst another server holds the enterprise database  Often a thin client - PC just for user interface and a little application processing. Limited or no data storage (sometimes no hard drive)

Three-tier architecture Thinnest clients Business rules on separate server DBMS only on DB server

Advantages  Scalability – middle tier can be used to reduce load on database sever by using a transaction processing monitor to reduce number of connections to server, and additional application servers can be added to distribute processing  Technological flexibility – easier to change DBMS engines – middle tier can be moved to different platform. Easier to implement new interfaces  Cost reduction – use of off-the-shelf components/services in the middle tier - also substitution of modules within application rather than whole application  Improved customer service – multiple interfaces on different clients can access the same business process  Competitive advantage – ability to react to business changes quickly by changing small modules of code rather than entire applications

Challenges  High short-term costs – presentation component must be split from process component – this requires more programming  Tools, training and experience– currently lack of development tools and training programmes, and people experienced in the technology  Incompatible standards – few standards yet proposed  Lack of compatible end-user tools – many end-user tools such as spreadsheets and report generators do not yet work through middle-tier services (see later discussion on middleware)

Middleware interoperate  Software which allows an application to interoperate with other software, without requiring the user to understand and code the low-level operations required to achieve interoperability  With Synchronous systems, the requesting system waits for a response to the request in real time  Asynchronous systems send a request but do not wait for a response in real time – the response is accepted whenever it is received.  6 Types of Middleware ->

 1. Asynchronous Remote Procedure Calls (RPC) - client makes calls to procedures running on remote computers but does not wait for a response. If connection lost, must re-establish the connection and send again. High scalability but low recovery  2. Synchronous RPC – distributed program using this calls services available on different computers – possible to achieve this without undertaking detailed coding (e.g. RMI in Java)  3. Publish/Subscribe (push technology) - server monitors activity and sends information to client when available - asynchronous, clients (subscribers) perform other activities between notifications from server.  4. Message-Oriented Middleware – asynchronous, sends messages that are collected and stored until acted upon - client continues with other processing.  5. Object Request Broker (ORB) – tracks location of each object and routes requests  6. SQL-oriented Data Access - translate generic SQL into the SQL specific to the database

Database middleware  ODBC – Open Database Connectivity - most DB vendors support this  OLE-DB - Microsoft enhancement of ODBC  JDBC – Java Database Connectivity - Special Java classes that allow Java applications/applets to connect to databases  CORBA – Common Object Request Broker Architecture – specification of object-oriented middleware  DCOM – Microsoft’s version of CORBA – not as robust as CORBA over multiple platforms

Client/Server security  Network environment has complex security issues. Networks susceptible to breaches of security through eavesdropping, unauthorised connections or unauthorised retrieval of packets of information flowing round the network. Specific security issues include:  System-level password security – user names and passwords for allowing access to the system. Password management utilities  Database-level password security - for determining access privileges to tables; read/update/insert/delete privileges  Secure client/server communication - via encryption – but encryption can negatively affect performance

DB access from clients  Partitioning to create 2, 3 or n-tier architecture - decisions must be made about the placement of the processing logic  Storage logic (the database engine) handled by server, and presentation logic handled by client  Part a) of Fig. depicts possible 2-tier systems, placing processing logic on client (fat client), on server (thin client) or partitioned across both (distributed environment)  Part b) depicts typical 3 and n -tier architectures  Some processing logic placed on the client if desired  Typical client in a Web enabled environment will be a thin client, using a browser for its presentation logic  Middle tiers are typically coded in a portable language such as Java

Processing logic distributions a) 2-tier b) 3 and n-tier Processing logic could be at client, server, or both Processing logic will be at application server or Web server

Open Database Connectivity (ODBC)  An API providing common language for application programs to access/process SQL databases - independent of particular RDBMS  Required parameters: ODBC driver needed, Back- end server name, Database name, User id and password  Fig. Shows generic ODBC architecture  Client application requests connection established with data source  Driver manager identifies appropriate ODBC driver  Driver selected processes requests from the client and submits queries to RDBMS in required version of SQL  Java Database Connectivity (JDBC) similar to ODBC – built specifically for Java applications

ODBC Architecture Each DBMS has its own ODBC-compliant driver Client does not need to know anything about the DBMS Application Program Interface (API) provides common interface to all DBMSs