Scribd
Importer
5 vues22 pages

DDB 1.

Distributed database notes1

Transféré par

mohdmohiuddin1409
Copyright
© All Rights Reserved
Nous prenons très au sérieux les droits relatifs au contenu. Si vous pensez qu’il s’agit de votre contenu, signalez une atteinte au droit d’auteur ici.
Formats disponibles
Téléchargez aux formats PDF ou lisez en ligne sur Scribd
5 vues22 pages

DDB 1.

Distributed database notes1

Transféré par

mohdmohiuddin1409
Copyright
© All Rights Reserved
Nous prenons très au sérieux les droits relatifs au contenu. Si vous pensez qu’il s’agit de votre contenu, signalez une atteinte au droit d’auteur ici.
Formats disponibles
Téléchargez aux formats PDF ou lisez en ligne sur Scribd
INTRODUCTION TO DISTRIBUTED DATABASES Short Que Qi. What is database management system? answer: ADBMS is a collection of programs that enables wssto create and maintain a database, @. Whatis distributed data processing? Aaswer: Distributed processing is the use of more than «re processor to perform the processing for an individual ok @. What are three advantages of distributed ‘processing? Answer: Advantages of Distributed Processing:- 4 Security/Encapsulation 4. Distributed database & Faster Problem solving 4 Security through redundancy SS Coker ive Processing ie eas Q4. What are the disadvantages of distributed data processing? Answer: Disadvantages of distributed data processing (bP) Complexity: Computers attached in DDP are difficult to ‘troubleshoot, design and administrate. Planning data synchronization is difficult: Doing the correct synchronization of data is Aifficult to develop. Sometimes data is updated in wrong Order. So administrators have to keep the focus on it before making a distributed network. Data security: If the unauthorized computer is connected toa distributed network then it can affect other computer performance and data can be a loss also. Q5. What is the goal of distributed database system. Answer: The goal of distributed database system is to achieve data integration and data distribution transparency, Q6. Define physical data independence. Answer: ‘The ability to modify the physical schema without causing application programs to be rewritten. Modifications at this level are usually to improve Q7._ Define logical independence. Answer: The ability to modify the conceptual schema without causing application programs to be [Link] done when logical structure of database is altered. ~ Q8. What is Top down design approach? Answer: It considers the data requirements of entire ‘organization and generates a Global conceptual model of all the information that is required. Q9. What are the two major strategies used for designing distributed databases. Answer; ‘Two major strategies that have been identified for designing a distributed databases are top-down approach and bottom-up approach. Distributed Databases m Quo. ‘What is Conceptual design? Answer: Itisa process by which the enterprise is examined to determine entity types and relationships among these Q11. Define Primary horizontal fragmentation. Answer: the primary horizontal fragmentation of relation is performed using predicates that are defined on that | Q12. Define vertical fragmentation. Answer: Vertical fragmentation partitions a relation into a set of smaller relations so that many of users |" will run on only one Q18. Write promises of DDB’s? i aeeieee ji) Easier system expansion joes to iil) See Q1. Whatis Distributed database system? Explain the goals of it. ‘Answer: A Distributed Database is a database that consists of two or more files located in differént sites either on the semenetows: A centralised distributed database management system integrates data logically soit can be managed as if it were all sored in same location. Goals of Distributed Database System: ‘The concept of distributed database was built with a goal to improve: ®)— Realibility: In Distributed database system, i one system fails doum or stops working for some time another system can complete the task. >) Availability: In Distributed Database system reliability can be achieved even if server fails down. Another system is available ‘oachieve the client request. ©) Performance: Performance can be achieved by distributing database over different locations, So the databases are available to «every location which is easy to maintain. 1 Distrisutep Datasase System QZ. What ae the types of distributed database system? Explain. Answer: ‘Types of Distributed Databases: The two types of Distributed Databases are as follows: @) — Homogeneous distributed database: Homogeneous distributed database system is a network of two or more databases (with same type of DBMS software) which can be stored on one or more machines. Example: c + that we have three departments using Oracle-9i for DBMS. If same changes are made in one ‘department then, it would update the other department also, ‘Waring: XevnPhotmcopying ofthis book ta CRIMINAL Act. Anyone found gully Is ABLE 6 fice LEGAL procedines ———— m14 ig ( Distributed Databases # Po Om con Distributed Databases 158 Explain the features of Distribute: = 2 databases. d | ili) Easier Expansion: Ina distributed environment expansion of the GEE system interms of adding more data, increasing pnewer: database sizes, or adding more processors much easier. Features of Distributed Databases: iv) _ Improved performance: In general, distributed databases include the following features. Location independent Distributed query processing Distributed Transaction management Hardware Independent Operating system independent Network independent ‘Transaction transparency DBMS Independent Advantages of Distributed Databases: Distributed databases basically proposed for various reasons from Organizational decentraliszation and economical processing to greater autonomy. i) Management of data with different level of transprarency: Ideally a database should be distribution transparent in the senxe of hiding the details of where each file is physically stored within the system. a) Network Transparency: ‘This basically refers to the freedom for the user from the operational details of network. These ‘22 of two types location and naming transparency. b) Replication Transparency: It basically made user unaware of existence of copies as we know that copies of data may be stored at multiple sites for better availability ‘performance and reliability. c) Fragmentation transparency: It is basically made user unaware about the ‘existence of fragments it may be vertical fragment orhorizontal fragmentation. ii) Increases Reliability and Availability: Reliability is basically defined a sthe probability ~ that a system is running at certain time whereas Availabilty is defined as probability thatthe system ‘iscontinuously available duringa time interval. ee reerr ts tv We can achieve interquery and intraquery parallelism by executing multiple queries at different sites by breaking up a query intoa number of subqueries that basically executes in parallel which basically leads to improvement in performance. Q4. Explain in detail distributed data processing? Answer: Distributed Data Processing: ‘As information requirements become increasingly complex, organizations must find new ways to distribute processing power, application programs and data. Distributed data processing allows distribution of applciation programs and data among interconnected sites to satisfy the information needs of organization. Depending on its requirements, an organization may choose to centralize or decentralize its data processing systems. Centralized Data Processing: Ina centralized system, one machine controls all file access and updates. ‘To respond to the needs of an organization, @ centralized system permits a high level of contol over application programme and data. Peat aes eae roe a et + Alldataiis shared across application programs. + Many endusers need access to same data and also required the most current data available. + Security has been established as the responsible of central site. “ In a Centralized data processing envitonment, more attention typically is given to direct access storage device than to data transmission. The following figure shows how one machine controls all file access and updates in a centralized system, ‘Warning -Xer/Photocopying of thie book sa CRIMINAL Act. Anyone found guilty is ABLE to face LEGAL proceeding Aes z Ble ee uae s Distributed Databases sf OPERATING SYSTEM eer See) , Host FILES eo MACHINE tnd tha wit cor ee : sy lo Decentralized data Processing: at Inadecentaledsytam, muple machine control fle access and updates fo serve the ered eedsof |g enduser. End users tnd to have more control over application development and operations in a decnevalnd | A environment. A decentralized system is useful when: > Palais primarily accessed at a single location. In this case, itis most efficient to store the data where itis used, > Security has been established as a local responsibility. > Data is used in highly specialized ways by end users. ‘Two kinds of networks: sp stibuted data processing networks make it possible tocombine the benefits of both centralized and “ecmaled sens These nwo canbe viewed nh wa, Z sei 11 cand target nodes: Under IDMS DDS, a single node accepts application requests for database services and either those requests directly by accessing a database Inder its controls or passes the requests to the node tfateontrols appropriate database, Distribute or Centralize data: The data available to applciations executing vuthin the network can be distributed among databases wuld by several different nodes. IDMS DDS is flexible: IDMS DDS permits flexibility in designing a atibuted data base netowrk. Any number of DUCE sgrems can participate in network, each node can be uated at remote site, orseveral nodes can be located atasingle site. Q@. Explain Distributed database system in detail. Answer: Distributed Database System: Distributed Database: It is a collection of multiple interconnected dhlabases, which are spread physically across various laations that communicate via a computer network. ADDBMSis a centralized software system that manages a database in a manner as ifit were all stored ina single location. Features: > Itisused to create, retrieve, update and delete distributed databases. + It synchronizes the database periodically and provides access mechanisms by virtue of which the distribution becomes transparent to users. > It ensures that data modified at any site is ‘universally updated. > Itis designed for heterogeneous’ database Platforms. > Itmaintans confidentiality and data integrity of databse. Factors Encouraging DDBMS: D Rese re read ceases. § Distributed Databases Most organization in the current times ae subdivided into multiple units that are physically distributed over the globe. Each unit requires its own set of local data. Thus, the overall database of the ‘organizaion becomes distributed. Need for sharing of data: The multiple organizational units often need to ‘communicate with each other and share their data and resources. This demands common databases or replicated databases that should be used in a synchronized manner. Database recovery: One of the common methods used in DDBMS is. replication of data across different sites. Replication of data automatically helps in data recovery if database in any site is damaged. Users can access data from other sites while damaged site is being reconstructed. ‘Support for Multiple Application Software: ‘Most organizations use a variety of application software each with its specific database support. DDBMS provides a uniform functionalitty for using the same data among different platforms. 1.1.3 Promises of DDBSs Q7. Discuss the Promises of DDBSs. Answer: AA distributed database management system is. then defined as the software system that permits the ‘management of distributed database and makes the distribution transparent to users. ‘The two important terms in the system are “logically interrelated” and “ Transparent management of distributed and replicated data. —> Reliable access to data throguh distributed transactions. ee Oi eke Distributed Databases # Q8. Explain in detail the problem areas of DDBS. ETO SOTO Answer: Problem Area of DDBS' The following are the problem areas of DDBS's: i) __ Distributed Database design ii) Distributed Query Processing iii) Distributed Directory Management iv) Distributed Concurrency Control ¥)__ Distributed Deadlock Management vi) Reliability of Distributed DBMS vii) Operating system support vii) Heterogeneous Databases ix) Relationship among Problems Distributed Database Design: + Design effects the distribution of database. Non replicated vs. replicated partitions > Twoissues: * Fragmentation * Distribution for op|timum Performance, + NP hard Problem: * Definition: The complexity class of decisions problems that are intrinsically harder than those that an be solved by a Turing machine in polynomial time. * The only tractable aj proach is t ‘pplvheuristics that operate in polynomial time. Distributed Query Processing: > Factors L * Distribution of data * Communication costs * Locality-of data > Take advantage of Parallelhum > NIP-Hard problem tocoptimize’ fetang ante approaches ‘make query processing Bly Distributed of Directory ‘asad > Thereare three levels of directories * Conceptual * Logical * Physical + Directories are consulted for most data operations, _ E > There are many issues that concem whether distributed or centralize the directories. Distributed Concurrency Control: > Syncthonization of access to distribute, databases — Maintain integrity of system. * Single distributed daabase * Multiple copies of of database. > Approaches * Pessimislic concurrency control * Optimistic concurrency control > Example approaches * Locking *Timestamps Distributed Deadlock Management: > Similar to operating system deadloc! ‘management > Well known solutions “Prevention 5 * Avoidance a) ~> Failure recovery among multiple sites > Make sure other systems are reliable &) consistent > Explores the ARIES algorithm. ~ _ Explores Two phase Commit algorithm. Heterogeneous Databases: istributed Databases dlock ERE Architectural Models for Distributed DBMS: Architectural Models for Distributed DBMS can be classified into three dimensions: a) Autonomy >) Distribution 4. Heterogenty Distribution: It states the physical distribution of data across the different users, and ses ne dtrbnson of contlo aaa srensand dope lo which ebh cone DBMS can | Seer Tcias ns uy eG of fee Fasc sere rambo caer Architectural Models: re Some of the common architetural models are . 8) client-server for DDBMS ae ATEN Peer-to-Peer for DDBMD asiak coohaaad © [Link] architecture ‘ee Rearing is ova GRIMINAL AS Ape ol ly MABE a AT pong Ve SE Distributed Databases DDBMS Arthitectural Model: The architecture of a yste defines its structure: Y ‘Components of system are identified. '* Function of eash component is specified. * The interrelationships and interactions among components are defined. E ‘Architecural Models for DDBM’S can be classified along three dimensions. E E B) ! ee multiple ss i , or provid tient r i) Assingle image of the, 3 oe eh ull databces Br ae evalble to any user who wants to share ) Semi autonomous Perspective, the data is logicaly pee The DI ae eabepe- 4 BMS can will make accessible to er'® dependently, Users of oth Each of these deten DBMS's. mines which parts oftheir own database they Total inckanea: wa The indivi ; . to communicate: lual systems are stand ‘ dati! ae 5 with them, slong DBMS's which Imow neithér of existence of other DE mensions of Autonomy: = fother DEMS' ‘nor how 5 Sates S06 an wit m Distributed Databases Design Autonomy: Each individual DBMsis free to use data models and transaction management techniques that it prefers. communication Autonomy: Each individual DBMS is free to decide what information to provide to the other DBMS's. Exeuctive autonomy: Each individual DBMS can execute the transactions that are submitted to it in any way that it wants to. ) Distributions: + _ Distribution refers to distributors of data. Here we are considering the physical distribution of data over Jrutiplestes, the user sees the data as one logical pool. There are two alternatives. i) Client/server distribution ii) Peer-peer distribution ) Client /server distribution: The client/server distribution concentrates data management duties at server while the clients fours on providing the application environment including the user interface. The communication duties are shared between cient machines and servers, li) Peer-to-Peer distribution: There is no distinction of client machines versus servers. Each machine has full DBMS functionality and can communicate with other machines to execute queries and transactions. © Heterogeneity: Itmay occur in various forms in distributed systems, ranging from hardware heterogeneity and differencies innetworking protocols to variations in data managers. + Itnot only involves the use of complexity diferent data access paradigms in different data models, but also covers differences in languages even when individual systems use the same dta model. ‘The dimensions are identified as : A (autonomy), D (distrubution) and H (heterogeneity). the altematives ‘long each dimension are identified by numbers as : 0, 1 or 2. Distributed. Federated DBMS Multi-DBMS DBMS. (Al, DX, Hy) (A2, Dx, Hy) (A0, D2, Ho) . For distributed DBMS, global schema refers to the union of all the local databases. For multi z Albalseghema refers toa subset of the union of al the local databases. HEBEMS. Wann oT IMSINAL Ach, Amvome found guilty io LIABLE ren Die Distributed Databases # a % . a Ti ould sytem asa loa hers | Client fncton Se Loosely - coupled system does not have a Might also include some data manage ae functions not just user interface. Bee Nee Name Client Server architecture is of two types, 1) Pet iG Tight integration ‘Twotier and three tier architecture, For % Semiautonomous systems __ The following figure shows the client gp, 1 MM lox Si architecture for relational systems. Client send gen For to database server and server process on that qa’ —— py, D, No distribution Result of query is passed to client machine » Qil. Dis fient/server system Answer: D, = Client 1 Client 2 Peer-to-Peer systems aa rs ee Application Application % Homnogeneous systems Programs Programs am a differents i Heterogeneous systems Cie |e era. Client licatic L eee Services Services perlicat SSS —— Commits Communic : Dp Q10. Explain in brief on common architecural manager manage: models. Conmataters Answer: wae Common Architecural Models: [malar Some of the common Architecural models of DDBMS are as follows: ie Sane i) Client-server architecture Server3 i) Peertt-peerarchitecture a] ii) Multi DBMS architecture i f Client Server Architecture: f Itis based on database reference model of DBMS, | Theuncfonaliyi ded ino to class sovor apt databan ¢,t? tiet architecture, the server perfom® | : comes fatabase functions and clients perform the userinte® | A clentis defined as requestor of services and a oe j server is defined as provider of services, The term fatclient refers to an arrangement wher? PB wor: Tipe late, the client perform the business functions. In these te Ikeasir to manage the ocr cian eture which makes | architecure, the chene ser i ee " mplexity of modem DEMS's Fae client performs the user interface and me aie cstriby fudions, a database server performs the databas* | inctions and separate computers called application Vt ep servers perform the business functions, i) and local processing 4° Use" interface Architecture : ; ing. Ofclient server Architecture: | ii) Server: ee Hort dvedicdene E Sra eae lorizontal and vertical scaling of resources » Provi and: ri q ee Ee oe to client nee eee > Better price/performance on client machines. Qi. archiving or datat nes i a base access, Ability to use familiar tools on client machines Answ tee "4 ~ Client access to remote data, lata managem¢ i Disadvantages: Optimization transaction ncting query way 2M * aa > Maintenance cost is more, oa " me ene | Xerox/Photoccnny, Letty lery, Sense 4138 & Distributed Databases Itsuffers from security problem as number of users and processing sites increases. Complexity increases. > yy PeerPeer architecture: For answer ESAQ -11 wy Multidatabase Systems: For answer ESAQ-12 Q11. Discuss in brief on Peer-Peer architecture. Answer: Peer-to-Peer Architecture for DDBMS: The physical data organisation on each machine may be different. The different software modules stroed at different siktes to communicate with each other to complete the processing required for execution of distributed applications. This architecture provides both client and server functionalities on each computer . Each node can access senices from other nodes as well as providing services to other nodes. External Eonal Schema 1 Shera? Tocal conceptual Toca conceptual Tocal concephaal schema 1 schema 2 schema N Local Internal Toca internal Tocal Internal schema 1 schema? schema N [pa] The above figure depicts the peer to peer type architecture: Location and replication transparences are supported by definition of local and global conceptual: and mapping in between. chee } Local internal Scheme (LIS) isan individual enternal schema definition at each site, 8) Global Conceptual Schema (GCS) describe the enterprise view of data. Local conceptural schema (LCS) describes the logical organisation of data at each site, © Extemal schemas (ESs) support user applications and user accesso the database. QU. What ié Multidabase system? Explain in detail. Answer: Multidatabase systems: “ Fag AMultidatabase system (MDBS) isa facility that allows users access to data located in multiple autonomous ‘stabase management ystems (DBMSS). eto : Warning XS STF his Took Isa CRIMINAL Act. Anyone found guilty s LIABLE to face LEGAL proceedings Distributed Databases @ may The following figure shows the taxonomy of multidatabase systems:, Mult Database systems Now: Federated MDBS, Federated MDBS Loosely Tightly Coupled Coupled Multiple si Federation A multidatabase system is a software layer on top of existing database systems, which is designed to manipulate ifnormation in, databases. ‘The MDBS software layer translates the queries formulated with the help of global data model and gobs! conceptual scheme into queries based on local data model. To execute global query, the MDBS first translate it into a number of sub queries and converts these sub ‘queries into appropriate local queries for running on local DBMS. After completion of execution, local results are merged and final global result for user query is generated, Aa MDBS controls multiple gateways and manages local database through these gate ways. snterac Sat #2 special program that simulates access from one database to anotehr by coding procotos« interactions by mapping query dialects and data types by maintaining catalog of target and so on. ; Multidatabase can be homogeous or hetergeneous. MDBS can be classified as non-federated MDBS and federated MDBS. 1.2.2 DDBMS Arcurrecrure Q13. Explain DBMS Architecture? 1.15 8 query fr databas connect table en know th b of 1158 —______ it aD isteibuite diDatatiases from a manufacturing client on local dat zi ce se and dept table on remote hq data et ™/g)can retrieve joined data from products table on local Fora client application the location and connected to database mig but want to access fable enables you to issue this query. Platform of databases are transparent. For example if you are data on daabase sq, creating a synonym on mig for remote dept Inthis way, a distributed system gives the a 4 Bho saow tat the daa they access eles ores apPEAFANCE on native database acces, User'son mf dom Manufoctrn Saas Heterogeneous b) Heterogeneous Distributed Database Systems: In this system, atleast one of data bases is a non-Oracle database system. To application this system appears as a single, local, create database. The local oracle server hides the distribution and heterogenity of data. ‘The Oracle database server accesses the non Oracle database system using Oracle heterogeneous services inconjunction with an agni. If you access the non oracle database datastore using an Oracle transparent gateway, then agent is a system specific application. For example if you include a sybase database in an oracle database distributed system then you need to obtain a sybase pecific transparent gatway so that oracle database in system can communicate with it. ©) Client/Server Database architecture: ‘A database server is the oracle s/w managing a database and a client is an application that requests information from a server. Each computer in a network is node that can host one or more database. Each node is distributed or both depending on the situation. In below figure, the host for the hq database is acting as a database saves when a statement is issued against its local data but is acting as a client when it issues statement against remote daa. Adient can connect indirectly to database server. A direct connection opecurs when a client connects to a server and avcesses the information from a dtabase contained an that server. For example if you connect to hq database and access the dept table asin figure you can issue the following: SELECT & FROM dept; : we This query is direct you are not accessing an object an a remote database. SELECT * FROM emp@sales: ‘This query is indirect because the object you are accessing isnot an database to which you are 3 DistRriBuTED DATABASE DESIGN Avrernarive DESIGN STRATEGIES Qi. Discuss in brief on Top down approach. Answer: Top down Design process for DDB: ‘Atopdown design approach Intopdown approach, first the c % ‘seem shut oma sate thetopdoun method nm ee tear defined after then details fo ays Sy rey AEOT Pets nih rg Niven Bo Warning :xo..3 1.16 = 8 & Distributed Databases Requitement Analysis Distribution deisgn Fragmentation = con H. Fragment |_| Freamentat Physical Schemea_ The above figure shows top down design process steps: Design of global schema Design of fragmentation schema Design of Allocations Schema Design of local schema Intopdown design process, distributed database design involves two places: Framentation , Allocation Conceptual design consists of entity analysis and functional analysis. can gly ana determines ents ad relaonshp among them in fancoal analy concept al design ss ote ‘of this book is a CRIMINAL Act. Anyone found guilty is LIABLE i face [BGAV > Distributed Databases @ my Q15. Discuss in brief on bottom up approach. Answer: Bottom-up Approach: ‘The desi stats with specifying requirements and capabilities of individual components and theo ipshation eld do emerge Out of teractions enmong continent components and besaigen eengonerise ‘environment. “This can be used for an existing system. This approach is based on integration of existing: p single, lobal schema. But requires that the following aspects have to be fulfilled. into) 2) Gloablschemad of databa i selected by using common database model b) Local schama is translated into common data model. ‘Integration of common schemata into a common Global Schema. ie “The following figure shows bottom-up approach. at >bal and, toa = SF ies 198 & Distributed Databases q16. Compare top down and Bottom up approach. Answer: ‘Topdown [Bottom-up j)ifsystem is built up from scratch, the top ') __Itsystems should match to existeing down method is more accepted, systems or some modules are yet oneday, the down up method is used. ii) Top-down approach flat defines general. concepts, | ii) First details modules are defined after the global framework after then details, then global framework. ii) Consider data requriements of entire organisation] ii) Consider the existing data distributed with sn organization. iv) Identifies data set end define data elements ™) Identifies data elements Design Issue: fragmentation, allocation 1¥)___Design of export and global schemas. 1.3.2 DistriBuTION DESIGN Q17. Discuss in detail the design issues. [DEUCE PERTIE) ‘Answer: Design Issues: i Distributed Daabase Design: One of the main questions that is being placed or addressed is how database and applications that run 2gainstit should be placed across the sites. These are two basic altematives to placing data partitioned and replicated. In Paratitioned scheme the database is divided into a number of disjoint partitions each of which is placed at differnt site. Replicated designs can be either fullup replciated (also called fully duplicated) where entire database is stored at each site or partially replicated where each partition of database is stored at more than one site, but ot atall the sites, il) Distributed Directory Management: Itcontains is information about data items in the database. Problems related to directory management are ‘similar in nature to database placement problem discussed in proceeding section. A ditectory may be global to entire DDBS or level to each site, it can be centralized at one site or distribued ver several sites, there can be a single copy or multiple copies. Distributed Query Processing: ‘Query processing data with designing algorithms that analyze queries and convert these into a series of data ™ainpulation operations. The problem is how to decide on a strategy for ececuting each query over network in ‘Most cost effective way, however cost is defined. Distributed Concurrency Control: os involves the synchronization of access to the distributed database, bly orca, orntinad Il, without any doubt one of extensively studied probleme s Pt vote q , synchronizing the exeuction of user request before oi eee er crocs cxmutenisicomaeains Mareen ee Waring: XeroyPhotocopyng of this book is a CRIMINAL Act. Anyone found guilty is LIABLE wo face LEGAL proceedings Distributed Databases & ith cd with “Two fundamental primitives that can be use’ both appraoches are locking, which s based on mutes execution starts, and optimistic, executing requests ané then checking if execution has ‘compromised the consistency of data base. “These are variations ofthese schemas as well as hybrid algorithms that attempt to combine the two basic mechanisms. Distributed Deadlock Management: ‘The deadlock in DDBMS is similar in nature to that encountered in OS. “The competition among users for access to set of resources can result in deadlock if synchromization ‘mechanism is based on locking. The will known alternative of preventio, avoidance ‘and detecting recovery also apply to DDBSs. vi) Reliability of Distributed DBMS: One of the original goals of building distributed systems was to make them more reliable than single processor systems. The idea is that ifa machine goes down, some other, machine takes over the job. highly reliable system must be highly available, but hat is not enough. Data e entrusted to the system ‘must not be lost or grabled in any way, and if files are stored redundantly on mutliple servers, all copies must beexcept consistent. Performance: ‘Always the hidden daa in background is issue of performance. Building a transparent, flexible, reliable distribued system, more important lies in its performance, FRAGMENTATION AND ALLOCATION Q18. Discuss in detail on Fragmentation. v) i: be made that the gragments are such that tehy can be iets Warning : Xerox/Photoconnan @ 1.29 reconstruct the original relation (i.e, there isn't any loss of data). Inthe fragmentation process, let's say, ifa table T is fragmented and is divided into a number of fragments say T1, T2, T3. .. TN. The gragments contain sufficient information to allovis the restoration of the original table T. This restoration can be doen by the use of UNION or JOIN operation on various fragments. This process is called data fragmentation. AL! of these fragments are independent. which means these fragments can not be derived from others. The users needn't be logically concemed about fragmentation which means they should not concemed that the data is fragmented and this is called gragmentation independence or we can say fragmentation transaparency. Advantages: * _Asthe data is stored close to the usage site, the efficiency of the database system will increase * Local query optimization methods are sufficient for some queries as the data is available local. * — Inorder to maintain the security and privacy of the database system, fragmentation is advantageous. Disadvantages: * “Access speeds may be very high if data from different fragments are needed * Ifwe are using recursive fragmentation, then it will be very expensive. ‘We have three methods for data fragmenting of atable: * Horizontal fragmentation * Vertical fragmentation Mixed or Hybrid fragmentation Let’s discuss them one by one: Horizontal Fragmentation: as Horizontal fragmentation refers to the process of living a table horizontally by assigning each row or (@ group of rows) of relation to one or more fragments. Be ser are then be assigned to different sides ae distributed system. Some of the rows of tuples laced able 2¢ placed in one system and the rest are plac tal Rae ce stems: The rows that belong tothe ‘are specified by a condition on one prmore attribujtes ofthe relation. Inteational algebo Leia ¢ntation on table T, can be represented amma >" 1218 Distributed Databases o,(1) where o istelational algebra operator for selection, Pisthe condition satisfied by a horizontal fragment Note that a union operation can be performed on the fragments to construct table T. Such a fragment conianingallthe rows of table Tis called a complete horizontal fegment, For example, consider an EMPLOYEE table (T); Eno Ename Design Salary i 101 5 abe 3000 1 102 B abe 4000 1 108 c abe 5500 2 104 D abe! 5000 2 105 E abc 2000 2 This EMPLOYEE table can be divided into different fragments like: EMP 1 = op,., EMPLOYEE EMP 2 = p,.2 EMPLOYEE These two fragments are: TI fragment of Dep =1 Eno Ename Design Salary Dep 101 A abe 3000 1% 102 B abe 4000 1 Similarly, the T2 fragment on the basis of Dep =2 will be: Eno Ename Design Salary Dep 108 c abe 5500 2 104 D abe 5000 2 105, Ey abc 2000 2 Now, here it is possible to get back Tas T = T1 UT2U...UTN Vertical Fragmentation: Vertical fragmentation refers to the process of decomposing a table vertically by attributes are columns. In this fragmentation, some of the attributes are stored in one system and the rest are stored in other eyteme, This fe because each site may not need all columns of atbale. In order to take care of restoration, each fragment must ‘ontain the primary key field (s) in a tbale. The fragmentation should be in such a manner that we can rebuild a table from the fragment by taking the natural JOIN operation and to make it possible we need to include a speci atrbute called Tuple-ito the schema, For this purpose, a user can use any super key. And by this, the topos: ‘ows can be inked together. The projection is as follows: Pajaro (T) . wend? aes Where, x isrelational algebra operaor m se al... anare the aatributes. of T Tis the table (relation) - For example, for the EMPLOYEE table we have Tl as: Distributed Database: Eno Ename 101 A 102 B 103 Cc "104 D E pepe 1 & 105 abe Forthe second, sub table of relation after vertical fragmentation is given as follows: Salary Tole 4000 Renme 5000 2000 e shee eg 5 ‘This is T2 and to get back to the original T, we joint these fragments T1 and T2 as Meyerover (Tq & T2). Atel crt Atle thea arama sme fegments scaled mixed hybrid fagmentation For defring tis type of fragmentation we use the SELECT and ‘the PROJECT operations of relational algebra. In some situations, the horizontal and the vertical fragmentation Beet enya cr mts" Pees acest sated ‘Mixed fragmentation can be done in two different ways: fac Te ist method so fst ceate a eto soup of horizontal agents nd then create vertical fgets _ from one or more of the horizontal au ‘The: Teac retention nen cette

Vous aimerez peut-être aussi