III. Current Trends

0
0
1791 days ago, 688 views
PowerPoint PPT Presentation
. 12.0 Content. 12.1 Objectives12.2 Overview of Networking12.3 Introduction to DDBMSs-Concepts-Advantages and Disadvantages-Homogeneous and Heterogeneous12.4 Functions and Architecture - Functions of a DDBMS-Reference Architecture for a DDBMS/Federated MDBS12.5 Distributed Relational Database Design-Data Allocation-Fragmentation.

Presentation Transcript

Slide 1

3C13/D6 III. Current Trends Distributed Databases and DBMSs: Concepts and Design

Slide 2

12.0 Content 12.1 Objectives 12.2 Overview of Networking 12.3 Introduction to DDBMSs -Concepts -Advantages and Disadvantages -Homogeneous and Heterogeneous 12.4 Functions and Architecture - Functions of a DDBMS -Reference Architecture for a DDBMS/Federated MDBS 12.5 Distributed Relational Database Design - Data Allocation -Fragmentation 12.6 Transparency in a DDBMS - Distribution Transparency -Transaction Transparency -Performance Transparency 12.7 Date's 12 Rules for DDBMs 12.8 Summary

Slide 3

12.1 Objectives In this Lecture you will learn: Concepts. Points of interest and disservices of circulated databases. Capacities and engineering for a DDBMS. Dispersed database outline. Levels of straightforwardness. Correlation criteria for DDBMSs.

Slide 4

12.2 Overview of Networking Overview of Networking Network: interconnected gathering of self-governing PCs, fit for trading data. Neighborhood (LAN) expected for associating PCs at same site. Wide Area Network (WAN) utilized when PCs or LANs should be associated over long separations. WAN moderately moderate Less dependable than LANs. DDBMS utilizing LAN gives substantially speedier reaction time than one utilizing WAN.

Slide 5

12.2 Overview of Networking Overview of Networking Network: interconnected accumulation of independent PCs, equipped for trading data. Neighborhood (LAN) proposed for associating PCs at same site. Wide Area Network (WAN) utilized when PCs or LANs should be associated over long separations. WAN moderately moderate Less dependable than LANs. DDBMS utilizing LAN gives considerably quicker reaction time than one utilizing WAN.

Slide 6

12.3 Introduction Concepts Databases and systems: An incorporated DBMS could be physically handled by a few PCs appropriated over a system There could be a few separate DBMS on a few PCs circulated over a system There might be a Distributed DBMS (DDBMS) made up of a few DBMSs disseminated over a system each with nearby self-sufficiency Each takes an interest in no less than one worldwide DBMS activity The DDBMS accordingly can work as a solitary worldwide DBMS

Slide 7

12.3 Introduction Concepts DDBMS to Avoid `islands of data' issue… A "Dispersed Database": is a coherently interrelated gathering of shared information (and a depiction of this information), physically conveyed over a PC organize. A "Circulated DBMS" (DDBMS): is a Software framework that allows the administration of the dispersed database and makes the appropriation straightforward to clients. Major Principle: make appropriation straightforward to client. The way that pieces are put away on various PCs is avoided the clients

Slide 8

12.3 Introduction Concepts DDBMS has taking after qualities: Collection of intelligently related shared information. Information split into sections. Sections might be reproduced. Pieces/copies apportioned to locales. Destinations connected by a correspondence organize. Information at each site is under control of a DBMS. DBMSs handle neighborhood applications self-rulingly. Every DBMS takes an interest in no less than one worldwide application.

Slide 9

12.3 Introduction Important distinction amongst DDBMS and dispersed handling ! Conveyed handling of unified DBMS DDBMS

Slide 10

12.3 Introduction Distributed Processing Distributed preparing of an incorporated DBMS has taking after qualities : Much more firmly coupled than a DDBMS. Database configuration is same with respect to standard DBMS No endeavor to reflect authoritative structure Much easier than DDBMS More secure than DDBMS No nearby independence

Slide 11

12.3 Introduction Important distinction amongst DDBMS and parallel database Parallel Database Architectures: Shared: a)memory b)disk c)nothing DDBMS

Slide 12

12.3 Introduction Why utilize a DDBMS? (!) Advantages: Reflects hierarchical structure Improved shareability and neighborhood self-sufficiency Improved accessibility Improved unwavering quality Improved execution Economics Modular development Disadvantages: Complexity Cost Security Integrity control more troublesome Lack of measures Lack of experience Database outline more mind boggling

Slide 13

12.3 Introduction Homogeneous & Heterogeneous DDBMSs Homogeneous: All destinations utilize same DBMS item. Significantly simpler to plan and oversee. Approach gives incremental development Allows expanded execution. Heterogeneous: Sites may run distinctive DBMS items, hidden information models. Locales executed their own databases - mix considered later. Interpretations required to consider Typical arrangement is to utilize entryways . Diverse equipment. Distinctive DBMS items. Diverse equipment and DBMS items.

Slide 14

12.3 Introduction Open Database get to and interoperability "The Open Group" shaped Specification Working Group (SWG) to give particulars that make database framework condition where there is: Common SQL API :permits customer applications to be composed that don't have to know seller of DBMS they are getting to. Regular database convention : empowers DBMS from one seller to discuss straightforwardly with DBMS from another merchant without requirement for a portal. Basic system convention: permits correspondences between various DBMSs.

Slide 15

12.3 Introduction Multidatabase framework (MDBS)! MDBS: DDBMS where each site keeps up entire independence Resides straightforwardly on top of existing database and document frameworks exhibits a solitary database to its clients. Permits clients to get to and share information without requiring physical database mix. 2 sorts: Federated MDBS : resembles a DDBMS for worldwide clients and a concentrated DBMS for nearby clients. Unfederated MDBS : has no "nearby" clients

Slide 16

12.4 Functions and Architecture of a DDBMS Functions and Architecture of a DDBMS

Slide 17

12.4 Functions and Architecture of a DDBMS Functions of a DDBMS Expect DDBMS to have at any rate the usefulness of a DBMS. Likewise to have taking after usefulness: Extended correspondence administrations. Amplified Data Dictionary. Dispersed question preparing. Developed simultaneousness control. Amplified recuperation administrations.

Slide 18

12.4 Functions and Architecture of a DDBMS Reference Architecture A reference engineering comprises of: Set of worldwide outside outlines. Worldwide calculated composition (GCS). Fracture outline and assignment mapping (see later … ) Set of patterns for every neighborhood DBMS complying with 3-level ANSI/SPARC. Correlation with unified MDBS: In DDBMS: GCS is union of all nearby applied patterns. In FMDBS: GCS is subset of neighborhood calculated constructions (LCS), comprising of information that every nearby framework consents to share. GCS of firmly coupled framework includes mix of either parts of LCSs or neighborhood outer outlines. FMDBS without any GCS is called approximately coupled .

Slide 19

12.4 Functions and Architecture of a DDBMS Distributed Relation Database Design

Slide 20

12.5 Distributed Relational Database Design Data Allocation ! Four option procedures in regards to situation of information: Centralized: single database and DBMS put away at one site with users conveyed over the system . Parceled: Database apportioned into disjoint pieces , every fragment alloted to one site . Finish Replication: Consists of keeping up total duplicate of database at each site . Particular Replication: Combination of parceling, replication, and centralization. Examination of techniques

Slide 21

12.5 Distributed Relational Database Design Data Allocation Four option methodologies in regards to situation of information: Centralized: single database and DBMS put away at one site with users circulated over the system . Apportioned: Database parceled into disjoint sections , every fragment alloted to one site . Finish Replication: Consists of keeping up entire duplicate of database at each site . Specific Replication: Combination of parceling, replication, and centralization. Examination of systems

Slide 22

12.5 Distributed Relational Database Design Fragmentation Why piece? Utilization: - Apps work with perspectives as opposed to whole relations. Productivity: - Data put away near where most much of the time utilized. - Data not required by nearby applications is not put away. Security: - thus not accessible to unapproved clients. Parallelism: -With sections as unit of circulation, T can be separated into a few subqueries that work on parts. Drawbacks: Performance & Integrity .

Slide 23

12.5 Distributed Relational Database Design Fragmentation ! Three Correctness of fracture standards: Completeness: If connection R disintegrated into parts R 1 , R 2 , ... R n , every information thing that can be found in R must show up in no less than one piece. Recreation: Must be conceivable to characterize a social operation that will reproduce R from the pieces. - for level discontinuity: Union operation -for vertical: Join 3. Disjointness: If information thing d i shows up in piece R i , then ought not show up in some other section. - Exception: vertical fracture. - For even fracture, information thing is a tuple. - For vertical fracture, information thing is a property.

Slide 24

12.5 Distributed Relational Database Design Fragmentation ! Four sorts of fracture: Horizontal: Consists of a subset of the tuples of a connection. - Defined utilizing Selection operation -Determined by taking a gander at predicates utilized by Ts. - Involves discovering set of negligible ( finish and applicable ) predicates. - Set of predicates is finished , iff, any two tuples in same piece are referenced with same likelihood by any application. - Predicate is significant if there is no less than one application that gets to sections in an unexpected way.

Slide 25

12.5 Di

SPONSORS