DDB分布式数据库系统 第三讲 Ddb-L3 design

Level of sharing
L3-1 DDBS Design -- 3
Design Approaches
Top-Down
Design systems from scratch Homogeneous systems
Bottom-up
When the database already exist at a number of sites
L3-1 DDBS Design -- 16
HF -Appl. Info. : Sel (m) & Acc (q)
Minterm
Selectivity: sel (mi)
number of tuples of the relation that would be accessed
priority.百度文库
L3-1 DDBS Design -- 2
Dimensions of the Problem
Access pattern behavior
Dynamic Static Complete Information
Data Data + Program Level of knowledge Partial Information
L3-1 DDBS Design -- 4
Top-Down Design Process
User Input
Requirement Analysis
Objectives
User Input
Conceptual Design GCS
View Integration
View Design ES’s
Access Information Distribution Design LCS’s Physical Design
L3-1 DDBS Design -- 9
Allocation Alternatives
Full
Replication.
Replication (Partitioned)
Each fragment reside at each site
Partial
Each fragment reside at some of the sites
– Mixed/Hybrid Fragments -combination of above two.
How
much to fragment
Too little -too much of irrelevant data access. Too much -too much processing cost. Need to find suitable level of fragmentation.
Disjointness
If relation R is decomposed into fragments R1, R2, …, Rn
and data item di is in Rj, then di should not be in any other fragment Rk (k!=j).
L3-1 DDBS Design -- 7
About fragmentation
How
should we fragment
Vertical Fragments -sub grouping of the attributes. Horizontal Fragments -sub grouping of the tuples.
EMP
L1
Eno, Ename, Title
L2
ASG
L3
Eno, Jno, Resp, Dur
L3-1 DDBS Design -- 14
HF -Appl. Info. : Simple Predicates
Predicates used in user queries.

Given R(A1 , A2 , ..., An ), with each Ai having domain of values Di Simple predicate pj defined on R has the form
L3-1 DDBS Design -- 8
Correctness Criteria
Completeness
-no loss of data.
Decomposition of relation R into fragments R1, R2, …, Rn
is complete if and only if each data item in R can also be found in some Ri .
L3-1 DDBS Design -- 12
Horizontal Fragmentation (HF)
Primary
horizontal fragmentation (PHF) based on predicates accessing the relation. Derived horizontal fragmentation (DHF) based on predicates being defined on another logically related relation. We shall first study algorithm for horizontal fragmentation, and then study issues related to derived horizontal fragmentation.
is an operator =, <, >, , , and ; Value Di
Pj : Ai Value
Example:
Jname=“maintenance” Budget 200000
Follow ``80/20'' rule
L3-1 DDBS Design -- 15
Reconstruction
If relation R is decomposed into fragments R1, R2, …, Rn
then there should exist some relational operator such that R = 1 i n Ri
L3: DDBS Design
1. 2.
Introduction Fragmentation
1) Horizontal fragmentation
2) Vertical fragmentation
3.
Allocation
L3-1 DDBS Design -- 1
Distributed Database Design
Allocation Information
requirements
L3-1 DDBS Design -- 6
Why fragment at all?
Unit
of distribution = unit of data application accesses. Reduce irrelevant data access. Facilitates intra-query concurrency. Can be used with other performance enhancing methods, such as, indexing and clustering Applications have conflicting requirements making disjoint fragmentation very hard problem. Multiple fragment access requires join or union. Semantic data control (integrity enforcement) could be very costly.
User Input
LIS’s
L3-1 DDBS Design -- 5
Distribution Design Issues
Fragmentation

Why fragmentation at all How to fragment How much to fragment How to test correctness
M = {mi | mi = Pj Pr p*j }, 1 i z, 1 j z where p*j = pj or ¬pj Example:
m1: Jname=“maintenance” Budget 200000 m2: NOT(Jname=“maintenance”) Budget 200000 m3: Jname=“maintenance” NOT(Budget 200000) m4: NOT(Jname=“maintenance”) NOT(Budget 200000)
L3-1 DDBS Design -- 11
Information Requirements
Four
categories
Database information.
Application information. Communication network information.
Computer system information.
Not-replicated
Each fragment reside at only one site
What
are the advantages and disadvantages?
L3-1 DDBS Design -- 10
Allocation Alternatives: Pros and Cons
HF -Appl. Info. : Minterm Predicates

Given a set Pr = {p1, p2, ..., pm} of simple predicates for relation Ri , the set of minterm predicates M = {m 1, m 2 , ..., m z } defined as
The
design of a distributed computer system involves making decisions on the placement of data and programs across the sites of a computer network. In the case of distributed DBMSs, the distribution of applications involves two things:
Full-Replication Partial -replication Partitioning Query Processing Directory Management Concurrency Control Reliability Reality Easy Easy or non-existent Moderate High Possible application Same Difficulty Same Difficulty Difficult High Realistic Easy Low Possible application
the distribution of DDBMS software distribution of applications
We
concentrate on distribution of data.
The distribution of DDBMS and applications are given a
L3-1 DDBS Design -- 13
HF -Database Information
The global schema
PAY Title, Sal
Owner and member relations
Cardinality of each relation
PROJ Jno, Jname, Budget. Loc