Wednesday, 22 February 2017

LESSON-11 TIPS ON DENORMALIZATION OF TABLES (Cont’d; Ashraf, Khan, and Mir) Posted on September 22, 2012


LESSON-11 TIPS ON DENORMALIZATION OF TABLES (Cont’d; Ashraf, Khan, and Mir)

LESSON-11 TIPS ON DENORMALIZATION OF TABLES (Cont’d; Ashraf, Khan, and Mir)

11.1 Performance Considerations:

a. Recognize the factors that affect performance
b. To be able to spot critical performance bottlenecks
c. To be able to apply some of the techniques that improve performance

When to de-normalize:

a. Long access paths: multiple joins, aggregation of low level data
b. Leading to: excessive I/O usage, excessive response time
c. De-normalized enduser views: setups, maintenance
d. Possible excessive storage requirements

Un-normalized data structures will also cost:

a. Extra Data Maintenance: update, insert, delete
b. Possible inconsistencies
c. Excessive storage requirements
d. Loss standard of instance MIS of retrieval supports

What time to De-normalize:

POINT NO-1:

WHY? Reduce Joins leading to; reduced I/O, reduced CPU

WHAT?

a. Infrequently – updated items
b. Items accessed consistently whenever a different table is referenced

HOW To Identify?

a. From Object (Table) form end of a relationship into Object          (table) at ‘Many’ End

c. Column (attribute) usage statistics, frequency of co-retrieval and frequency of updates.

De-normalization:

a. Making Joins efficient: reduce joins leading to reduce I/O, and CPU
b. What to de-normalize: column accessed consistently whenever another object (table) is referenced
c. Columns duplicated from one object (table) at ONE end of relationship in table at MANY
d. Look at application / column usage information for frequency of co-retrieval/ update.

POINT NO-2:

a. WHY? Reduce Aggregation
b. How? Maintain summary data

What?

a. Frequently aggregated items

Note: If aggregation is periodic consider object (table) populated when required

How To Identify?

a. Summarize object (table) at ‘MANY’ End of a relationship into Object table) at ‘ONE’ End
b. Column (attribute) usage statistics, frequency of co-retrieval and frequency of updates

De-normalization:

a. Reducing aggregation; maintain summary data
b. What to de-normalize; frequently aggregated items

Note: If aggregation is periodic, consider a temporary table, populated when required

c. Summarize column information at MANY End of relationship in
object(table) at ONE End.

*****

Click To MAIN PAGE: OBJECT- ORIENTED RELATIONAL MODELLING E-MONOGRAPH, By J. Ashraf, M. Khan, and H. Mir

*****

 MESSAGE: DEDICATED TO ANONYMOUS COMBATANTS OF KNOWLEDGE

مکتب علم الل مہد منل لحد   Learning continues from birth to death

FOR PROMOTION OF LEARNEDNESS SHARE WITH FRIENDS ABOUT;
   
  1. CAUSE AND EVENTS – SPATIAL And TEMPORAL RESPECTIVELY:                         https://be4gen.wordpress.com/2012/11/16/cause-and-events-spatial-temporal-respectively/

  1. WHETHER ‘BRAIN-POWER’ OR ‘HEART-IMPULSE’ PLINTHS ‘HUMAN GEN’? http://sunedu.wordpress.com/2012/09/17/whether-brain-power-or-heart-impulse-plinths-human-gen/

************

No comments:

Post a Comment