Transcript An Introduction to DB2

ADABAS Versus DB2
An Evaluative Study by Butler Bloor Group
The DBMS Marketplace
 The OLTP Database

used for production data
 designed for the production logic flow
 The Operational Data Mart

similar to the OLTP DB in Structure & Volume
 used for heavy production-oriented query activity
 The Data Warehouse

Data Volume is much larger than in OLTP DB
 Neutral Data Structures for Data Mining/Trawling
ADABAS Ratings
Performance-large OLTP
10
Performance-large query
10
Performance (General)
Database Functionality
9
7
Data Integration
9
Distributed Features
9
Server Architecture
9
Portability
EASE OF USE ***
UTILITIES *****
DEVELOPMENT
TOOLS *****
8
1 2 3 4 5 6 7 8 9 10
IBM Mainframe Database Ratings
DB2 Ratings
DB2
Performance-large OLTP
8
Performance-large query
9
Performance (General)
Database Functionality
UTILITIES ****
7
6
Data Integration
7
Distributed Features
7
Server Architecture
Portability
EASE OF USE **
DEVELOPMENT
TOOLS **
8
6
1 2 3 4 5 6 7 8 9 10
IBM Mainframe Database Ratings
Performance
 DB2 has a history of performance problems and IBM
has historically addressed it by throwing more
hardware resources at DB2, specifically memory
 ADABAS’ major strength has always been
performance, regardless of platform. We have
numerous success stories and references to support
this claim.
Note the performance ratings on Butler Bloor study.
Resource Consumption
 Resource consumption is a major weakness of DB2.
DB2 is not only mainframe and memory hungry, it is a
DASD eater as well because it stores data
uncompressed
 ADABAS has always proven to be a cost effective
database solution, requiring limited hardware
resources. Unlike DB2, the ADABAS nucleus needs
only 1 region to run in and, because of ADABAS’
unique ability to compress data, it is stingy on DASD
resources
Relational Flexibility
 This is one area where DB2 shines. From the end
user’s perspective, as well as for data modeling, the
relational model greatly simplifies the concept of data
storage and access
 ADABAS also supports the relational model.
Moreover, ADABAS can support a wide variety of data
models which can add flexibility and performance
Development Productivity
 Most of the DB2 related software products currently on
the market focus on performance and tuning, system
monitoring and database administration. CASE tools
have failed miserably to provide productivity gains.
Other offerings are limited to query tools
 Software AG has two different solutions for the
productivity issue with NATURAL for DB2 as well as
NATURAL for ADABAS
Desktop Access
 DB2 offers a superior solution in this category. Due to
DB2’s market presence, and its support of SQL, there
are a wide variety of desktop solutions for accessing
DB2 data
 With the ADABAS SQL Server and ODBC driver
interface we are able to provide direct access to
ADABAS data via ODBC compliant desktop tools
Other Issues
 MUs and PEs: ADABAS has them, DB2 doesn’t
 Binding: Precompiling and binding a program
package for DB2
 Read-thru-locks: An option with DB2, the norm with
ADABAS
Other Issues Cont’d
 Field lengths: Can’t change them in DB2, can in
ADABAS as long as it is not FIXED format
 Database Reorganizations: In DB2 you cannot
reorganize at the table level, only at the DB level.
ADABAS allows you to do both
 Reorders on Simple Tablespaces in DB2 don’t help
anyway because the rows from other tables are
intermingled
Other Issues Cont’d
 Reorganizations (cont’d): When you drop a tablespace
in DB2, that space is not available until the next
Reorganization. If you delete a file in ADABAS, the
space can be reused immediately
 Row overflow: Rows can span a page in DB2, but
causes overhead if it does. Records cannot span
blocks in ADABAS
Other Issues Cont’d
 The 3 “R’s” of DB2: Reorganization, Runstats,
Rebind. ADABAS needs only the first “R” and you’re
ready to go
 Backups: DB2 does not copy indexes with backups.
You must recover them separately. ADABAS
includes both index and data in the backup
Other Issues Cont’d
 Data Compression: DB2 doesn’t do it, unless
hardware compression is involved. Indexes are not
compressed in DB2, nor is the catalog. DASD saving
compression has always been a part of ADABAS
architecture