Chapter 1.

Introduction

DB2 is Universal Database:

It was the first multimedia, Web-ready relational database
management system, strong enough to meet the demands of large
corporations and flexible enough to serve medium and small-sized business.
DB2 Universal Database powers the most demanding e-business
applications, such as electronic commerce, enterprise resource planning,
customer relationship management (CRM), supply-chain management
(SCM) and Web self-service. It is all about an available and scalable,
industrial-strength database that should be the foundation of your e-business
data management strategy. DB2 includes free support for leading edge
Internet technologies, such as storage and manipulation of extensible
Markup Language (XML)-based documents, high speed in-memory
database techniques for “speed of thought” response times over the Web,
Simple Object Access Protocol (SOAP).

DB2 Universal Database Works With Data:

It means using data assets to make better business decisions. It is
about data access, data analysis and decisions that help control costs uncover
new opportunities and increase customer loyalty. A free built-in foundation
for creating, populating, storing and maintaining Online Analytical
Processing (OLAP) cubes and Data Warehouses/Data Marts helps users to
build small repositories of information and grow their most valuable
resource-their data-into virtually unlimited data stores. This data can then be
leveraged by the enterprise to make faster better decisions that grow the top
line. Data management is more than simply running queries and
applications. Universal Database is a grand name. It implies a product that
is designed to be used for all types of purpose and in all kinds of
environments.

DB2 Architecture:
In this chapter we briefly discuss the DB2 architecture.
The DB2 process model
The DB2 memory model
The DB2 storage model
DB2 process model
Figure depicts the DB2 Process Model. In this figure, rectangles
represent processes while ellipses represent threads. The main DB2 process
is called db2sysc. Under this process there are several threads, the main one
is also called the db2sysc. This is the main thread that spawns other threads.
When a remote application tries to connect to the server using an SQL
CONNECT statement, the remote listeners for the communication protocol
will receive this request and contact a DB2 coordinator agent (db2agent). A
DB2 agent is like a little worker that performs operations on behalf of DB2.
When the application is local, that is, running on the same server as DB2,
the steps are very similar, only that a db2ipccm agent handles the request
instead of the db2tcpcm thread. In some cases, such as when parallelism is
enabled, a db2agent may spawn other agents which appear as db2agntp
threads. Other agents shown in the figure such as db2pfchr, db2loggr,
db2dlock may also be used for different purposes.

DB2 memory model

The DB2 memory model consists of different areas in memory at the
instance level, database level, and application and agent level as shown in
Figure We will not explain in detail each of the different areas in memory in
this book, but just provide a brief overview. When an instance is started, the
database manager shared memory is allocated. This normally does not take
much space. When you first connect to a database, the Database Global
Memory is allocated. In this block, the buffer pool is one of the most
important parts, especially for improving query performance. The size of the
buffer pools will determine the size of the entire Database Global Memory.
Agent private memory is the memory used by each DB2 agent. Without
using the connection concentrator, each connection requires one agent.
Typically an agent uses approximately 3 to 5 MB. With the connection
concentrator, several connections can use one agent, therefore reducing the
need for more physical memory.

DB2 storage model

In this section we will describe the following concepts:
Pages and Extents
Buffer pool
Pages and Extents
A page is the minimum unit of storage in DB2. Allowed page sizes
are: 4K, 8K, 16K and 32K. An extent is a grouping of pages. Working with
one page at a time in DB2 would be costly in terms of performance;
therefore, DB2 works with extents at a time instead. The page size and
extent size are defined when working with buffer pools and table spaces as
we will see in the next sections.

Buffer pools
A buffer pool is a real memory cache for table and index data. It
improves performance by reducing direct sequential I/O and it promotes
asynchronous reading (pre-fetching) and writing. That is to say, DB2
anticipates what pages will be needed and pre-fetches them from the disk to
the buffer pool so they are ready to use. Buffer pools are allocated in
memory units of 4K, 8K, 16K, and 32K pages. There should be at least one
buffer pool per database, and at least one matching buffer pool for a table
space of a given page size.

Control Center:
The control center is DB2 central GUI-based administration tool and
it can be used to administer DB2 databases from pervasive devices such as
Palm Pilots and cellular phones, to clusters of symmetric multiprocessing
(SMP) workstations, to massively parallel processing (MPP) systems like
the IBM z-series servers. The Control center has a setup communications
function that allows you to display the protocols and configuration
parameters that a server’s instance is configured to use. It also allows you
to modify any parameter values of a configured protocol, as well as add or
delete supported protocols.

DB2 and its Companion Products:

Data Servers: Provide software services for the secure and efficient
management of data and enable the sharing of information across multiple
platforms.

Data Warehousing and Business Intelligence Business Intelligence:

Help customers collect, prepare, manage, analyze, and extract valuable
information from all data types to help them make faster, more insightful
business decisions.
Enterprise Content Management & Discovery: Management &
Discovery Manage content, process, and connectivity. The content includes
both structured and unstructured data, such as e-mails, electronic forms,
images, digital media, word processing documents, and Web content.
Perform enterprise search and discovery of information.

Information Integration: Bring together distributed information from

heterogeneous environments. Companies view their information as if it were
all residing in one place.

Characteristics of DB2:
Integrated: DB2 software has built-in support for both Microsoft and Java
development environments. It is also integrated into WebSphere, Tivoli,
Lotus, and Rational products. In addition, the DB2 family has cross-platform
capabilities and can be integrated natively with Web services and message-
queuing technologies. It also provides support for heterogeneous data
sources for both structured and unstructured information, including
pureXML support.
Open: DB2 software allows for different technologies to connect and
integrate by following standards. Thus, it provides strong support for the
Linux operating system and for Java, XML, Web services, grid computing,
and other major industry applications.
Virtualized: Grid computing technology, a type of distributed computing,
collects and shares resources in a large network to simulate one large, virtual
computer. DB2 software products support grid computing technology
through federation and integration technologies. Both of these are discussed
in more detail later in this chapter.
Autonomic: An autonomic computing system manages, repairs, and
protects itself. As systems become more complex, autonomic computing
systems will become essential. DB2 provides self-tuning capabilities,
dynamic adjustment and tuning, simple and silent installation processes, and
integration with Tivoli for system security and management.