Survey of Database Systems

Introduction
• You have probably used a variety of computer
applications: a word processor, an Internet
browser, games, and so on
• These are relatively small programs, designed to
be installed on a single computer’s hard drive, to
run within that computer, and generally to
interact with only a single user at a time
• That’s the traditional computer application
model, and still it is the largest category of
computer programs
• But there is another category of computer
applications – called enterprise software – and it
is becoming more popular
• Enterprise software can be any type of large-scale
program that is meant to handle the needs of
many users
• One hallmark of enterprise software is that it is
almost always networked – made available to
users via a large private network or the Internet
• Another characteristic of many enterprise
applications is the fact that they are based on
very large, complex databases
• In some organizations, thousands of users may
access the same enterprise application, but
for a wide range of different purposes
 Enterprise Software
• Generally, the word enterprise is used to
describe any organization
• However, people who work with computers
use the term “enterprise” to mean
organizations with large hardware installations
running large software applications
• In other words, an enterprise system is a very
large-scale computer system
• Enterprise software is typically a suite of
applications used by hundreds, or even
thousands, of people at the same time, which
handles millions of records, or both
• Because enterprise systems are so large, it is
natural to think of them as mainframe tools,
and not a major concern for PC users
• However, today’s PCs are as powerful as
mainframes were years ago
• And, many large-scale DBMS applications –
such as Oracle, IBM’s DB2, and Microsoft’s
SQL Server – run just fine on PC servers and
networks
• So, it is quite possible that in a business or
educational environment, you might run into
enterprise software, particularly where a
network is used
• And even if a single PC does not have a huge
application running on it, that PC might be a part
of a group of PCs working together to execute
various parts of a single huge application
• The Internet, and improvements in
communications and security, make it possible for
applications to be divided into parts that reside
on different PCs – even if those PCs are in
separate locations around the world
• An application that executes on separate PCs
like this is called a distributed application
• The word tier is another term used to describe
distributed applications
• For instance, if you divide the job of managing
a database between two computers (a client
and a server), it is called a two-tier
application
• You may also find three-tier applications: one
computer displays a visible user interface for data
entry; a second computer executes the business
logic (figuring out what data is needed from the
database, how to connect to the database, and so
on); and a third computer holds the database
itself
• Some database applications – called n-tier
applications – divide among even more than
three machines
 Meeting the Needs of Many Users
• Some enterprise applications need to handle a
great deal of traffic, from users who may be
located all over the world
• For example, two users with identical software
can access the same data source and perform the
same tasks, even if they are on opposite sides of
a town – or opposite sides of the world
• It is also possible for users to work with different
software but still access the same central data
source
• One example of this is – when people using
different Internet browsers place orders from the
same online shopping database
• One popular type of enterprise software is
electronic document management (EDM) system
– It tracks documents, keeps related ideas together, and
aids in facilities management
– Using such an application, a maintenance worker can
display line drawings of a factory floor to see what
changes are necessary for an impending project
– The same software also allows a secretary to
record text notes from a project meeting and link
them to the line drawings of the factory floor
– Engineers using the software can add red-line
corrections to the factory floor drawings, and
project supervisors can display the meeting notes
– As work on the project progresses, an engineer
may use the software to add a bill of materials
while a company purchasing agent tracks
inventory as it is consumed
• The idea of one application meeting the needs
of many diverse users applies even if the users
are doing different kinds of work
• Consider an EDM system that is used to
manage patient records in a hospital
– A receptionist may use the software to record
appointments and patient insurance claims, while
a doctor may use the same patient records to find
X-rays and medical histories
 – In another part of the hospital, an administrator
can monitor the number of beds in use and
determine staffing needs accordingly
• In other words, enterprise software is large
and each user interface may look different
depending on the needs of the particular user,
but it is all tied together underneath by a
DBMS
• There is another issue to consider when many people
need to use the same database – High traffic can
potentially slow things down
• One popular solution – If someone wants to view a
table of data, it is often best to detach that table from
the database and send this disconnected table to the
user
• For example, many thousands of people surf the
Amazon product catalogue at any given time, so it
would be impractical to maintain a direct open
connection between each of them and the Amazon
database
• Microsoft Access database application can only
efficiently maintain between 10 to 30
simultaneous updating user connections
• Database administrators can solve this bottleneck
to a degree by using disconnected datasets
• Here’s how it works
– The user connects to the database, but only long
enough to “check out” the list of products he/she is
interested in reviewing
– The connection to the database is broken, but the
user can spend as long as he/she wants viewing
the dataset, the portion of the database the user
requested
– Finally, if he/she made any changes to the
database (such as typing in a review of a product),
a new brief connection is established with the
server and the review is submitted
– After being examined for appropriateness by an
employee, the review is then formally added to
the database
– During all these activities, connections to the
database itself are only maintained briefly, only
for as long as it takes to extract a set of data, or
update a table with new data
 Databases in Business
• Databases have become essential to most
modern businesses – to the point that most
companies cannot remain competitive if they do
not computerize at least some of their operations
• Obvious productivity boosts result from using
databases for inventory management, payroll,
order fulfillment, and other clerical tasks
• Perhaps less obvious are the “sometimes”
startling ideas that bubble up when a new
database analysis technique called data mining is
used
• Knowledge discovery is a type of database utility
designed to analyze data and report back with
useful information
• Data mining is one type of knowledge discovery
where a program searches a database,
discovering hidden information
– Using statistics, artificial intelligence (AI), modeling
techniques, and other tools, data mining can predict
trends or relationships that likely would never be
noticed by a human analyst
– Data mining can provide a company with suggestions
about improving efficiency, predict trends, and
otherwise offer wise and useful ideas
– Data mining does not just provide facts; it provides
actual ideas
– Data mining searches a database all by itself,
discovering “gems” of information and bringing them
to attention
– For example, data mining might tell you that 78% of
the shoppers who bought tea kettle from your online
shopping site returned to buy a tea strainer within
two weeks
– With information like this, you can store those
items together in the warehouse, saving your
company some money
– Also, you may modify your online shopping
catalogue to suggest to your tea kettle customers
that they might want to purchase the strainer at
same the time, saving your customers shipping
and handling costs – everybody wins
 Common Corporate Database
Management Systems
• Over the years, database management systems
have come and gone
• In the early 1980s, dBASE was the dominant
commercial database system, but today
enterprise systems dominate the scene – Oracle
Corporation’s Oracle Database, IBM Corporation’s
DB2, and Microsoft Corporation’s SQL Server
– Also, MySQL is introduced; it can be used for free, and
is especially popular for new Internet database work
 Oracle Database System
• First introduced more than 25 years ago, the
Oracle database system has become the most
popular enterprise-level DBMS in the world, with
around 40% of this market
• Oracle – like DBMS competitors Microsoft and
IBM – offers a large suite of applications, utilities,
and languages that can be used together to solve
business problems and manage data
• However, Oracle is losing market share to rivals
IBM and Microsoft
• Oracle has a reputation as the most flexible
enterprise DBMS, but with this flexibility
comes its equally famous complexity
• In addition, Oracle is capable of running on
various operating systems – Windows, UNIX,
Macintosh, and Linux – a capability sometimes
known as platform independence
 DB2
• IBM’s DB2 has versions that run on Windows,
Linux, and various versions of UNIX operating
systems and currently enjoys second place in
popularity with around 34% of installed DBMSs
• IBM is the most venerable of computer
companies and has been developing relational
database systems since 1970
• IBM also developed SQL, the language now used
almost universally for database queries and, in
some cases, even commands that manipulate
data
• As with Oracle and Microsoft, IBM offers a family
of DBMS software, called DB2 Universal
Database, and a suite of associated applications
to assist with data mining, analysis, and
integration, as well as enterprise management
and data warehousing
• DB2 – like Oracle and SQL Server – offers a stable,
robust system that is capable of high-volume
activity, and includes the guarantees and support
that you would expect from a major company like
IBM
 SQL Server
• Microsoft’s SQL Server is the fastest-growing DBMS, though
it currently has only about 11% of the enterprise-level
database management market
• In many ways, it is similar to Oracle, though a key difference
is SQL Server’s requirement to run on only Windows
operating systems and Linux, where Oracle runs on various
operating systems
• Nonetheless, SQL Server is gaining in popularity because it
offers improved language features and greater speed &
efficiency on some benchmark tests
• SQL Server 2019 currently comes in five different versions,
offering various performance and price levels to meet the
needs of various organizations
• The ability to adjust to changes in scale – to
offer stability and efficiency when expanding
from, say, only 100 to 1 million transactions –
is known as scalability
 MySQL
• MySQL is the most popular enterprise DBMS for
the “open-source” community
• Many business enterprises embrace Linux – an
“open-source” operating system and MySQL is
the leading DBMS for the Linux platform
• MySQL is generally simpler to use, but has fewer
features than the competitors, such as DB2 or
SQL Server
• However, an important consideration for many
businesses is MySQL’s lower cost
• MySQL is often recommended for new Web
applications, particularly where cost is a major factor
• MySQL works less well with legacy (older technology)
databases
• Also, MySQL is not as scalable as SQL Server
• So, if you expect your website to have surges of
hundreds of simultaneous visitors interacting with your
database, you need more power than MySQL currently
offers
• In that case, you are better off choosing Oracle or SQL
Server