Arden University

Advanced database Concepts

Advanced database Concepts

COM6005D

STU58009

Mr. Mohammed Al-Kated

4076 words without Titles and References

Table of content
Task 1 …………………………………………………………………………………………………………..1
Part A 1
Part B 1
Part C 2
Part D 3
Task 2 …………………………………………………………………………………………………………..5
Part A 5
Part B 6
Part C 7
Task 3 …………………………………………………………………………………………………………..9
Part A 9
Part B 9
Part C 10
Part D 12
Task 4 …………………………………………………………………………………………………………..14
Part A 14
Part B 15
 Task 1

A)
Data warehouse (DW) gathers summarised, historical and consolidated data from different sources at
a central platform or repository. This powerful data model is also referred to as a complete
unswerving data store (Hobbs et al., 2005). While on the other hand OLTP technique is about
detailing daily routine transactions of data. These are short online transactions and store current data
in transaction databases such as 3NF. It supports OLAP (online analytical processing) for analysing
multidimensional data while facilitating effective data mining. OLAP has different functional and
performance requirements than online transaction processes (OLTP) (Brar, 2020). Data designed by
DW enables the businesses to accommodate a wide range of ad hoc data analysis and queries. To
collect and summarise the data records, OLAP queries need read-only access instead of locking and
logging mechanisms of OLTP (REDDY et al., 2010).

DW is market-oriented, which provides extensive data storage (100GB-2TB database) and access to
millions of records simultaneously, while OLTP gives access to only tens or hundreds of records.
DW/OLAP uses complex queries for processing historical data from OLTP systems (Cios et al.,
2007). DW and OLAP are becoming the need of the hour due to the provided advantages. DW will
solve the issue of storage and organising large amounts of data available and result in significant
retrieval of data for History Uncovered. Data security problems, accuracy, and availability will also be
resolved as all the data will be available at a central repository (REDDY et al., 2010).

B)
Data warehouse (DW) uses tools to extract data from different internal and external sources. The
below figure shows the architecture of a data warehouse.

Source: (REDDY et al., 2010)

1|Page
These tools clean, transfer, and integrate the collected data and load it in DW. It provides periodical
updates of data and sources while purging the DW data and loading it to archival storage. There are
departmental data marts, while data in both DW and Data marts are managed and stored by DW
servers. Query tools, analysis tools, data mining and report writer tools enable multidimensional
picture and view of data. Lastly, a central repository stores and manages this metadata and monitors
and administers the warehouse system (REDDY et al., 2010). Below are the steps/order used to
design and roll out a DW:

● Firstly, define the warehouse's architecture by performing capacity planning and

selecting tools, database, and servers of storage and OLAP.
● Secondly, all the selected servers, client tools and storage are integrated.
● Thirdly, schema or plan or diagram and views of the warehouse is designed.
● Next, the data access method, data placement and partitioning of the physical
warehouse are defined.
● External and internal sources connect through drivers, gateways and supplementary
wrappers.
● Data cleaning, transferring, loading, and refreshing techniques are designed and
implemented.
● The schema and view scripts, definitions and other relevant data are used to populate
the repository.
● Finally, end-user applications are designed and implemented while rolling out the
warehouse.
Extract, transfer, and load is an integration function of data used to extract data from external sources.
It reshapes data into helpful information and stores it in the central repository of DW. There is a need
to have techniques for data; loading, coordination, hygiene, and transformation along with data
populating technologies (Goar et al., 2010). It is also used to integrate data obtained from various
OLAP databases to OLTP systems. To fulfil these requirements of clients/servers, ETL is used as
without proper data cleansing, extracting, transferring, and loading (integrating), the query process of
the data warehouse would fail. With the help of query cache, it also increases the speed of data
cleaning, transferring, and loading, which improves the response time and performance of the data
warehouse (Goar et al., 2010). Not all ETL are equal in quality and performance. Therefore,
optimisation of ETL is necessary. Only an effective ETL leads to effective decision making while
yielding extraordinary business outcomes (Kakish and Kraft, 2012).

C)
Data marts are the subset of the data warehouse, and they come from using existing data warehouses
of an organisation (REDDY et al., 2010). It consists of focused/specified data sets assisting business
areas/specific subject(s), for example, sales, finance or marketing. They are designed through
complex procedures using advanced technologies and tools (IBM, 2020).

2|Page
These focused data subsets help/aid the originations for different departments/groups and areas to
access data insight without going through the entire warehouse or enterprise data (IBM, 2020). These
data marts will enable History Uncovered to meet its strategic needs by accelerating its business
processes and procedures by providing fast and efficient access to relevant information to each
department. A data mart contains only focused data for a particular line of a business/organisation.
For example, the marketing team can create a data mart without searching the required data in the
entire data warehouse. It will save the efforts and time, and cost while giving quick and efficient
access to critical insights.

D)
Normalisation arranges and organises/reorganises data in the data warehouse. There are two main
reasons; it reduces data redundancy and ensures dependency while maintaining data integrity.
Normalisation occupies small disk space, improves performance, and enhances speed, accuracy, and
database efficiency (Microsoft, 2020). Data is in tables and columns forms, where each table contains
only relevant data. For different types of data, more than one table can be formed or created. It helps
in minimising data modification issues, simplifies queries and minimises the null values.
Normalisation is performed using and adding constraints on data retrieval, up-gradation and deletion
and is of particular importance for OLTP systems. The data warehouse and OLAP systems do not
always deem it necessary to use normalisation (Westerlund, 2008). Data ware uses denormalisation to
improve the performance of queries. The main reason is performance and the query requirement of
access to a complete set of joined data, running complex queries and implementing certain constraints
that are not possible otherwise (Shin and Sanders, 2006).

The multidimensional data model used in a data warehouse consists of star schema, snowflake schema
and fact constellation schema. Star schema, the most common schema, consists of one fact table
placed at the centre and several dimension tables surrounding the centre or fact table. Following is an
example of a star schema where the fact table represents sales and dimension tables represent time,
brand, location, and item. If we look closely, there is only one table for each dimension of the fact
table, while each dimension table contains a set of attributes. For example, the location table contains
city, street, province, and country (Han et al., 2012).

3|Page
On the other hand, the snowflake schema is a variant of the star schema where some of the dimension
tables are normalised and divided into further tables, which form the shape of a snowflake. The
example below illustrates the formation of a snowflake schema (Han et al., 2012).

4|Page
 Task 2

A)
On-line analytical processing provides business with the facility to analyse database information for
business decisions. OLTP is a transition-oriented application system that administers daily routine
transactions of a business. OLTP and OLAP's main difference is its function as OLTP is data
processing while OLAP is about data analysis (Yıldırım, 2020).

Source: (Franco, 2019)

OLAP architecture consists of the following two components:

 OLAP server: it provides a storage facility while performing necessary operations

and formulating multidimensional models.
 OLAP client: it enables the user/client to perform data analysis while
manipulating data.
OLAP uses the ETL technique to simplify data management while reducing absorbed efforts. With
the help of OLAP queries, users can also analyse aggregations, trends, and averages. OLAP provides
a platform to solve/fulfil all analytical needs of a business while operating complex queries. High
cost, departmental cooperation and complicated implementation processes are some of the
disadvantages of OLAP. Despite these issues, OLAP provides data security by imposing restrictions
on user access to data and protects confidential, sensitive and vital data from being exploited (Plattner,
2009). However, OLAP has an insignificant client base compared to OLTP. OLTP preserves data
accuracy and integrity while providing a high speed of data transaction in contrast to OLAP, which
gives consistency of information and calculations. Moreover, the OLAP system helps businesses in
identifying trends and patterns of user interaction. Thus, History Uncover will handle large volumes
of data without compromising on data security and will better understand user behaviour, which will
help identify.

5|Page
 ● Essential periods of history.
● Most/least demanded history genre, e.g. political, military, social, or economic
and most popular digital formats.

B)
A data cube is part of the data warehouse and consists of facts and dimensions. In a data warehouse,
these data cubes are multidimensional structures and are used to store data. The data cube also
represents aggregated data. There are two data cubes; multidimensional and relational data cubes
(Han et al., 2012). Data cubes store precomputed, summarised, aggregated, and optimised data.
Usually, a three-dimensional structure comes into one's mind when it comes to a cube, but in data
warehouse data, an n-dimensional data cube can be implemented (Han et al., 2012). One such
example is Microsoft SQL Server 2000 Analysis Services which gives users 64 dimensions. These
data cube dimensions may comprise angels, attitudes and other entities depending upon the format of
storing enterprise data. It is not easy to visualise these many dimensions of data cubes, so we tend to
look at only three dimensions at a time despite constructing or creating data cubes with many
dimensions. Data cubes are of much importance as one or more of their dimensions can be indexed.
The following figures show an annotated diagram of a data cube of AllElectronics Sales (Han et al.,
2012).

The data of data cubes does not require further calculations as it is already in summarised form. The
figures show unit, regional and store sales along with net sales profit. These data cubes help
organisations to establish data trends over time. History Uncovered can use the data of data cubes to
not only explore data trends along with performance analysis within the company but also can analyse
third-party data for this purpose.

Through OLAP operations of background knowledge of data domain, data can be represented at
various abstraction levels, which accommodates different viewpoints of different users. These
operations or functions are as follows. Drill-down; Data is divided into small parts by moving the
concept hierarchy down and increasing dimension. Roll up; is the opposite of to drill down function as

6|Page
concept hierarchy is moved up and dimensions are reduced. The following figure shows the
implementation roll-up function.

 Slice and Dice: one or more dimensions of the data cube is picked and forms a sub-data cube.

 Pivot: Data cube is rotated to view another view of data.

Following aggregation functions are applied on the data cube along with the above operations (Activ,
2020):

 Sum: returns the total amount of selected data.

 Average: gives the average value of selected data

 Minimum: provides lowest data value

 Maximum: returns a maximum data value of selected data

 Count: number of fields/rows included in your data set.

 Unique count: Returns distinct/ unique values of the data.

C)
Data mining looks at hidden and useful data patterns. It operates at a detailed level while providing
future prediction and insight, whereas OLAP summarises data and provides data forecasts (Abdellatif
et al., 2011). Data mining also provides information about previously unknown/unsuspected
relationships of data. However, both OLAP and Data mining can complement each other and give
more insight into data. Prediction of unusual patterns of user interaction will help this client in better
understanding customer behaviour while providing an insight of the regions/clients producing higher
or lowest revenue or profit for the organisation and thus will help in identifying the client profile

7|Page
along with location and most demanded subjects/content/histrionic periods. It also provides details if
any suspicious activity is performed and protects confidential, sensitive data from hackers.

There are three techniques commonly used for data mining.

 Classical Statistics: It used to study data, relationships, and large numerical data
of a database. It comprises regression analysis, discriminant analysis and cluster
analysis.
 Artificial Intelligence (AI): It is used for processing statistical problems of data.
AI uses the following techniques: fuzzy logic, neural computing and genetic
algorithms.
 Machine Learning: it is used to analyse large data and knowledge discovery
from the data (Siguenza-Guzman et al., 2015). It combines advanced statistical
methods and AI heuristics which includes swarm optimisation, symbolic learning,
and neural networks methods.
Alongside these techniques’ regression, near-neighbour methods, multivariate analysis, survival
analysis, tree-based models and generalised linear models are used in data mining (Hosking et al.,
1997). Various hypotheses are developed to validate these techniques and models. Following are
specific steps need to implement data mining:

● Understand business and client objectives.

● Link/inline business goals with data mining goals.
● Collect data and data properties.
● Ascertain data quality and acquire missing data (if any).
● Next, data is made ready by cleaning, transforming, formatting, and constructing.
● Then Mathematical models are used to determined and used for determining data
patterns.
● Identified patterns are analysed and evaluated against defined business objectives.
● Finally, data mining is deployed by shipping discoveries to day-to-day operations.

8|Page
 Task 3

A)
An object can be anything in an enterprise, such as a desk or a file. In computer science, an object is a
variable, a method, a function and data structure. It is the value of memory of a computer system
referred to by an identifier (Tupper, 2011). It has two characteristics:

 State (data)
 Behaviour (code)
For example, the service delivery department of History Uncovered has objects such as documents,
video clips, photographs, and audio clips. The department also supports behaviour such as requests for
original materials, subscription packages and schedule one-to-one online meetings such as Skype.
Each object has a unique identifier that remains unchanged throughout an object's life cycle
(Braumandl et al., 1998). In a relational database model, an object can be an association of data, a
column or a table. For example, relating compliance of laws and regulations of a particular area to
business units operating there. The most significant advantage of using objects in a relational database
is that it enables users to define data types, functions, and operators. This, in return, increases the
functionality and performance of a database (García et al., 2015). It also provides benefits of
encapsulation, inheritance, extensible and flexible indexing while handling different types of data
such as audio, video and image files compared to a relational database (Weikum et al., 1999; Paterson
et al., 2006).

O/RDBMS bridged the gap between relational and object-oriented database as it combines both
(García et al., 2015). It represents data and relationships are in the forms of inter-related tables, while
in OODBMS, data is represented in object form. O/RDBMS stores complex and comparatively
smaller data, whereas OODBMS stores complex and extensive data and methods to use it. In
O/RDBMS primary key is used to identify objects in a table, whereas OODBMS uses object
identifiers to define an object/entity (Sabau, 2007).

B)
UML class diagram is used to visualize object-oriented systems. It is a blueprint that is used to create
objects and/or sets of objects. The following diagram represents elements of the UML class diagram.

9|Page
The class diagram of UML is used to describe system structure through the use of

 Classes
 Attributes or State
 Operations, Behaviour or Methods
 Relationship of objects.

Following is a UML class diagram that defined the objects of History Uncovered.

10 | P a g e
 C)
Multimedia means computer information in digital forms of text, audio, video, and still/moving
objects(animations). The text, audio, video, and still/moving objects are known as multimedia data
(Steinmetz, 2012). One can define it as anything or any type of data except conventional
alphanumeric data. The information conveyed and represented through multimedia can represent
anything from the real world compared to traditional data (Vries et al., 1999). A Multimedia database
is a controlled collection of multimedia data supporting applications like video-on-demand and
multimedia document authority systems (Prabhakaran, 1997). Multimedia database management
system provides support to varying data types and facilitates multimedia database in creation, access,
query, storage and control of the multimedia data in addition to traditional DBMS functions.
MMDBMS also facilitates efficient storage and data manipulation and fulfils the following purposes:
Integration to ensure data is not duplicated when different programs are requesting data access.
Application programs are separated from management and database systems to maintain data

11 | P a g e
independence. Concurrency control ensures consistency with rules. MMDBMS ensures persistence,
privacy, recovery, and integrity control (Adjeroh and Nwosu, 1997; Zhang, 2002).

The following diagram shows components of a typical MMDBMS (Prabhakaran, 1997).

The physical storage view shows how huge data/objects of the multimedia database are stored in the
system. The storage and retrieval of multimedia data require the use of different techniques. The
conceptual data view provides interpretations of the stored data. The interpretation is derived from the
physical storage view of data. The indexing technique is used in this view which provides fast access
to stored data. Multimedia objects can be stored in different systems, and a distributed view is used to
access this data through communication networks such as computer networks. Queries are used to
provide users with a filtered view of only the required objects. Finally, the objects retrieved from the
database are presented in the user-view component of MMDBMS. The below diagram shows the
requirements of MMDBMS.

12 | P a g e
 D)
Some possible applications and types of multimedia for this client are as follows:

 Text:
The briefing documents of the historical events of a particular period consists of text. The documents
are saved in pdf format in the system. These secret sensitive documents are scanned and converted
into a digitised format for user access and security and data backup. The multimedia application
enables the user to have a quick lookup to required information and an instant display of information
about a certain topic or historical event that he/she wants. The document's text can be viewed in any
colour or format, or font size, conveying the text's exact meaning. Windows Help Engine is one such
application that helps make the accessed information easier and more understandable for the users
(Pavithra, 2018).

 Images:
They are a large set of high definition digitised original photographs to support the information/data
provided. It supports such information, which is difficult to explain otherwise because "a picture is
with a thousand words."

 Interviews, Video and Audio Clips:

13 | P a g e
These are another type of multimedia and are useful in providing additional information to the users.
MPEG is a common storage format for such multimedia types. The size, resolution and sound quality
of these clips play an important part. Multimedia applications and video editing software improve the
quality of sound, video resolution, encoding and decoding of these media files (Dochev et al., 2000).
GIS stands for Geographical Information System which is a framework for gathering, managing, and
analysing data. Although primarily used in geography, it can be used in other fields. History
Uncovered can use GIS to add more quality to their history packages, e.g. they can use the GIS to
show the area covered by Nazi Germany or the area covered by past civilizations. The GIS can be
used in virtual reality or animation. For example, how did the Romans conduct their trade or how did
Europeans colonize the US. While the evidence is available in books, linking it with GIS can offer a
better experience. The GIS can be used to remap/reconstruct cities and experience ancient life.

14 | P a g e
 Task 4

A)
According to Tan (2009), a distributed database (DDB) is an integrated collection of physically
distributed databases across sites in a computer network. Another term for sites is nodes. The nodes
are connected through a computer network. Thus, the DDBMS is a software/system that manages the
DDB. To form a DDBMS, the files must be logically interrelated, structured, and distributed across
multiple sites. The below picture shows a DDBMS. The nodes are the headquarters, manufacturing
unit, and sales department which are connected through a network.

Source: Oracle

As the DDBS aims to manage the DDB, it must have one interface for accessing/managing the stored
data. Thus, the DDBMS will aid the Board in its expansion plans. Without a DDBMS, it will be
challenging to manage the data from the regional offices. As a result, there will be a negative impact
on the decision-making. Tan (2009) states that as the data's storage is in places where it is needed,
there is a significant reduction in costs and communication gaps. For example, the regional office will
not need to contact the head office for data all the time. Instead, it can access it from its system.
Another benefit is that in any case of failure, other sites can access the data. In a centralized system, a
failure can bring the entire organization to a halt. Thus, if History Uncovered suffers a cyber-attack or
any other failure related to a centralized system, it can take a considerable time to go back online.

15 | P a g e
 B)

Replication (Active & Passive)

Replication

Replication is mainly used in DDBMS to provide fault tolerance. The replication process reproduces
the data at the intended targets.

The technology of SQL server replication design for copying and distributing data and database
objects from one DB to another. Also, keep the integrity of the data.

Replication Types

Replication is further divided into two types:

1. Active Replication
2. Passive Replication

Active replication is a process which has passed the same request to every replica.

 Each client’s request is sent to all the servers.

 Replication first named state machine replication.
 Introduced by Leslie Lamport.
 Process shown in diagram called deterministic.

Passive Replication

Passive replication is a process which has passed every request to single replica. Then transfer the
results to the other replica. As shown below:

16 | P a g e
  This type of replication using a single server that processes the client’s request.
 After processing the request, the primary server updates the other server and send
back the response to the client.
 The process is called non-deterministic.
 The disadvantage of passive replication is a slow response from/to the server.

Fragmentation
One of the important features of DDMS is Fragmentation or Data Fragmentation. Fragmentation is
useful to save space, avoid duplication, and better query response time.

 Fragmentation means to break or split the DB tables

 Using unique ID like FK, PK
 To Avoid Duplication
 Efficiently retrieve data from the DB
 Remove duplicate data from the DB

Fragmentation is divided into:

1. Horizontal Fragmentation

The table is divided horizontally into tuples or rows. For example, a company has 2500
records(employees) which can be fragmented (horizontally) into 25 fragments. Each fragment
consists of 100 employees.

2. Vertical Fragmentation

In vertical fragmentation, tables is divided into columns which are referred as a set or site. Every set
must have at least one mutual column, e.g. primary key attribute. The reason for this is that whenever
there is a requirement, the fragmented columns can be converted into a parent table. For example, the

17 | P a g e
employee table of History Uncovered contains the following details: ENO (PRIMARY KEY),
ENAME, ADDRESS, EMAIL & SALARY. Divide the table into two sets such as Set1 and Set2. As a
result, set1 Columns include ENO, ENAME & ADDRESS, whereas set 2 columns are ENO, EMAIL
& SALARY.

3. Derived Fragmentation or Hybrid Fragmentation

Derived Fragmentation using the concept of both Horizontal and Vertical Fragmentation to form
tables in DDBMS

Two-Phase Commit [Transaction Management]

In DBM, commit is known as saving data, while rollback is known as undoing changes. In a single
server the transaction management is easy, but it will much complex in distributed DBMS. Multiple
DBs with their unique IDs in different location and having their own data in this case the process can
be trickier.

A two-phase commit is used when a commit must be broken into two separate parts. It is a
standardized protocol used for this purpose. In a DDBMS, there is a requirement for a special object
known as a coordinator. It is responsible for arranges the activities, while synchronizing between
distributed servers.

Phase 1:

 For each server that needs to commit data to write its record means you need to log in
to the system
 If the server successful logged alert shows “OK”
 Else will show “Invalid Entry / Failure”

Phase 2:

 After a successful login to the system, Phase 2 begins.

 A signal is passed to each server with commit instruction by the coordinator.
 The coordinator receives a message about a successful implementation of the commit.
 In case of failure, the coordinator will send a rollback instruction.
 The servers send a message that rollback is complete.

18 | P a g e
 References

Abdellatif, T.S., Elsoud, M.A., and Ali, H.A. (2011) Comparing Online Analytical Processing and Data
Mining Tasks In Enterprise Resource Planning Systems. 8 (6), 14

Activ, D. (2020) ‘How To Setup A Data Cube’. [ 2020] available from

<https://powerbase.dotactiv.com/knowledge-base/how-to-setup-a-data-cube/> [2 April 2021]

Adjeroh, D. and Nwosu, K. (1997) ‘Multimedia Database Management - Requirements and Issues.’ IEEE
MultiMedia 4, 24–33

Brar, K. (2020) OLAP and OLTP — The Two Different Data Techniques [online] available from
<https://medium.com/analytics-vidhya/olap-and-oltp-the-two-different-data-techniques-
78c91e0ab9d6> [1 April 2021]

Braumandl, R., Claussen, J., and Kemper, A. (1998) ‘Evaluating Functional Joins Along Nested Reference
Sets in Object-Relational and Object-Oriented Databases’. in In Proc. of the Conf. on Very Large
Data Bases (VLDB. held 1998. 110–121

Cios, K.J., Swiniarski, R.W., Pedrycz, W., and Kurgan, L.A. (2007) ‘Databases, Data Warehouses, and
OLAP’. in Data Mining: A Knowledge Discovery Approach [online] ed. by Cios, K.J., Swiniarski,
R.W., Pedrycz, W., and Kurgan, L.A. Boston, MA: Springer US, 95–131. available from
<https://doi.org/10.1007/978-0-387-36795-8_6> [1 April 2021]

Dochev, D., Koprinska, I., and Pavlov, R. (2000) Multimedia Data Management  Characteristics and
Requirements. 14

Farh, J.-L.L. and Pi, X. (n.d.) Huawei the Ren Zhengfei Way: The ‘Tough Guy’ and His Corporate
Philosophy ^ CB0099 [online] available from <https://store.hbr.org/product/huawei-the-ren-zhengfei-
way-the-tough-guy-and-his-corporate-philosophy/CB0099> [2 April 2021]

Franco, T. (2019) OLTP vs OLAP: Main Differences [online] available from

<https://www.imaginarycloud.com/blog/oltp-vs-olap/> [1 April 2021]

García, S., Luengo, J., and Herrera, F. (2015) Data Preprocessing in Data Mining [online] Intelligent
Systems Reference Library. Springer International Publishing. available from
<https://www.springer.com/gp/book/9783319102467> [1 April 2021]

Goar, V., Sarangdevot, P.S., Tanwar, G., and Sharma, D.A. (2010) ‘Improve Performance of Extract,
Transform and Load (ETL) in Data Warehouse’. International Journal on Computer Science and
Engineering 2

Han, J., Kamber, M., and Pei, J. (2012) Data Mining: Concepts and Techniques [online] ed. by Han, J.,
Kamber, M., and Pei, J. The Morgan Kaufmann Series in Data Management Systems. Boston:
Morgan Kaufmann. available from
<https://www.sciencedirect.com/science/article/pii/B9780123814791000046> [1 April 2021]

Hobbs, L., Hillson, S., Lawande, S., and Smith, P. (2005) Oracle 10g Data Warehousing [online] Elsevier.
available from <https://linkinghub.elsevier.com/retrieve/pii/B9781555583224X50000> [1 April 2021]

19 | P a g e
Hosking, J.R.M., Pednault, E.P.D., and Sudan, M. (1997) ‘A Statistical Perspective on Data Mining’.
Future Generation Computer Systems 13 (2–3), 117–134

IBM (2020) What Is a Data Mart | IBM [online] available from <https://www.ibm.com/cloud/learn/data-
mart> [1 April 2021]

Kakish, K. and Kraft, T. (2012) ETL Evolution for Real-Time Data Warehousing. held 3 November 2012

Microsoft (2020) Database Normalization Description - Office [online] available from

<https://docs.microsoft.com/en-us/office/troubleshoot/access/database-normalization-description> [1
April 2021]

Paterson, J., Edlich, S., Hörning, H., and Hörning, R. (eds.) (2006) ‘From RDBMS to OODBMS’. in The
Definitive Guide to Db4o [online] Berkeley, CA: Apress, 363–392. available from
<https://doi.org/10.1007/978-1-4302-0176-2_13> [1 April 2021]

Pavithra (2018) MULTIMEDIA AND ITS APPLICATIONS. 10, 271–276

Plattner, H. (2009) ‘A Common Database Approach for OLTP and OLAP Using an In-Memory Column
Database’. in Proceedings of the 2009 ACM SIGMOD International Conference on Management of
Data [online] held 29 June 2009 at New York, NY, USA. Association for Computing Machinery, 1–
2. available from <https://doi.org/10.1145/1559845.1559846> [1 April 2021]

Prabhakaran, B. (1997) ‘MMDBMS Architecture’. in Multimedia Database Management Systems [online]

ed. by Prabhakaran, B. The Springer International Series in Engineering and Computer Science.
Boston, MA: Springer US, 177–182. available from <https://doi.org/10.1007/978-1-4615-6235-1_8>
[1 April 2021]

REDDY, G.SATYANARAYANA, SRINIVASU, RALLABANDI, RAO, M., and RIKKULA, S. (2010)

‘DATA WAREHOUSING, DATA MINING, OLAP AND OLTP TECHNOLOGIES ARE
ESSENTIAL ELEMENTS TO SUPPORT DECISION-MAKING PROCESS IN INDUSTRIES’.
International Journal on Computer Science and Engineering 2

Sabau, G. (2007) ‘Comparison of RDBMS, OODBMS and ORDBMS’. Informatica Economica XI (4), 83–
85

Shin, S.K. and Sanders, G.L. (2006) ‘Denormalization Strategies for Data Retrieval from Data
Warehouses’. Decision Support Systems 42 (1), 267–282

Siguenza-Guzman, L., Saquicela, V., Avila-Ordóñez, E., Vandewalle, J., and Cattrysse, D. (2015)
‘Literature Review of Data Mining Applications in Academic Libraries’. The Journal of Academic
Librarianship 41 (4), 499–510

Steinmetz, R. (2012) Multimedia: Computing Communications & Applications. Pearson Education India

Tan, K.-L. (2009) ‘Distributed Database Systems’. in Encyclopedia of Database Systems [online] ed. by
LIU, L. and ÖZSU, M.T. Boston, MA: Springer US, 894–896. available from
<https://doi.org/10.1007/978-0-387-39940-9_701> [1 April 2021]

Tupper, C.D. (2011) ‘21 - Object and Object/Relational Databases’. in Data Architecture [online] ed. by
Tupper, C.D. Boston: Morgan Kaufmann, 369–383. available from
<https://www.sciencedirect.com/science/article/pii/B9780123851260000218> [1 April 2021]

20 | P a g e
UNDP (2015) Sustainable Development Goals [online] available from
<https://www.undp.org/content/undp/en/home/sustainable-development-goals.html> [2 April 2021]

Vries, A.P. de, Apers, P.M.G., and University of Twente (1999) Content and Multimedia Database
Management Systems. Enschede: Centre for Telematics and Information Technoloy (CTIT)

Weikum, G., König, A., Kraiss, A., and Sinnwell, M. (1999) ‘Towards Self-Tuning Memory Management
for Data Servers’. IEEE Data(Base) Engineering Bulletin, v.22, 3-11 (1999)

Westerlund, P. (2008) Business Intelligence: Multidimensional Data Analysis [online] available from
<https://www.diva-portal.org/smash/get/diva2:1137039/FULLTEXT01.pdf>

Yıldırım, A. (2020) What’s the Difference between OLTP and OLAP? [online] available from
<https://medium.com/@yildirimabdrhm/whats-the-difference-between-oltp-and-olap-bdcafdffb1c3>
[1 April 2021]

Zhang, A. (2002) ‘Introduction to Multimedia Database Systems’. in Readings in Multimedia Computing

and Networking [online] ed. by Jeffay, K. and Zhang, H. The Morgan Kaufmann Series in Multimedia
Information and Systems. San Francisco: Morgan Kaufmann, 387–390. available from
<https://www.sciencedirect.com/science/article/pii/B9781558606517501200> [1 April 2021]

21 | P a g e