Content uploaded by Mahmoud H. Qutqut
Author content
All content in this area was uploaded by Mahmoud H. Qutqut on Oct 23, 2019
Content may be subject to copyright.
IJCSNS International Journal of Computer Science and Network Security, VOL.19 No.4, April 2019
47
Manuscript received April 5, 2019
Manuscript revised April 20, 2019
Overview of the Current Status of NoSQL Database
Yasmin Rasheed, Mahmoud H. Qutqut and Fadi Almasalha
Faculty of Information Technology, Applied Science Private University, Amman, 11931 Jordan
Summary
Nowadays with the accelerated development of the Internet and
Cloud computing, the fast growth of technology generates a
massive amount of data. Businesses and people generate these
data by using web apps, social media, and new technologies.
These data, in general, could be structured, semi-structured or
unstructured. Because of the different types of data and the big
data that is generated, there is a need for a database to be able to
store and process these data effectively to enhance the
performance when reading and writing. So, there is a need for a
new design for the database, and it is not suitable for storing,
analyzing and performing data in a relational database for big
data. In addition to that, many new challenges faced the
traditional relational database; especially in the applications that
required large scale and high concurrency such as search engines.
In response to that, NoSQL has developed to solve these types of
problem. NoSQL database has many advantages that make it
gain significant popularity over the last few years and used
widely. It read and write the data quickly, expands easily, low
cost and many other features. In this paper, we overview the
NoSQL database and its characteristics in the field of the Internet
of Things (IoT). We also provide two representative use cases of
using the NoSQL database in current technologies.
Keywords:
NoSQL; database; relational database; IoT.
1. Introduction
Relational database management system (RDBMS) has
been adopted since the’70s. That is why it is a mature
technology for storing data and their relationships [1]. Also,
every RDBMS must ensure four properties or
characteristics in the transaction that is known as an ACID
(Atomicity, Consistency, Insolation, Durability). Atomicity
where all the tasks of a transaction are executed or none of
them will be executed. Consistency is the operation takes
of the database from one consistent state to another equally
consistent. Where Insolation describes the effect of a
transaction is not visible to other transactions until it is
committed. Durability defines the changes made by
authorized transactions are permanent. These
characteristics have a cost, and it will generate a cost so
that they are guaranteed [2]. In addition to that, recently
big data analysis becomes the core of modern science and
commercial. It is generated by users from online uses such
as emails, videos, audios, images, logs, posts, search
queries, health records, social networking interactions,
science data, sensors and mobile phones and their
applications. These data are stored in databases; they are
either structured or unstructured form of data, so we face
some problems like as how to capture, store, manage, share,
analyze and visualize them via common database [3].
The main issues for researchers are that the data growth
rate exceeds their ability to design appropriate cloud
computing platforms to evaluate data and update the
workload problem. Because of these problems, there is a
need to modify the database as traditional relational
databases have proved to be weak for the distributed
environment. To solve the main issues and for better
performance and scalability, a new database released by
scientists [3] called NoSQL. NoSQL has appeared because
it is a flexible, scalable and schema-free database. NoSQL
means “Not Only SQL.” It provides storage and retrieval
mechanism with less constrained consistency models than
traditional relational databases [3].
Regarding NoSQL design, some concepts must be taken
into consideration such as the CAP theorem or Brewer’s
theorem. The designers of distributed systems suffer a
fundamental tradeoff limitation that they must choose only
two out of these three properties CAP, which are data
consistency, system availability or tolerance to network
partitions. This theorem states that it is impossible in a
distributed system to guarantee the following properties
simultaneously [1].
• Consistency: If this attribute is satisfied, once the
data are written, it guarantees that this data is
available and up to date for all users using the
system.
• Availability: This attribute means that the service
is offered continuously and without interrupt or
degradation within a specific time.
• Partition Tolerance: This property means that a
transaction or a process can be done entirely even
when there is a part of the network failed.
Eric Brewer in 2000 estimates that at any specific time
only two out of the three properties that we mentioned can
be guaranteed. After that, Gilbert and Lynch proved this
estimation; they conclude that in distributed systems only a
specific combination can be created; AP (Availability-
Partition Tolerance), CP (Consistency-Partition Tolerance)
or AC (Availability-Consistency). To this end, we provide
an overview of the NoSQL database and its characteristics
IJCSNS International Journal of Computer Science and Network Security, VOL.19 No.4, April 2019
48
in the field of the Internet of Things (IoT). We also present
two representative use cases of using the NoSQL database
in current technologies.
The structure of this paper is as follows. We will start with
a brief background of the NoSQL database and its
categories in Section 2. We will determine the difference
between the RDMS and NoSQL and the benefits of the
NoSQL database over RDBMS in Section 3. In Section 4,
we provide two representative use cases and applications
of the NoSQL database in current technologies followed
by a conclusion in Section 5.
2. Background
Over the last few years, NoSQL databases have gained
high popularity with both, developers who make new
systems and organizations who want to improve their
business. Both are trying to adapt their information
systems to meet today's data requirements. The leaders of
NoSQL databases were massive web companies such as
Google, Amazon, and Facebook to promote them build
and support their businesses. After they made the NoSQL
public and open source, other giant web companies such as
Twitter, Instagram and Apple started to use them [4].
• Developers are dealing with applications that
create high volumes of rapidly changing data
types; structured, semi-structured, and
unstructured.
• There are no more twelve to eighteen months in
the waterfall development cycle. Now a group of
teams works on sprint-agile, which has iteration
and generates code every week or two.
• Applications used by broad audiences required to
be always-on, accessible from many different
devices and scaled globally to millions of users.
• Instead of large monolithic servers and storage
infrastructure organizations are now moving to
scale-out architectures using open source software,
commodity servers, and cloud computing.
In these situations, relational databases were not designed
to deal with the scale and agility challenges that face
modern applications [3]. There are different reasons to
move to NoSQL database. Some reasons are mentioned
below [3].
• The growing of Big Data and high data velocity,
data variety, data volume, and data complexity.
Fig. 1 Types of NoSQL Database.
• Data is always available.
• Reallocation transparency.
• A new era of transactional capabilities.
• Data architecture is flexible.
• High-performance architecture.
• Highly Intelligence.
There are different types of NoSQL database which are
represented in Figure 1 and discussed below with an
example in each kind.
2.1 Key-Value Database
These databases assign a key to a value or set of values [5].
The keys are unique and atomic. These keys are used to
query for entries in the key value storage databases. The
Key-value store provides a hash table structure with key-
value pairs spread across several remote servers in a
distributed cluster. So, they can achieve the required
efficiency by providing fast random read/write requests
and flexibility to store data in the schema-less format.
Since pairs of different key values save a group from
irrelevant data, it avoids SQL join and group by operations
as well as foreign key references. Twitter use key-value
stores to store tweets using a unique Twitter ID. The
corresponding values may include the original message,
User ID and time of sharing. One of the cases of key-value
stores Amazon Dynamo DB. Figure 2 shows the key-value
database system.
Fig. 2 Key-Value Database.
IJCSNS International Journal of Computer Science and Network Security, VOL.19 No.4, April 2019
49
Amazon Dynamo DB - Dynamo DB uses key-value data
structures that are designed to scale quickly with a flexible
schema. Also, it supports querying, updating and storing
documents. We can write applications that store JavaScript
Object Notation (JSON) documents directly in Dynamo
DB tables by using the Amazon Web Services software
development kit (AWS SDK). Each item in Dynamo DB
(row) is a key-value pair which has a primary key attribute
that uniquely identifies each item. Dynamo DB is fast and
has high adaptability. Here, data encryption is not
supported; instead of that, they use https protocol for the
communication between the client and the server. Dynamo
and requests control authentication and approval must be
marked using Hash-based Message Authentication Code
(HMAC)-SHA256. Amazon dynamo gives a consistent
replication mechanism with consistency levels that are
what makes it a different approach [6]. Dynamo DB has a
unique feature which is that the task of the database
administrator is automated. It monitors issues of scalability,
provisioning, load balancing, reliability, elastic map reduce
integration and makes sure that there is replication
synchronously to ensure no data is lost [7]. Dynamo DB
has many features as described below [7].
• The users can specify how they want their
performance to be in terms of the number of reads
and writes per second. Dynamo DB will provide
services based on their requirement.
• Eventual consistency and durable consistency are
to types of readings offered by Dynamo DB. The
reads are eventually consistent by default, but the
user can choose strong consistency that after a
write operation the updated value will be visible.
• When there is any modification or change in the
level of provisioning, this does not lead to data
loss or disruption on the application program.
• Because Dynamo DB stores data using solid-state
drives rather than hard disk drives; retrieving the
stored data is much faster than other NoSQL
databases.
In the following, we describe cases that we can use
Dynamo DB in them.
• Applications that have simple create, update,
delete operations performed over an extensive
data set (e.g., online gaming).
• In Amazon's cart, they use it while doing online
shopping to store items.
However, Dynamo DB is not recommended for the
following cases.
• For applications that have many relational JOIN
operation or normalization of data.
• When the number of reads and writes per second
in any application get change very fast in a way
that passes the read/write specified limit.
2.2 Graph Database
These databases store the data in a graph structure. Data is
represented by edges and nodes, each with their features
and attributes. Most graph databases provide a useful
graphical traversal, even when the nodes are on separate
physical vertices. Lately, the graph database has received
much attention because of its applicability to social media
data. This opens the way for new implementations to
accommodate the current market. However, several
authors exclude graph databases from NoSQL because
they do not fully align with the relaxed model constraints
typically found in NoSQL implementations. However,
others include it because they are mostly non-relational
databases and have many applications nowadays [1].
Graph database includes Neo4j. Figure 3 shows the
concept of the graph database.
Neo4j-Nowadays many companies and organizations used
neo4j in different industries such as government, financial
services, technology, energy, retail, and manufacturing.
Neo4j is an open-source NoSQL database that provides an
ACID-compliant transactional back-end for applications
[8]. Neo4j uses graph model. Nodes and edges have
associated properties with them. The nodes can also be
linked with labels, and they classify according to their
roles [9]. They have used Neo technology for developing
Neo4j, the implementation was in Java, and other software
are written in other languages can access using Cypher
Query Language (CQL) by using a transactional HTTP
endpoint, or through the binary “bolt” protocol. The main
features of CQL are described below [4].
• The way it works to extract information or modify
the data that matches patterns of nodes and
relationships in the graph.
• It deals with parameters, restricted elements, and
variables that indicate named.
• CQL can create, update, and remove nodes,
relationships, labels, and properties.
• CQL manages indexes and constraints.
Fig. 3 Graph Database.
IJCSNS International Journal of Computer Science and Network Security, VOL.19 No.4, April 2019
50
2.3 Document Database
Document-oriented model is used to store data as a record
with its linked data as a single data structure called a
document [4]. Each document contains multiple related
attributes and values. Documents can be retrieved based on
attribute values by using the various APIs or query
languages provided by the DBMS system their schema-
free organization of data characterizes them. Which mean
that the record does not follow a specific structure, i.e.,
different records may have different attributes. The types
of values of individual attributes can be different for each
record. To store the records; document stores use a format
such as JSON or XML. This helps the records to be
processed directly in applications. The documents are
stored and retrieved by using a key [4]. One of the
examples of a document database is MongoDB. Figure 4
shows the document databases system.
MongoDB - MongoDB is document free and open source
database published under the GNU Affero General Public
License [10]. The data in Mongo is stored in flexible
JSON-like documents. The fields could be different in
each document, and over time the structure of the data can
be modified [10]. Replication in MongoDB called replica
set; it is a collection of MongoDB servers that maintain the
same data set and provide data redundancy [4]. We
summarize the features of MongoDB below [7].
• MongoDB provides high performance by offer
indexing of every attribute in a document.
• It can scale-out without disrupting application,
and it supports sharding mirroring and load
balancing of data across nodes.
• It uses capped collections which are like the
concept of circular buffers that provide high
throughput performance.
Fig. 4 Document Database.
2.4 Column Database
This type of databases instead of saving data by row (as in
relational databases), they store data by column [1].
Therefore, some rows may not contain a portion of the
columns, providing flexibility in data definition and
allowing to apply data compression algorithms for each
column. Columns that are not frequently used or queried
together can be distributed across different nodes.
Cassandra DB is an example of a column database.
Cassandra-Cassandra is free and distributive open-source
broad column database developed by Apache [3]. The
language used to write Cassandra is java, so it can be used
by any platform that has a Java virtual machine (JVM) [3].
Cassandra is intended to handle enormous amounts of data
across several commodity servers [3]. So, this will provide
high availability with no single point of failure. To handle
fault tolerance, the data in Cassandra replicated
automatically to multiple nodes. The replication is
performed via various data centers. There is no downtime
when replacing failed nodes [11]. The following are some
features of Cassandra DB [7].
• Linear scalability is offered by Cassandra even if
the workload is enormous and the throughput
performance will not change while crunching an
extensive set of data.
• Gossip protocol is used by Cassandra to
communicate an update message to all replicas
simultaneously.
• In Cassandra reading, writing and updating are
simple. It uses built-in queries; hence, it provides
a good experience for the user.
Cassandra DB is highly recommended and should be used
in the following cases.
• In the case of apps that the number of reads is
more than the number of writes such as Twitter.
• There are some applications where immediate
consistency is not a significant concern. In this
case, it is recommended to use Cassandra.
• An application needs high maintenance of code.
• There are web applications that provide dynamic
schema and content to users such as Netflix.
3. NoSQL and the Relational Database
The NoSQL database maintains consistency models that
are constrained only to the relational database contrast to
collect and retrieve data [3]. Because of real-time web
applications and big data; NoSQL has been optimized for
use in various fields of engineering and traditional
industries. The primary process for NoSQL is to simplify
retrieval and to attach extensive data using data processing
IJCSNS International Journal of Computer Science and Network Security, VOL.19 No.4, April 2019
51
operations. NoSQL key features are horizontal scalability,
the simplicity of design and more delicate control over the
availability. These technical features demonstrate effective
and reliable results [12]. The relational databases load the
data vertically as shown in Figure 5. This is not the case in
NoSQL since it does not distribute data in this style. The
relational database does not spread logical entities across
multiple tables, as they are stored in one place. Relational
databases do not guarantee referential integrity between
logical objects. This feature enables them to distribute data
across a significant number of database nodes and write
independence [3].
NoSQL is an appropriate approach to deal with big data.
Companies like IBM, Amazon, Facebook, Twitter, Google,
Oracle, are now applying high performance conveyed
NoSQL arrangements. Comparing NoSQL with RDBMS,
NoSQL DB is more scalable and provide high
performance, and their data model addresses several issues
that the RDBMS is not designed to address [3]. Regarding
the Internet of Things (IoT) domain, over the last few
years until now the IoT has been used in many areas. IoT
concept indicates numerous smart devices that are
connected to the Internet [12]. IoT applications need to
serve a high number of users, quick response to all users
that are globally distributed, available all the time (no
downtime), deal with different types of data, semi- and
unstructured data [13]. These applications will generate a
massive amount of data and with the heterogeneous data
that are created; the problems to store, transfer and manage
the data efficiently will appear. If we use the RDBMS that
uses Structured Query Language (SQL) with these
applications, we will face its static schema which is the
main limitation that makes the RDBMS not suitable for
IoT applications. So, the NoSQL database is a schema-free,
no joins, and horizontally scalable database [7].
t
hat
Fig. 5 Relational DB vs. NoSQL.
SQL Databases: to store the data in the SQL database; the
relational data model is used. This model uses tables and
stores the data in rows and columns, and these tables can
be linked.
NoSQL Databases: NoSQL use non-relational data model
which is schema-free that store the data in different forms
as we mentioned earlier; document, graph, key value, and
column. NoSQL has gained significant reputation because
the features provided, such as high scalability, easy access,
and distributed architecture [12]. Table 1 shows the
differences between the SQL and NoSQL databases based
on the IoT point of view. We describe the aspects in the
following.
• Scalability: vertical scalability means that more
resources will be added to the node such as
memory or processors to enhance the
performance of that node. But in horizontal
scalability, the system load is divided into several
nodes (servers) without the need to have more
resources. Necessarily, in IoT application, the
database will need to expand, therefore choosing
the database that can grow will be a practical
choice.
• Data retrieval: In SQL the tables are connected.
To retrieve or to search for data from several
tables, the JOINS statements are used by the user
to view the data. On the other hand, the data in
NoSQL is stored in the form of objects that
contain all the related data. In this case, the
process of combining and then view the data will
be eliminated.
• System Maturity: SQL has been used for a long
time, and it is practised technology. Therefore,
most of the obstacles and issues have been solved.
Security features like authentication, data
confidentiality and integrity are incorporated in
SQL. On the other hand, NoSQL considered as a
new and not mature technology hence security
issues not solved yet, and that may generate more
security issues.
Table 1: SQL vs. NoSQL Databases.
Aspects
SQL
NoSQL
Scalability
Vertical scalability
Horizontal scalability
Data
Retrieval
Time consuming
process
Save response time
System
Maturity
Mature technology
hence highly secured
Not mature technology,
hence lack of security
4. Current Use Cases and Applications of
NoSQL Database
In this section, we present two current representative
applications and use cases of using the NoSQL database.
IJCSNS International Journal of Computer Science and Network Security, VOL.19 No.4, April 2019
52
4.1 Using a NoSQL graph-oriented database to store
accessible transport routes
Some organizations aim to enhance public transport
accessibility for people with disabilities, and this is what
was proposed by the World Health Organization in its
“World Report on Disability 2011” [14]. Different
websites and applications solve the problem of public
transport and its accessibility. But none of them offers
general mechanisms that achieve available transportation
data. There was a significant shortage of open and reusable
data concerning public transport and its accessibility.
Based on that, they proposed to develop a technological
framework that can process manage and exploit open data
that aim to raise the ability to access to city public
transport under the scope of the Access@city project [14].
They concentrated on the design and storage of accessible
transport routes, by using crowdsourcing techniques, in a
NoSQL graph-oriented database. So, they defined an open
data repository for convenient public transport under the
Access@city project. They used a NoSQL database to
develop it because of its high ability to deal with and
manage vast volumes of information in addition to its
scalability and flexibility. They have chosen a NoSQL
graph-oriented database specifically; because they are
going to deal with highly connected data and there will be
queries that are more efficient in a graph-oriented database.
By using a methodological approach, they developed and
designed the graph-oriented database from scratch. They
have selected Neo4j which is the most popular graph-
oriented database according to the database ranking. For
that, they have developed a native Android application; the
application users can register for the generation of
accessible routes. There is starting a route option that they
can use which have choices about the particular need
(wheelchair, bike, baby stroller, etc.) they will have in the
journey. The application will periodically register the GPS
position. There are two options the users have when the
trip is finished either discard the route or save it.
Comments about the paths taken may be added by the
users in addition to that they can tell about possible
incidents and/or including photos. In this case, they have
decided to choose the graph-oriented database for the
development of big data repository, that is because the data
of the routes is highly connected and there are many ways
will be used for querying the data [14].
4.2 Distributed architecture of mobile GIS
application using the NoSQL Database
As a college student that on campus, you would like to
know about every event, or any occasion happened on the
campus. The idea here is to benefit from the distributed
architecture of a mobile GIS (geographic information
system) application [15]. A group of students built a GIS-
based app for their campus. The app aims to provide
location-based information to the students on the campus,
information such as events, maps or any other useful or
related information. The application keeps track of the user
(student) current location. Once the user enters a
predefined polygon structure such as a building that is
already stored in the database, any relevant information
about that building will be shown to the user. It could be
an event in that building or a workshop hosted in a
particular floor or a room. For the database design, they
looked for a database that has high flexibility, ease of use
and quick deployment. So, they chose NoSQL MongoDB
because of the significant popularity of NoSQL databases.
They decided to benefit from the advantages of this
relatively new technology for their application. In their
case, the college campus is periodically changing the
layout of building they require a flexible database to store
the building location information so that is why the choose
NoSQL database because it will offer a straightforward
future change to their data models as the campus changed
over time. They used MongoDB, and it was preferable
over SQL because of easy scalability, quick startup
development, and ease of creating flexible data models
[15].
5. Conclusion
In this paper, we overview the current status of the NoSQL
database. We described all NoSQL database types and
presented their applicability to different domains along
with an illustrative example of each type. Also, the
applicability of the NoSQL database for the Internet of
Things (IoT) domain is discussed. We detailed two current
representative use cases of the NoSQL database.
Furthermore, we have noticed that people and
organizations are moving on SQL to NoSQL because of
the high performance and scalability and other features
provided by NoSQL. NoSQL database has many features
in the perspective of the massive amount of storage
management and their utilization. However, security is a
significant concern for IT infrastructure. Security in
NoSQL databases is weak; authentication and encryption
are non-exist or very weak. There should be a new solution
to enhance security to improve the use of resources in the
future.
Acknowledgment
This work made possible by financial support from
Applied Science Private University in Amman, Jordan.
IJCSNS International Journal of Computer Science and Network Security, VOL.19 No.4, April 2019
53
References
[1] A. Corbellini, C. Mateos, A. Zunino, D. Godoy, and S.
Schiaffino, Persisting big-data: The NoSQL landscape,
Information Systems, 63, pp. 1–23, 2017.
[2] F. Oliveira, A. Oliveira, and B. Alturas, Migration of
relational databases to NoSQL-methods of analysis,
Mediterranean Journal of Social Sciences, 9 (2), pp. 227–
235, 2018.
[3] A. Haseeb and G. Pattun, A review on NoSQL:
Applications and challenges. International Journal of
Advanced Research in Computer Science, 8 (1), 2017.
[4] W. Hauger and M Olivier, NoSQL databases: Forensic
attribution implications, SAIEE Africa Research Journal,
109 (2), pp. 119–132, 2018.
[5] Amazon official website, Amazon web services (AWS) -
cloud computing services, https://aws.amazon.com/.
[6] A. Corbellini, C. Mateos, A. Zunino, D. Godoy, and S.
Schiaffino, Persisting big-data: The NoSQL landscape,
Information Systems, 63, pp. 1–23, 2017.
[7] P. Srivastava, S. Goyal, and A. Kumar. Analysis of various
NoSQL database, in International Conference on Green
Computing and Internet of Things (ICGCIoT), pp. 539–544,
IEEE, 2015.
[8] Neo4j official website, what is a graph database and
property graph - neo4j, https://neo4j.com/developer/graph-
database/.
[9] A. Gupta, S. Tyagi, N. Panwar, S. Sachdeva, and U. Saxena,
NoSQL databases: critical analysis and comparison, in
International Conference on Computing and
Communication Technologies for Smart Nation (IC3TSN),
pp. 293–299, IEEE, 2017., 2015.
[10] MongoDB official website, MongoDB for giant ideas,
https://www.mongodb.com/.
[11] Apache Cassandra by intracluster official website:
https://www.instaclustr.com/.
[12] S. Rautmare and D. Bhalerao. Mysql and NoSQL database
comparison for IoT application, in IEEE International
Conference on Advances in Computer Applications
(ICACA), pp. 235–238, 2016.
[13] Couchbase official website, why NoSQL database?
https://www.couchbase.com/resources/why-nosql.
[14] B. Vela, J. Cavero, P. C ́aceres, A. Sierra-Alonso, and C.
Cuesta, Using nosql graph oriented database to store
accessible transport routes, in EDBT/ICDT Workshops, pp.
62–66, 2018.
[15] J. Rodriguez, A. Malgapo, J. Quick, and C. Huang,
Distributed architecture of mobile gis application using
NoSQL database, Journal of Computing Sciences in
Colleges, 33 (3), pp. 68–68, 2018.
Yasmin Rasheed is an MSc student at the Faculty of Information
Technology at Applied Science Private University in Amman,
Jordan. Yasmin received her BSc degree in computer science
from Applied Science Private University in 2017. Her research
interest focus on security for massive IoT.
Mahmoud H. Qutqut is an Assistant Professor at the Faculty of
Information Technology at Applied Science Private University in
Amman, Jordan since October 2014. He received his Ph.D.
degree from the School of Computing at Queen’s University in
Canada 2014, under the supervision of Prof. Hossam Hassanein.
He received his MSc degree in Telecom Systems from DePaul
University at Chicago, Illinois in 2007 and BSc degree in
computer systems from Applied Science University in 2015. He
has served as a TPC co-chair and a technical program committee
member for several IEEE international conferences. His research
interests include mobile heterogeneous small cells networks,
Internet of Things (IoT) enabling technologies, and smart cities
enabling services.
Fadi Almasalha is an Associate Professor at the Faculty of
Information Technology at Applied Science Private University in
Amman, Jordan; received his M.S. in Computer Science from
New York Institute of Technology, in 2005 and Ph.D. in
Computer Science from the University of Illinois at Chicago, in
2011. In fall of 2011, he joined the Department of Computer
Science at the Applied Science University. Dr. Fadi Almasalha
received his Associate rank in 2016, during his appointment as
the head of the computer science department. Dr. Fadi has
published more than ten technical papers, journals and book
chapters in refereed conferences and journals in the areas of
multimedia systems, data mining, and cryptography.
