cs. 3 newlgu.edu.pk/research/images/pdf/computer-science-and-information-technology/volume-2...aftab...

Awais Salman Qazi ([email protected]), Waqas Ahmad

Performance Comparison of Real-Time Video Conferencing In MPLS and IP Networking Environment 7-9

Aftab Ahmad Malik, Mujtaba Asad, Waqar Azeem

Artificial Intelligence A Byproduct of Natural Intelligence and Their Salient Features 1-6

Muhammad Hassaan Rafiq, Shiza Gul Niazi, Subaika Ali

Comparative Analysis of Urdu Based Stemming Techniques 11-14

Muhammad Arslan Tariq, Rafaqat Alam Khan

A Review on Skin Cancer Data Using Image Processing 29-32

Rafaqat Alam Khan, Muhammad Arslan Tariq

A Survey on Wired and Wireless Network 19-28

Muhammad Zunnurain Hussain, Muhammad Zulkifl Hasan, Zaka Ullah, Taimoor Hassan,

Vehicular Data Cloud Services 15-18

Areej Fatima, Sagheer Abbas, Muhammad Asif,

Cloud Based Intelligent Decision Support System for Disaster Management Using Fuzzy Logic 33-42

Waqas Ahmad, Awais Salman Qazi,

Statistical Power Profiling of Various Network Switches 43-47

LGURJCSIT

Volume 2 Issue 3 July - Sept.

1. INTRODUCTION

This paper mainly is concerned with artificial intelligence based on natural intelligence. To develop the thesis of this paper, we briefly introduce a few preliminaries first. It is described in [1] that due to the advent of I.T and modern computing facilities the efficiency, reliability and accuracy of machines and related systems has enhanced. The tasks performed by humans are formulated in logical order to constitute Algorithms to make the machines work. In order to facilitate the further discussion

and presentation of the thesis of this paper, we first introduce some preliminaries related to Artificial Intelligence, Natural intelligence and finally the "plant intelligence [2]". According to [2], synapse is at the intersection of neuron and target cell, where the communication between these cells occurs by releases a chemical to send the message to receptor. The target cell is also called the secretary cell, the communications of the neurons is also carried out by electrical synopses, also termed as electrical connections. According to [1], behind machines the human knowledge, intelligence and wisdom works. The

Artificial Intelligence A Byproduct of Natural Intelligence and Their Salient Features

1 2 3Aftab Ahmad Malik , Mujtaba Asad , Waqar Azeem1 Professor Department of Computer Science, Lahore Garrison University (LGU), Pakistan

2 PhD Scholar, School of Electronics Information & Electrical Engineering, Shanghai Jiao Tong University Shanghai China

3Senior Lecturer, Department of Computer Science Lahore Garrison University (LGU), Pakistan1 2 [email protected] , [email protected] , [email protected]

Abstract

This paper mainly focuses on the creation of Artificial Intelligence (AI) using natural intelligence but the question to be considered whether the natural intelligence can be created using artificial intelligence or not. The Artificial intelligence is the outcome of functionality and capabilities of human brain called neural Network. In this paper, it is presumed that the artificial intelligence is a by-product of natural intelligence and then we discuss some relationship between both of these, especially the working of natural intelligence. Some other important questions are raised to understand a deep linkage between natural and artificial intelligence. There exists lot of non-material phenomenon created by dint of natural intelligence (not created by human) causing to produce systems run by artificial intelligence theorems and algorithms working at backend. The software based on Knowledge Based Systems (KBS) derives its power from human wisdom and natural intelligence. There are several limitations on artificial intelligence. In creation of natural intelligence there is a great role of spirituality. Humans are creator of artificial intelligence with limited abilities. Actually AI started with invention of machines. The applications of creation of natural intelligence are vastly and abundantly known to humans of 21st Century, which are incorporated in the areas of Space Science, Anatomy, and motion of Plants, spin of electron, Electronics, plant intelligence and Neural Science etc. The working of machines depending upon the artificial intelligence doesn't provide creativity or self-motivated innovations, within the meaning of natural intelligence.

Keywords: Artificial intelligence, Natural intelligence, Spirituality

LGU R. R.J. Computer Sciences IT 2(3) LGURJCSIT

LGURJCSITISSN: 2519-7991

Ahmad et al LGURJCSIT 2018

LGU R Jounral forComputer Sciences IT

Vol. 2 Issue 3, July - September 2018

machines cannot suo-moto produce creativity, which is a logical activity and is beyond reason. However, a machine follows the scheme prepared by human engineers and performs its functions in a prescribed manner what we call artificial intelligence. Human intelligence works with Natural intelligence. Artificial intelligence is used when human brain cannot per-form difficult jobs.

2. THE ARTIFICIAL INTELLIGENCE STATE OF AFFAIRS

Artificial intelligence has been designed and created by human beings. The Artificial Intelligence effectively embraces the areas such as machine learning and Human Computer Interaction. The major working areas of artificial intelligence [5] are Knowledge representation, Knowledge bases Systems (KSB), Knowledge in learning, Knowledge agents, intelligent agents and their behaviour, searching techniques, optimal and decision making process in games. However, the area based on theory of probability is frequently applied, what is called probabilistic reasoning over time to make single and complex decisions. Other areas of most significance in Artificial Intelligence [5] are Image formation and processing, communication, natural language processing, Robotics and Robot application. The machines have no compassions of their own. The designed systems in Robots can make them feel, but they have no passion of love. Entirely, the Artificial Intelligence is the outcome of human abilities and therefore machine intelligence is a result of Artificial Intelligence. According to [6], the usefulness of inductive logic programming for the problems of learning, reasoning and knowledge representation is of immense importance. The framework of learning programs is highlighted in [6]. The neural network ensembling has been discussed in [7] where several networks are simultaneously utilized to solve a problem and the activity of ensembling has been described with use of genetic algorithms. The useful 'Bagging and boosting' ensembling approach has been discussed in [7]. The applications of integrated Artificial Intelligence in the area of machine learning and Robotics have been explained in [8] using engineering theory in electrical and mechanical. Further [8] elaborates the use and integration of electrical engineering and mechanical engineering. Fortunately, nowadays a new discipline has emerged known as

Mechatronics Engineering. The theory and applications of "Argumentation" in Artificial intelligence plays an important role which is highlighted in [9] particularly with respect of "Similarities" and "Differences", which formulates the bases of Artificial Intelligence. The famous algorithms in Artificial Intelligence algorithm are Nearest Neighbour, Naive Bayes, Decision Trees, Linear Regression and algorithms for Support Vector Machines (SVM) and Neural Networks. The scientists are replicating the natural human intelligence to convert to Artificial intelligence, where, special algorithms are working in driving a car without driver or flying an airplane without pilot. The important machine learning Algorithms have been designed for unsupervised, semi-supervised, supervised learning, and reinforced learning. The systems working with Artificial Intelligence Algorithms have fabulous speed, operate 24 four hours seven days a week. They are less biased and work with more accuracy and prediction; for example in stock market applications. The Artificial Intelligence is f r e q u e n t l y a p p l i e d i n t h e a r e a o f Communication, E-commerce, Human Resource Management, healthcare Cyber security, Logistics, Supply Chain and most impotently manufacturing and assembling, for example Computer and Electronics hardware and Car Assembling.

3. HOW HUMAN BRAIN WORKS

The brain performs five well known functions planning, handling feelings, emotions, controlling social interactions and most importantly the element of creativity. It controls all functionalities of the human body.

It is elaborated in [3] about the ensembling of neural networks whereas [4] highlights several aspects of artificial intelligence, working of human brain, comparison of human brain and computer. The Surface area of the adult human brain is nearly 2500 cm2 which works with about twenty billion neurons and 240 trillion synapses having 15 um size of neuron, size of one synapse is 1 um. It is well known that motor system of the brain plays pivotal role and undertakes the major tasks such as movement control and send signals to motor neurons to activate the muscle actions. The muscle connections pass through the spinal cord and action messages travel to torso and limbs.The following Figure (1) shows the different parts of human brain of diverse actions.

LGU R. R.J. Computer Sciences IT 2(3) LGURJCSIT 2

Figure 1: Structure of Human Brain

3.1. The Element of Natural Intelligence

The element of natural intelligence exits in all creations such as planets, earth, human beings, animals, plants and other species living in sea have been created and induced by the creator of the universe. The difference between the alive and dead body is the “soul” present or absent. The soul makes the living beings to perform their usual functions. Science has not been able to design and make a living cell capable reproduction of another living being. The DNA and RNA can be perhaps synthesized but having no traces of soul. The natural intelligence works with limitations varying from one being to another.

3.2. Guilford Structure of intelligence

According to [10] and [11] there are 150 component of intelligence regarding operations, products and the contents represented by each cube in the Guilford Model shown in Figure(2) .

Figure 2: Guilford Structure of Intelligence [10],[11]

3.3. Sternberg's Triarchic Theory

This Theory deals three abilities for intelligent behaviour such as analytical, creative and practical. It advocates more importance and focus on environment and learning.

Figure 3: Sternberg's Triarchic Theory [15], [16]

3.4. Carroll 's Three-Strata Model of Intelligence

In Carroll's Model three are mainly three Stratum or layers such as general, broad and narrow abilities. Their components are shown in Figure (4).

Figure 4: Carroll's Three-Strata Model of Intelligence [14], [15]

3.5. Some important Questions for the reader of this paper

In order to establish the thesis of this paper, authors are obliged to raise the some questions in this section. The purpose of these questions is to think inside of the creation of Natural Intelligence, which gives rise to artificial intelligence. Most of questions are on the products of natural intelligence. These phenomenon shown in Table 1 are the state of brain. For example, during the laughter endorphins and cortisol are released in the brain and one feels happiness. But the question still remains to be answered that who designed these and using what?

Table 1: Human Brain Phenomenon and different states of brain.


The above mentioned elements in Table 2 are not made up of any matter but have deep impact on our life and that give rise to artificial intelligence. In order to further explain the idea of Natural Intelligence, following questions are raise having trivial answers.

4.1. Cell Destiny

The life originates from parent's Cells and sometimes from single parent cells by the germ cells and hence by the combination and formation of Zygote which later on becomes embryo. According to our basic knowledge, the embryo after development is responsible for the various organs. The multi-cellular organism issues the instructions to cells and decides what limb is to be prepared. The genes which determine what the cell is to become, are known as Hox-Genes. About 10 to 20 Million Cells are born per second and the same number of old cells is automatically destroyed.

4.2. Who designed, made and Controls the following natural systems

Q: Ten 10 billion Galaxies and one hundred Billion stars per Galaxy?Q: Who induced Spin and revolutionary motion about an axis in the Planets?Q: How the planets, stars and galaxies remain in order while travelling in Universe?Q: Why planets and stars don't collide with one another? Do they have intelligence? Q: How the resultant force of gravity of a planet is balanced with other planets?Q: How the angle of inclination of Earth, while rotation is maintained at 67.5O?Q: How the trajectories of the revolution of the planets are disciplines as per prescribed route?

Q: How the motion of Electrons about nucleus of atom in various states is maintained?Q: How the nucleus of Uranium atom has been designed to contain enormous amount of energy? Who has designed it? One kilogram of U-235 possesses 2 to 3 million times more energy as compared to 1 Kg coal or oil. The energy released from 1 Kg U-235 is approximately 24X106 KWh.Q: Who has programmed the bee to fly over valleys suck the fluid from flowers and convert it to mouth of other bees in their house on mountains and trees? The bees keep of transferring the fluid to other bees until its moisture is reduced from 37% to 20%. This how pure honey is produced.Q: Who has programmed the seed of the plant to possess a database and a mechanism for the embryo to grow when appropriate condition of temperature, moisture and suitable soil is available? How the seed knows when to grow? What sort of knowledge it possesses to decide how tall would be the stem and what type fruit or flower to produce.Q: Who has designed the Knowledge, wisdom and intelligence which is induced in protoplasm (the embryonic) material placed in eggs of humans, animals, birds and other species? Q: Can artificial intelligence design a living cell for without using natural cell?Q: How natural intelligence is designed to work to produce a living cell, DNA and RNA?Q: How the parts of human, animal or bird's babies are automatically developed in the womb or Egg?

4.3. Genetic Modification

The Gene is fundamental functional unit [13], which determines the heredity. It has been designed and created from DNA. The knowledge of genetic manipulation and modification of a living organism, using the processes prescribed in Genetic Engineering and biotechnology has expanded and efficiently producing results. For example cloning technique is highly developed.

Figure 6: Structure of Chromosome


Table2: Natural Intelligence and Statistical Information about Human Body [20],[21]

4

4.3.1. Cloning Technique- a combination of Artificial and Natural intelligence

In order to further elaborate the preposition of the paper, now we discuss systems where a combination of Artificial Intelligence and Natural Intelligence work together, for example Cloning Technique. This enables [12] us on to produce multiple copies of the same person, animal or other species. The Cloning and genetic engineering has been successfully applied on plants, using natural sources. The Soybeans and corn resulted in lecithin as a food experiments called Soybeans cloning.

Figure 7: Similarities in different human shapes Source: The Centre for Bioethics and Culture (CBC)

Using reproductive cloning, identically the same specie through the technique called somatic cell nuclear is transferred to the incubator. In this manner [12], a new embryo is developed to be injected into uterine environment, what we call incubator. Similarly, using genetic engineering meaningful changes in the nucleus membrane of the cell can be made to produce parts of human body or limbs. Human cloning is therefore, producing genetically same copy of the Human. Using this method, cells and tissues can also be produced for transplantation.

4.4. Shapes and Colours of humans and other species are due to Natural Intelligence

The basic driving force in Natural Intelligence is not sensibility, therefore, it is evident in ants, hyenas and humans [13]. It is important to visualize how the different colourful and symmetrical designs are formed inside the mother's womb or egg? Following are some relevant images shown in Figure:

Figure 8: Shape and Structure of different Species


[6] Mark Law, Alessandra Russo Krysia Broda (2018), "The complexity and generality of learning answer set programs", Artificial Intelligence, Volume 259, Pages 110- 146. https://doi.org/10.1016/j.artint.2018.03.005Get rights and content.[7] Z h i - H u a , Z h o u J i a n x i n , WuWeiTang(2002) "Ensembling neural networks: Many could be better than all", Artificial Intelligence, Volume 137, Issues 1-2, Pages 239-263 https://doi.org/10.1016/S0004-3702(02)00190-XGet rights and content[8] KannaRajan, Alessandro Saffiotti (2017), "Towards a science of integrated AI and Robotics", Artificial Intelligence, Volume 247, J u n e 2 0 1 7 , P a g e s 1 - 9 . https://doi.org/10.1016/j.artint.2017.03.003Get rights and content. [9] T.J.M.Bench-CaponPaul E.Dunne (2007) , "Argumentation in artificial intelligence", Artificial Intelligence, Volume 171, Issues 10-15, July-October 2007, Pages 619-641[10] Guilford, J.P. (1950), "Creativity". American Psychologist, 5, 444-454.[11] Guilford, J.P. (1967), "The Nature of Human Intelligence". New York: McGraw-Hill[12] Robert P. Lanza, , Jose B. Cibelli & Michael D. West (1999)," Human therapeutic cloning", Nature Medicine volume5, pages 975-977 (1999) |[13] N a t u r a l I n t e l l i g e n c e . www.cs.bath.ac.uk/~jjb/web/uni.html[14] Sternberg, R.J.(1977)," Intelligence, Information Processing, and Analogical, Reasoning. Hillsdale, NJ: Erlbaum.[15] Rober t J S tenburg , "Cogni t ive Psychology, 4th Edition Chapter 13.[16] What is a gene? - Genetics Home R e f e r e n c e - N I H , https://ghr.nlm.nih.gov/primer/basics/gene[17] Karpi?ski S, Szechy?ska-Hebda M. Secret life of plants: from memory to intelligence. Plant Signal Behav 2010; 5:1391 - 4; http://dx.doi.org/10.4161/psb.5.11.13243; PMID: 21051941[18] Cvrcková F, Lipavská H, Zárský V. Plant intelligence: why, why not or where?. Plant S i g n a l B e h a v 2 0 0 9 ; 4 : 3 9 4 - 9 ; http://dx.doi.org/10.4161/psb.4.5.8276; PMID: 19816094 ; [Taylor & Francis Online[19] Michael Marder (2013) , "Plant intelligence and attention; Journal of Plant Signaling and behavior, Volume 8, Issue 5 .https://doi.org/10.4161/psb.23902[20] Elaine N. Marieb , Patricia Brady Wilhelm , Jon B. Mallatt , "Human Anatomy (7th Edition), ISBN-13: 978-0321822413; ISBN-10: 0321822412[21] Human Body, A Visual Encyclopaedia, Published by D.K Publishing

4.5. Plant Intelligence

The secrets of plant life, intelligence, memory and their signal behavior has been discussed in [17] [18] and [19]. The plants exhibit intelligent behavior such as perdurance, modulation, selectivity and attention subject to their cognitive ability. The seeds of plants are produced fully intelligent to grow and adapt their destination according to database information they contain. They follow up prescribed procedure to produce the outcome. Their conduct is intelligent due to built in components. They possess communication components, for example, the Sunflower possesses the intelligence to move its flower towards the direction the sun.

5. CONCLUSION

The deliberations are arguments presented here have been directed to prove the basic philosophy of the thesis of this paper, which is that the Artificial Intelligence is created by Natural Intelligence where as several components of Natural Intelligence have not been created by man. Few questions in support of the assertion have been raised for the learned readers. More dominant is the Natural Intelligence.

Acknowledgment

The authors acknowledge Mr. Kaukab Jamal Zubari Director DRFSC for his motivation.

References

[1] Acharya Prashant, "Artificial ntelligence and Real intelligence" Advait.org.in [2]https://www.medicinenet.com/script/main/art.asp?articlekey=9246.[3] Zhi-Hua Zhou, Jianxin Wu,| Wei Tang (2002) ," Ensembling neural networks: Many could be better than all ", Artificial Intelligence , Volume 259, June 2018, Pages 110-146[4] Engr. Mujtaba Asad , " Functionality Ar t i f i c i a l I n t e l l i gence and Cur r en t Technologies", A Research Paper presentation, Department of Electronics and Electrical Engineering, Shanghai Jiao Tong University, Shanghai Minhang Campus, China.[5] Stuart J Russel, Peter Norvig (2018), "Artificial Intelligence"; Pearson.


1. INTRODUCTION

In today's technological world, real time applications are becoming very popular. There is an enormous growth in the multimedia content of voice and video traffic over the Internet. The voice and video traffic demands intelligent utilization of resources because of delay sensitivity, less Quality of Service (QoS) requirements and limited bandwidth [3]. So it is imperative to determine suitable networking mechanism which provides efficient routing of packets and best performance over both data and real time applications. (MPLS) Multiprotocol Label Switching is such a mechanism that brings lower delays in the delivery of packets over the network.

In this paper, a research work is done to compare the performance of MPLS and traditional IP networks (Open Shortest Path First) OSPF in video conferencing applications considering similar network topologies simulated in OPNET. Performance evaluation of comparing throughput & end-to-end delay shows that MPLS outperforms the traditional IP networks.

2. SURVEY OF RELATED WORKS

Performance of MPLS and Non MPLS is measured in [1] by applying traffic engineering on both the networks in the heavy traffic scenarios. It was shown that MPLS extends the improvement scale of network performance for variable heavy traffic platforms.

In study of performance evaluation[2] comparing constraint based Label Distribution Protocol (CR-LDP) and Resource Reservation Protocol (RSVP) of MPLS, simulations provided the results which confirmed that the use of CR- LDP protocol in the MPLS network would result in better performance and could overcome the drawback of scalability in Resource Reservation RSVP) protocol.

Taking File Transfer Protocol (FTP) into cons ide ra t i on , IP, MPLS and ATM performances are evaluated on OPNET simulator in [3] and confirmed the out performance of ATM and MPLS [4] in comparison to pure IP by judging delay and response time parameters.

Performance Comparison of Real-Time Video Conferencing In MPLS and IP Networking Environment

Awais Salman Qazi ([email protected]), Waqas Ahmad([email protected])Department of Computer Science Lahore Garrison University (LGU), Pakistan

Abstract

Modern techniques in networking have evolved a diverse range of opportunities to resolve the challenging task of traffic engineering. With a principle focus on video conferencing that has gained popularity worldwide, we contrive to optimize video traffic by using Multiprotocol Label Switching (MPLS) rather than conventional Internet Protocol (IP) networks. MPLS can be a promising technology for real-time applications with low network delays and efficient utilization of network resources. To substantiate this, in this paper we model certain scenarios of both IP and MPLS networks using Optimized Networking Engineering Tool (OPNET) to simulate and provide comparative performance statistics of throughput and end-to-end packet delay. We illustrate that MPLS has the ability that makes it stand out to IP by analyzing the simulation results of both networking mechanisms.

Key Words: Index Terms—IP, MPLS, OPNET



Qazi et al LGURJCSIT 2018



7

3. PROBLEM STATEMENT AND MAIN CONTRIBUTION

The real time applications like video conferencing will face delay or packet loss in the traditional networks. The aim here is to reduce high latency rate and overcome packet drops. With an aim to accomplish the required quality of service, the research question was outlined as “which of these networking mechanisms, MPLS or IP would perform better in terms of throughput and delay of the received packets in video conferencing applications”.

We hypothesized that MPLS would be a better performing networking mechanism compared to IP in terms of throughput and delay in video conferencing applications.

The main contribution of the paper is to evaluate the best networking mechanism for real-time video conferencing applications and to validate by OPNET Simulator that MPLS serves as a best platform against Conventional IP. Parameters chosen to analyze relative performance of video conferencing applications are throughput and end-to-end delay to accomplish higher performance and lower latency rate.

4. PROBLEM SOLUTION

4.1. Implementation

The network topology is designed and simulated by using OPNET. The MPLS network is designed using 7x Label Switching Routers (LSR) and 2x Label Edge Routers (LER). The LSR's are interconnected using PPP_DS3 Link with 44.7 Mbps speed. LER's are connected to LAN clients and video server using eth4_fddi4_tr4_switch_adv switch. The switch is connected to LAN client and video server using 10 base T Link with 10 Mbps speed. The configuration of applications is executed in profile configuration and these applications are mapped to clients and servers accordingly. MPLS parameters like Forward Equivalence Class (FEC) and Label Switched Path (LSP) are configured using MPLS configuration. CR-LDP is set as the routing protocol. In Fig.1, each packet enters at ingress label edge router and is assigned with labels that go along the LSP. The FEC determines the type of labels assigned at each LER. The addition and deletion of labels

are performed at ingress and egress edge routers respectively. The parameters in the simulation are mentioned in Table I.

Figure 1: MPLS Model

The IP network model is the replica of the MPLS model. The LSRs and LERs in the MPLS model are replaced by the IP Routers. And the routing protocol used is Open Shortest Path First (OSPF).

4.2. Results

The simulation is set to start at the 30th second and it is shown in an output graph. Fig.2. shows the end-to-end delay of MPLS network and the conventional IP network with end-to-end delay on vertical axis and simulation time on horizontal axis.

Figure 2: Video conferencing packet end-to-end delay.

The end to end delay threshold for video conferencing should be 150 ms, but it is acceptable up to 400 ms according to [5]. It is observed from the end-to-end delay graph that IP network crosses the threshold at 1 minute


References[1] M. K. Porwal, A. Yadav and S. V. Charhate, "Traffic Analysis of MPLS and Non MPLS Network including MPLS Signaling Protocols and Traffic Distribution in OSPF and MPLS," in 1st International Conference of Emerging Trends in Engineering and Technology, NAGPUR, 2008, pp. 187-192.[2] M. A. Rahman, A. H. Kabir, K. A. M. Lutfullah, M. Z. Hassan and M. R. Amin, "Performance analysis and the study of the behavior of MPLS Protocols," International Conference on Computer and Communication EngineeringICCCE, KUALA LUMPUR, May 2008, pp. 1-4.[3] H. M. Asif and M. G. Kaosar, "Performance Comparison of IP, MPLS and ATM based network cores using OPNET," International Conference on Industrial Information Systems, SRILANKA, Aug. 2006.[4] Madhulika Bhandure, Gaurang Deshmukh, Prof. Varshapriya J N / International Journal of Engineering Research and Applications ( I J E R A ) I S S N : 2 2 4 8 - 9 6 2 2 , U R L : http://www.ijera.com/. Vol. 3, Issue 4, Jul-Aug 2013, pp. 71-76[5] J. N. Boshoff and A. S. J. Helberg, "Improving QoS for real-time multimedia traffic in Ad-hoc Networks with delay aware multi-path routing," in Wireless Telecommunications Symp., CALIFORNIA, Apr. 2008, pp. 1-8.[6] Awais Salman Qazi, Waqas Ahmad, Performance Estimation of Real-time Video Conferencing in MPLS and Non MPLS Environment, International Journal of Networks and Communications, Vol. 8 No. 2, 2018, pp. 34-36. doi: 10.5923/j.ijnc.20180802.02.[7] Rohit Mishra, Hifzan Ahmad, Comparative Analysis of Conventional IP Network and MPLS Network over VoIP Application, International Journal of Computer Science and Information Technologies, Vol. 5 (3) , 2014, pp 4496-4499.

whereas the MPLS network crosses after 2 minute 10 seconds approximately.

In Fig. 3 green, red and blue lines indicates the average number of packets sent from source, average number of received packets in MPLS network and average number of received packets in conventional IP network respectively.

Figure 3: Video packets throughput.Throughput is an accomplishment rate of successful content delivery over a media channel. Average number of packets sent and average number of packets received in both the networks are shown in Fig. 3. The throughput graph shows that the average number of packets received in the MPLS network is more than average number of packets received in the conventional IP network for the same number of packets sent in both the networks [6], [7].

5. CONCLUSION

The performance of MPLS and IP networks are analyzed in real time video conferencing. The routing mechanisms are configured and compared. It is concluded that MPLS mechanism performs much better than IP as the throughput is enhanced and end-to-end delay is decreased in real time video conferencing applications.

The future work of our project can be extended to performance analysis of MPLS in real time video conferencing using high definition video mode.

Acknowledgment.

We would like to offer special thanks to our University colleagues and our parents for their support and guidance in making this project successful.


1. INTRODUCTION

A stemmer provides base, stem or root form from different modified forms of a words, for instance word “mixed”, and “mixing” will be reduce to stem i.e “mix” in English stemming. Stemmers are important for a number of application which requires the use of base form instead of derived form of words. Word Sense Disambiguation, Information Retrieval systems, word count studies, text mining, automatic discretization are few of the applications that make the use of stemming. Stemming of morphologically rich languages (MRLs) provides a greater impact as compared to less inflectional languages. For example an Urdu word has more than 60 inflectional forms. Urdu is spoken by approximately 200 million people and if we include Hindi speakers who speak Urdu in their Hindi words in their vocabulary then it will become second largest language with 588m speakers in the World. This sort of Urdu language will be known by linguists as Hindi-Urdu. The diversity and Morphological richness of Urdu language introduce a number of challenges in stemming. Vocabulary of Urdu contains many words taken from different languages that result into unavoidable complexity because these borrowed words have a tendency to be handle according to the grammatical principle of its source language. Some of the most prominent source language of Urdu vocabulary are Persian, Turkish, Sanskrit, Arabic, Hindi, and English. Here are some of

few examples explaining this characteristic of Urdu vocabulary. The words selected for these examples are normally used in daily routine Urdu language.�) ��)baag, garden(2) Sanskrit, (umang, aspiration)3) English, (television, television)�) ��)khatoon, woman(5) Arabic, (manzil, destination)6) Hindi, (aag, fire)

These examples taken from “six different languages” based on “six Urdu words”. The majority of Urdu vocabulary words belong to different foreign languages. Owing to these facts it is essential to adopt a customized approach for Urdu stemming in order to improve performance and accuracy. Hence, this approach is selected to handle the challenging nature of Urdu stemming. In Urdu stemming, every Urdu vocabulary word is stemmed manually according to the type of word. Before this selection, a thorough analysis of various Urdu stemming approaches has been performed. The possibility of further enhancement in its performance is also described. Rest of the paper is organized as follows: Section 2 presents the related work, in section 3 design approach of dictionary based stemmer is given and section 4 draws a conclusion.

Comparative Analysis of Urdu Based Stemming Techniques 1 2 3Muhammad Hassaan Rafiq , Shiza Gul Niazi , Subaika Ali

1 2 [email protected] , [email protected] , [email protected]

Abstract

Stemming reduces many variant forms of a word into its base, stem or root, which is necessary for many different language processing application including Urdu. Urdu is a morphologically rich and resourceful language. Multilingual Urdu words are very challenging to process due to complexity of morphology. The Research of Urdu stemming has an age of a decade. The present work introduces a research on Urdu stemmers with better performance as compare to the existing Urdu stemmer.



Rafiq et al LGURJCSIT 2018



further duplicated for transitive and causative forms and all these inflected forms make a total of more than sixty variant forms. A language with this much morphological variations and with a vocabulary that contains the majority of its words borrowed from more than half dozen languages cannot be handled successfully without a precise and more Urdu specific approach.

Table 1: Analysis of year wise related work.

3. PROPOSED METHODOLOGY

After the careful analysis of Urdu vocabulary and the research in Urdu stemming etc., the dictionary-based approach is selected as the most appropriate one. It is also called the lexical lookup, table look-up or the brute force approach. In this approach, a dictionary of word-

2. LITERATURE REVIEW

Stemmers can be developed by using a number of approaches such as rule-based, hybrid and statistical. Statistical approaches are usually not fully dependent on the prior linguistic knowledge of the concerned language; they make use of corpus analysis to calculate the occurrences of stems and affixes. On the other hand, rule based and hybrid approaches process the text according to framed rules derived from the grammar of the concerned language. Current state-of-the-art Urdu stemmer Assas-Band, developed by Akram et al., claims 91.20% accuracy which is the lowest among all stemmers as comparison is given in the Table. On the other hand Assas-Band is standing at the first position among all the reported Urdu stemmers since 2009 till 2016. This fact denotes the challenging nature of multilingual Urdu vocabulary. Table 1 provides a comparative view of various proposed solutions deploying statistical, rule-based and hybrid approaches regarding Urdu stemming. Statistical approaches presented in is, in fact, two studies on the basis of one work so, there is only one statistical approach proposed by the author. Authors in table 1 used word pattern matching schemes to handle infixes and they first time addressed the problem of Urdu words having infixes to some extent. However, the average accuracy of their results is just 77.39%. In the most recent work presented in 2016 authors used a rule based approach. Their approach is capable of handling Urdu loanwords and Urdu compound words. The overall accuracy of their stemmer is 88.91%. They declared the “Urdu Light Weight Stemmer” as current state-of-the-art, while it offers the lowest performance regarding accuracy. However, the discussed approaches do not provide remarkable results for Urdu stemming as shown in figure. The prime concern of this issue is the flexibility of Urdu language that introduces numerous exceptions against the framed rules (for rule-based approaches) and derived patterns (for statistical approaches). Following description could be helpful to unders tand the morphological richness of Urdu language. Urdu verbs have various inflections for habitual, infinitive, past, non-past and imperative forms and these verbs also inflect to show agreement for the case, respect, gender, and number. These twenty inflected forms of a regular verb are


stem pairs is prepared manually. Manual stemming is performed in a word-stem pair fashion according to the selected vocabulary.

Table 2: Urdu words with their Stemms

Along with the stemming, a stop word list is also developed. Stop words are non-content words or functional words that are usually ignored in search queries because no one uses them as query words and they can possibly index the whole corpus and hence make the search results entirely useless. Interestingly, there is no predefined criterion for stop words in Urdu. Different studies show different lists of stop words. A stop word in a language can be generally defined as a token that does not provide any linguistic meaning. Aqil Burney et al also mentioned that no proper work has been done on stop words in the Urdu language so they simply translated 421 English stop words in Urdu to use them as Urdu stop words. There are no guiding rules to identify stop words. Different studies proposes stop words lists of different numbers. In a study a list of 150 words declared as stop words list while mentioned a list of 200 stop words in their study. To avoid such vague approach regarding Urdu stop words in my study, I framed out some rules to decide either a word should consider as a stop word or not. The most important rule among all is that a stop word must neither be an inflected word nor a stem word because such type of stop word selection will create contradictions in any stemming approach. A sample list of collected stop words is presented in Table 3.

Table 3: Sample list of Urdu stop words

A list of special characters is also dveloped to filter out such characters from the input text in the text pre-processing stage.

Table 4: List of special creators

4. CONCLUSION

Stemming is important and basic component in many natural language processing applications. The accuracy of the stemmer strongly affects the results of the system in which it is used. Various stemmer development techniques are being explored and studied for different languages across the world. This paper surveyed the stemmers available for Urdu language. This paper presents a rule based stemming method for Urdu language. Proposed method has the ability to generate the stem of Urdu words as well as loan words (words belong to borrowed languages i.e. Arabic, Persian, Turkish, Hindi, etc).

REFERENCES:

[1] Akram QU, Naseer A, Hussain S. 2009. “Assas-Band, an affix-exception-list based Urdu stemmer” In Proceedings of the 7th Workshop on Asian Language Resources, Association for Computational Linguistics, 40-46.

[2] Gupta, V., Joshi, N., Mathur, I. (2016) Design and Development of a Rule-Based Urdu Lemmatizer Proceedings of International Conference on ICT for Sus ta inable D e v e l o p m e n t p p 1 6 1 - 1 6 9 , d o i . 10.1007/978981-10-0135-2_15

[3] Husain, Mohd Shahid, Faiyaz Ahmad, Saba Khalid. 2013 "A language Independent Approach to develop Urdu stemmer." Advances in Computing and Information Technology. Springer Berlin Heidelberg, 4553.

[4] Ali.M, K.S, S.H.M, 2014 “A Novel Stemming Approach for Urdu language “ISSN:20904274 Journal of Applied Envoironmetal and biological sciences, J. Appl Envoiron. Biol. Sci., 4(7S)436-443

[5] Khan, Sajjad Ahmad, et al 2012. "A Light



Weight Stemmer for Urdu Language: A Scarce Resourced Language." 24th International Conference on Computational Linguistics.

[6] Ali M, Khalid S, Haneef M, Iqbal W, Ali A, Naqvi G. 2016 A Rule based Stemming Method for Multilingual Urdu Text. International Journal of Computer Applications. 2016 Jan.

14

1. INTRODUCTION

Cloud computing is the process in which hardware and software will deliver a service through internet. The user can access any files of the computer through internet. Cloud services has projected a frame work to start cloud services computing which will declare a policy for a group which is based on information such as traffic restriction and controlling. A band of vehicles with corporate exchange information and practical work will be allocated by authorized users. Vehicular data cloud is used for the particular purpose with the large amount of Mobile devices and GPS devices. Vehicular data is used for networking, communication to the other vehicles and exchanged information with the other environment such as Internet protocol. The advancement in IOT has received a lot of focused which will facilitate transport industry. Location tracking and monitoring will be available to drivers. In past vehicular ad hoc networks is used for wireless communication in vehicular transportation system for security purposes.V2V is used to exchanging information for neighboring vehicles in roadside safety. I2V and V2I is used to transferring data through signals. Wireless Networking, Mobile Computing and Cloud Computing are key Technologies that enables for application that will support traditional cities to smart. Smart

Transportation safety in smart cities is used in attracting attention in industry research because of incident involving vehicular theft as well as violent crimes on urban buses. Crime Details are usually obtained in video-feeds and camera. Smart transportation safety is based on Process ing, s torage, mul t ip le media transportation for storing data through vehicles.

KEY WORDS

PAAS,SAAS,IAAS,VANETS, IOT, V21/12V,V2V.


In past decades wireless technology is used for the development of vehicular network. The idea was found to communicate roadside infrastructure & is used by wireless network. If you build network operation for routing intended result have developed a vehicular network called vehicular ad hoc network[2].The structure of VANET is hybrid architecture. The focus on VANET application is to improved driver security and function which is based on traffic control. The domain of cloud which has framework to vehicular services. Cloud computing is to create vehicular cloud such as various sensors.

Vehicular Data Cloud Services1Muhammad Zunnurain Hussain , [email protected],

2 2Muhammad Zulkifl Hasan , [email protected], Zaka Ullah , [email protected] Hassan , [email protected]

1 2Bahria University Lahore Campus , Lahore Garrison University

Abstract

The advance cloud computing has provided an opportunity to resolve the challenges which effects by increasing transportation issues. Two methods of cloud services are available these are parking and mining. Mobile cloud computing has improved the storage capacity, stand by time of mobile terminals by migrating data processing to the remote cloud. The introduction of smart phones, cloud computing the automotive system is shifting toward the internet of vehicles.



Hussain et al LGURJCSIT 2018



15

A modular approach is applied to vehicular cloud computing to divide a not understandable systems into small sub system. A modular approach is used to build vehicular Cloud services. We can divide a module cloud services by traffic administration, Information processing and mining[1]. Cloud computing includes three services Such as

Ÿ SAASŸ PAASŸ IAAS

These are used to build cloud services. Cloud services are subdivided into parts such as private, public and hybrid cloud. User information query are public cloud; traffic administrator should be hosted by private cloud. Vehicular cloud and hybrid cloud are developed by taxonomy method.

There are three layers of V-Cloud to combine vehicular cyber security system.

Figure 1: Achievement for Iot based Vehicular data cloud

In present, Vehicular cloud was made to interlinked automobiles network into architecture.

SERVICES OF IOT BASED VEHICULAR DATA CLOUD:

In the data especially information for VANET explained into two things of Vehicular to veh icu la r and V21/12V exchang ing information. Investigate systematically of V2V communication focus on how to achieve timely data delivery through mobile vehicular on roads over signal. The data result framework with interlinked buffering.

Aps to improve the quality of data dissemination services. At network Level network resource were managed to requirement for real time and non-real time traffic. Rate less coding technology also applied to improve the efficiency of data dissemination. The limited buffer size of Roadside wireless app and intermittent connectivity between the wireless apps Affect the data dissemination performance. When no vehicle is available can deliver the Data static nodes locates at road intersection keep data and forward routing path available. The mailing and accepting large amount of data from automobiles to signals. A wireless measurement study under different driving condition carried out Vehicle cloud considered as a cloud of vehicles which make available their underutilized with each other to provide services to authorized users. These resources needs to be analyzed dynamically.

The Intelligent transportation system (ITS) is a basic element for sustaining a smart city and IOT play an important role in intelligent transportation system. The advantage of IT'S is reducing traffic accidents. The growth of mobile applications and IOT the question arises how to deal with big transportation data is still an open problem in IOV. Mobile Support of IOV is not been considered cloud computing frameworks. The IOV is composed of vehicular cloud ad hoc n e t w o r k s p r o v i d i n g s o l u t i o n t h e s e heterogeneous network is not considered in services data. The mobile app services to facilitate the location of vehicular at the same time which is involved central cloud services. Fog computing provides an offload computation from cloud to local fog server. A vehicular fog computing is proposed in which vehicular with resources are used in computational infrastructure and the burden of resource limited vehicles are reduced. Software defined network is composed as the fog computing which is basis of IOV[3].


The introduction of smart phone, mobile devices, internet and cloud computing the automotive community is transfer to wards the IOV. The Auto 3.0 ecosystem show robotics vehicular a device analysis and transfer information with computing system and other vehicles. Vehicles are the source of IOT system. The core networking technology couples with computation on vehicular sensor data are the part of IOT. The sensor data as a part of smart city which is used to be traffic management System and pollution monitoring. The vehicular sensors generate multi-modal data, support different encoding formats, heterogeneous domain and communicate using different


domains. Vehicular network mostly used IP version 6.6 version of internet protocol not support Mobility, not data centric. Automatic management of interlinked vehicles and their physical resources. Data servers is important for physical delivery and application found. Mobile servers are interlinked with vehicles.

The implementation on trust based management of Vehicular social network and its feedback is based on Vehicular cloud computing architecture[4]. A general trust Management targeting vehicular social network. A Vehicular cloud computing has three layer of communication Architecture for vehicular social network These work continue our work to research a long trust management to save data on vehicular social network and result a feed back plan for a management framework. The Small cell technology is a solution for an advanced growth of signal data services and requirements of signal reached[5]. Small cell base station are low power and low range. On the other hand, long cell base station are high power and high range. The 5G mobile network are the fastest network. The net work of 5G is high speed services [6].The data traveled is fastest then other network. The mobile cloud the data is stored in large amount and data traveled in vehicles is fast then other network[7]. A lot of small base station such as fem to cells are installed in spot hot areas to improve coverage and small cell users [8]. There are many multi users to worked with clouds. Mobile users are play a vital role in vehicular cloud services [9].

3. METHODOLOGY

The methodology used Vehicular ad-hoc networks (VANET). The vehicular ad-hoc network were arranged to backing the connection mid various automobiles and the connection between automobiles and road side infrastructure. The vehicular ad hoc network basically a hybrid structure. It accommodates VANETS for intelligence transportation System. The VANETs application were developed for vehicular manufacture, government Agencies and industr ia l organization. The main points on Vehicular ad-hoc network is driver safety improvements and offered traffic monitoring, updates emergency warning and road assistance. The non-safety VANET application such as gaming had been developed. The Research of info dissemination

17

depends on its scalability to handle changing number of vehicles. Vehicular clouds handle a vehicular traffic, traffic hurdles caused by special situations. As new things reached made competent internet of things middleware that support integration of new things. As vehicular data cloud is often the vehicular networking is often reliable. More new Mechanism are needed to communication reliability with reduced traffic overhead. Some concerns of security and privacy in vehicular data clouds due to lack of Established infrastructure of authentication and authorities. A low security issues of Vehicular data cloud is unacceptable for vehicular services regarding transportation Safety. Global standard is compulsory to avoid conflicts between locally developed Vehicular data cloud. There are number of stakeholders also exist which have Challenges to established global standard to lower complexity and make vehicular Data cloud cost effective.

References[1] G. Y. a. L. D. X. Wu He, Developing Vehicular Data Cloud Services in the IoT Environment, vol. 10, no. 2, p. 9, 2014. [2] H. L. RyangsooKim, Prefetching-Based Data Dissemination in Vehicular Cloud Systems, vol. 65, no. 1, p. 15, 2016. [3] J. h. Soumya Kanti Datta, VEHICLES AS CONNECTED RESOURCES, p. 10, 2017. [4] A Cloud-Based Trust Management Framework for Vehicular Social Networks, vol. 5, no. 2017, p. 14, 2017.[5] Q. W. Y. X. Ducheng Wu, QoE and Energy Aware Resource Allocation in Small Cell Networks With Power Selection, Load Management, and Channel Allocation, vol. 66, no. 8.[6] J. J. R.-M. P. A. Jonathan Prados-Garzon, Modeling and Dimensioning of a Virtualized MME for 5G Mobile Networks, vol. 66, no. 5, p. 13, 2017. [7] L. W. XIAO CHEN, A Cloud-Based Trust Management Framework for Vehicular Social Networks, vol. 5, p. 14, 2017. [8] Q. W. Y. X. Ducheng Wu, QoE and Energy Aware Resource Allocation in Small Cell Networks With Power Selection, Load Management, and Channel Allocation, vol. 66, no. 8. [9] Y. L. a. H.-H. C. Xing Liu, Wireless Resource Scheduling Based on Backoff for Multiuser Multiservice Mobile Cloud Computing, vol. 65, no. 11, p. 13, 2016.


in VANETS is explained into two Categories vehicular to vehicular, vehicular to infrastructure and infrastructure to Vehicular. The focus on research of vehicular to vehicular how to achieve a data services by using signal. The focus on research of infrastructure to vehicular are the limited buffer size of road signals and Inter linked between wireless app and mobile vehicles. The data pouring algorithm proposed with intersection buffering. The vehicles at intersection sent the data by the source node in their Buffers and again rebroadcast it to other vehicles passing the interaction. The route information of vehicles which is available through GPS enable Navigation system in the vehicles apps to improve the quality of data dissemination services. At network Level network resource were managed to satisfy the quality of services Requirement for real time and non-real time traffic. At pocket level High transmission rate was proposed. Rate less coding technology also applied to improve the efficiency of data dissemination. The limited area of edge of way signal app and intermittent connectivity between the wireless apps Affect the data dissemination performance. A static node was proposed by Hybrid data dissemination. When no vehicle is available can deliver the Data static nodes locates at road intersection keep data and forward routing path available. The sending and receiving large amount of data from Vehicle to roadside app is wireless transmission. A wireless measurement Study under different driving condition carried out. The info-circulation method for vehicular cloud services has same with web caching action accept for provincial approach to internet capacity via proxy servers build upon on network topology the web caching action.

4. CONCLUSION

The framework was presented using vehicular data services which uses internet of things. The data loss is automatically stored in the cloud. During data transfer vehicular data cloud is used. It plays an important role in transferring data.

5. FUTURE WORK

Internet of things based vehicular cloud computing must be competent way, secured previous used at extensive. The vehicular cloud

18

1. INTRODUCTION

The network is a set of communication devices connected by media links. There are two types of network wired network and wireless network [1]. Bob Metcalfe and D.R. Boggs are the two engineers who developed the Ethernet (Wired network) [2].They started their work in 1972 and established their development in 1980 under the standards IEEE (802.3).Wired network defines as a low-level transfer of data and for its usage, they build the cards and cables, through which data can be transfer from one PC to another computer. The term wired refer some solid thing which consists of the cable. Wired network means that connection through cables, modem or any other source. The data transfer from one place to another through the cables. The cables are consists of copper, fiber optic, and twisted pair. In a wired network internet connection is taken up from only one source (single wire), modem or any other kind of means. The wired network also includes Ethernet in it. Ethernet wire has been using for a long time period. The wired network usually transfers the data up to 10 Mbps. Fast Ethernet and Giga Ethernet also used in the wired network. The speed of fast Ethernet is 100 Mbps and the speed of Giga Ethernet is 1000 Mbps and

CAT 5, CAT 6 both wires used for fast and gigabit Ethernet [3]. The wireless network was also established by IEEE in 1997 with standard 802 [4], its first connection was of 2 Mb and that time it was not so much advance and familiar to anyone but later with the generating of new versions of wireless, it become famous over the world. Wireless network refers to a medium such as electromagnetic waves or infrared waves through which data passes. All the wireless network devices have antennas and sensor in them [5]. The wireless network is based upon on frequency without using any kind of wire .it is an open source for every person. In a wired network, the region is bound for the user to use the internet and to communicate with another computer. But because of the wireless network region is not bound for the user and connectivity to become easy. Wired network is inexpensive and has high reliability and high bandwidth with high speed. Whereas wireless network is expensive, have lower quality and lower bandwidth. Wireless network infrastructure requires little more than the single access point. On the other hand, the wired network has more difficulty and complexity with the cable connections. Using wireless network many people can access the internet but in wired

A Survey on Wired and Wireless NetworkRafaqat Alam Khan, Muhammad Arslan Tariq

Department of Computer Science Lahore Garrison University (LGU), Pakistan

Abstract

The wireless industry is going very fast nowadays. We can easily see the evolution from 2G to 3G and now advance to the 4G and 5G network. Before wireless networks, wired networks were commonly used in every field. But there were some disadvantages regarding mobility, quality of service and connectivity. Wired network bounded the region of the working area for the internet and it requires multiple wires to connect computer from one device to another. While on the other hand wireless network is an open source for everyone to use the internet. There is no limitation of the region and no issue regarding connectivity because data is transfer through signal which includes frequency in the form of waves. But there are also some disadvantages of wireless network regarding cost, speed, coverage, bandwidth etc. If we talk about the better network so it depends on the situation and problem.

Keywords: Wired, Wireless, Network, Security, Internet



Khan et al LGURJCSIT 2018



19

different policies due to which the data is secure and cannot be easily hacked by the other [8].

2.3. Type of Connection

There are two main and important connections [9] of the network as follow:Ÿ Point to pointŸ Multipoint

2.3.1.Point to Point Connection

A point to point connection provides the path between two devices. The transmission of the data is reserved between those two points. It uses the length of the cable.

2.3.2.Multipoint Connection

A multipoint connection is used to connect two or more devices by a single link. The channel shared the capacity with each other in the environment.

2.4. Physical Topology of Network

The word physical topology refers to the way in which network is carried out physically, when two or more devices connect with each other with the help of some certain path then it forms a topology [10]. It is basically a geometric representation between the devices.

2.4.1.Types of Topology

Ÿ MeshŸ StarŸ BusŸ Ring

2.5. Network Models There are two main models in networks:Ÿ Open system interconnect(OSI)Ÿ Internet model

These models are depending upon the standard. Standards are needed to connect the h e t e r o g e n e o u s n e t w o r k w i t h o t h e r computer/devices.

2.6. Categories of Networks

Ÿ LAN (higher bandwidth and limited in size up to few kilometers)Ÿ WAN (lower bandwidth, have long distance

network additional user need the additional wiring. There are some problems in the wired network which we are going to discuss in detail in a wireless network which works as a solution.


A network is a device which is used to connect by a communication link. The device through which it is connected is also called Node which can be a printer, scanner, and computer. It may be any kind of devices which have the ability to send or receive the data generated by the other devices on the network [6]. All the network work upon the protocols (which are the set of rules made for the network). all the networks must have the following:

Ÿ ResourceŸ Transmission medium

2.1. Reasons for Using Network

There are the following main reasons for using the network:

Ÿ Provide serviceŸ Reduce equipment costŸ Sharing the files from one to another mediumŸ Sharing the printers and other devices Ÿ Manage the security of the resourcesŸ · Support the network application

2.2. Performance of the Network

The most important performances of the network are:

Ÿ ReliabilityŸ Security

2.2.1.Reliability

The network is reliable [7] because if the data is lost due to the failure then the network has the capacity to recover the data. It provides the accuracy and failure of the network can be measured by the frequency.

2.2.2.Security

It provides the facility of recovering of the lost data. It secures the data from damage and protects it from external resources. It implies the


and multiple of different LAN)Ÿ MAN(size between LAN and WAN and accessible inside the town)Ÿ Now let discuss in detail about the network types on is wired network and the other is a wireless network.

3. ORIGIN OF WIRED NETWORK (ETHERNET)

Robert Metcalfe develops the first Ethernet (wired) system at Xerox PARC [11]. It connected more than 100 workstations by using the 1 km cable length at the speed of 2.94Mbits/sec. In 1983 the IEEE 802.3 gave it the specification. It is used for the long distance and the fiber optical cable is used for this purpose for two decades. Now a day's Ethernet gave us the reliability, higher speed, higher bandwidth. But it is the beginning because Ethernet is becoming faster and faster up to 400Gbit to 1Tbit.in future it will further simplify by automated service and delivery etc. The speed of the Ethernet (wired) depends upon the cable we used.

4. WIRED NETWORK

The word wired [12] refers to any kind of physical medium which is consisting of the cable. The cables made up of the copper, fiber optics, and twisted pair. Wired network is mostly used to carry different types of signals in the form of electricity from one medium to another. In a wired network, only one internet connection is used in the cable. Only one device is attached to one internet cable and data is shared among the different devices by using this same concept of wire network.

Figure 4.1: Diagram of Wired Network

4.1 Protocols of Wired NetworkWired network has the following protocols and let's discuss this protocols one by one:

Figure 4.2. Protocols of Wired Network

4.1.1. Ethernet

It is one of the protocols of a wired network and it is mostly used in the world. Ethernet uses the access method which is called CSMA/CD (carrier sense multiple access/collision detection). This is basically a system in which each computer before sending some information/data through the network must listen to the cable. In this only one by one link is made. If there is multiple networks in a single line then computer have to wait until the line is clear after that it sends the data to the receiver. If two devices send the data through the cable, this work is done by one by one. First, one device receives the data from the sender and the other device have to wait until the first data transfer completed and wait for his turn and try again later. When the first receiver receives the data the second will start. Because of this sometimes collision occurs and the computer has to wait and lots of time consumes. But this collision is occurring for a very small time and it does not affect the transmission of the network. The Ethernet protocol uses the method of bus, star and tree topologies [13]. Data can be transferred through the coaxial wire, optical fiber wire at a speed of 10Mbps up to 1000Mbps.

4.1.2. Fast Ethernet

It is another type of protocol which is used to increase the speed of transmission and it develops the new standards that have the speed of 100Mbps [14].this is called fast Ethernet. Fast Ethernet requires the hub, networks interface cards and category 5(CAT5), fiber optics and twisted pair cable is required. This Ethernet is commonly used in school nowadays.


Ÿ EthernetŸ Fast EthernetŸ Local talkŸ Token ringŸ FDDI

21

4.1.3. Local talk

Local talk is a wired network protocol which was developing by the Apple computer for Macintosh computer [15]. The data is transfer by using the special twisted pair cable. It allows the linear bus, star or tree topologies using twisted pair cable. Disadvantages of the local talk are the low speed of transmission which is 230Kbps.

4.1.4. Token ring

The token ring [16] protocol was developing by IBM in 1980 mid. The information is carried out in the form of token and move round in the circle/ring. Two computer cannot connect to each other until the will finish the task. After that, it connects with the other system and its application on 2,23 computers system and rarely used because its performance is very low and lots of time is consumed and lots of trouble is there .its speed is 4Mbps 0r 16Mbps.

4.1.5. FDDI

Fiber distributed data interface (FDDI) [17] is another protocol of a wired network that is used to connect two or more LAN. This is far away over a long distance. It is used as the token ring method. But it uses the two rings for the transfer of the data. If one system is busy then the other systems are automatically activated and transmit the data. Its advantages are high speed and it works on the fiber optic cable at 100Mbps.

4.2. Advantages of the wired network

Ÿ Power is not consumed so much.Ÿ Wired network is easy to use, plug in the wire and ready to use the internet.Ÿ There are varieties of cables available in the market nowadays. The user can choose according to its need and budget.Ÿ It provides the constant, stable and faster speed because it provides one to one connection.Ÿ Security is very strong in a wired network.Ÿ It is beneficial for small area network like at home or at the office.Ÿ It is most reliable than a wireless network.Ÿ It is good for business, home, office, school etc.Ÿ Bandwidth is higher.Ÿ It can extend to the longer distance by using the optical fiber.

4.3. Disadvantages of the wired network

Ÿ It is not useful for the mobile or Smart-phone. Because it requires some physical sort of connection to use the internet.Ÿ One wired network cable is attached to only one computer so it does not facilitate another computer by single wire.Ÿ Cable can be easily damaged after some time, so the user has to be very careful while arranging and connect the cable with PC and protect the cable from any cuts and water.Ÿ It messes the room where you are gathering the wireŸ If you want to connect more devices with each other you need an Ethernet connection. But it also requires more wires to connect and if you are doing this then it is difficult for you to find out that which wire is connected to which PC and it takes your lots of time.Ÿ If you want to expand your network then you need more cables and it becomes more costly and takes lots of time to establish the network. If you want to expand your network widely then you have to rewire all the devices and then again establish the wider network.Ÿ There is no freedom of movement for users.Ÿ Wired network is not suitable for openly public usage.Ÿ Lots of cables need to connect to the certain port.Ÿ Set up is difficult sometimes but it's expensive.

5. WIRELESS NETWORK

The wireless network was also established by IEEE in 1947 with a standard 802. It first connection was of 2Mb and that time it was not so much advance and familiar to anyone but later with the passage of time and generating the new version of wireless, it becomes famous over the world. Wireless word is used to refer to medium which is made up of electromagnetic waves or infrared waves. All the devices, which are wireless that has sensor or antennas embedded in them. It includes mobile, wireless sensor, TV remote, laptop etc. It does not use the wire for the connection between two devices or to transfer the data. It uses the radio frequency waves. Fiber optic and broadband ADSL are also used.


Figure 5: Diagram of the Wireless Network

5.1. Types of wireless protocols:

There are three protocols of wireless network:Ÿ Long range (measured in miles)Ÿ Medium range (measured in tens or hundreds of feet)Ÿ Short range (less than 10 feet)

5.1.1. Long range

Long range protocols are used for speed to transfer the data over the longer distance. It may be used as back-haul between two sites such as Smart-phone etc. GSM (Global System for Mobile Communication) is the most important protocol of wireless network which is using in all world and its connections are between the cellular phones and mobiles.

5.1.1.1. LTE

Before newer Smart-phone, older generation used GPRS, EV-DO for the communication. Because of that companies and industry have to spend lots of money to upgrade it and made it supportive for 4G. LTE (long-term evaluation) is used to improve the low data rates and other issues that occur in older phone . The protocol can carry the 100Mbps of data which is divided into the users to use and each user gave the 10Mbps.

5.1.1.2. 60 GHz protocol

Most of the video which is running on the computer needs 60GHz and they are using it as well. It requires a lot of bandwidth. There are two different standard called wirelesses HD and WiGig. It gives the best high-quality definition for the video streaming.

5.1.2. Medium Range

WLAN is usually used for the medium range protocol which is used for the communication between the computers to enhance wired LAN or replace it. These all protocols are the parts of IEEE (Institute of electrical and electronics engineers) 802.11 standards.

5.1.2.1. Wi-Fi

Wi-Fi is mostly used nowadays because of its range and access to another device. It provides the facility of a hotspot as well. It becomes popular in 1990 for the hardware usages. Wi-Fi can be controlled in the environment according to the range. Its speed is lower as compared to the other wireless network protocols speed but mobile device easily support the Wi-Fi and LTE and give the flexibility to the user.

5.1.2.2. WAP

The wireless application protocol is standardizing a protocol for communication. It is used to provide the security and privacy to the network. There are other types of protocols .one is WEP and the other is WPA. Both used for security.

5.1.3. Short range

Wireless Personal Area network or WPAN is also called short wireless protocols. Which work on the lower frequencies between the devices which are just a few feet away from each other? Bluetooth is an example of a short range protocol. Its common usage is that it allows the wireless headset to communicate with a portable phone. Infrared data association or IRDA is older and used for every short range protocol.

5.1.3.1. Bluetooth

It is the oldest wireless network protocol which is commonly used now a day. It transfers the data from one device to another device. It needs a lower amount of power to work then Wi-Fi and from most other wireless protocols. It is a short distance wireless network protocol. Wi-Fi has been replaced with the Bluetooth [18] but some Smart-phone still has the features of Bluetooth in their system.


Ÿ We can also use the network in business because everyone can easily access each other; no matter where they are now and what are they doing. Because of this, they can run they're smaller to smaller business and larger also and earn lots of profit.Ÿ We can use the wireless network in mobile communication. By using this technology multimedia approach, interconnection and transfer of data and all other things related to wireless are in your control and range.Ÿ We can also use this technology in voice communication. It gave the facility to in contact with two or more users via video calls or text messaging.Ÿ We can also use the technology in the remote control. There are many different uses of a remote control system such as doorbell at home, TV, car, remote, garage opener etc.Ÿ We can use this technology in entertainment and also in the navigation system.

Figure5.3. open source Wi-Fi usage

5.4. Properties of wireless network

Ÿ HomeŸ SpaceŸ PerformanceŸ Wireless network element

5.4.1. Home

Wireless technologies are effective for sharing printer, scanner, and high-speed internet connection. It saves the cost and time and creates the mobility for the devices which are connected to the devices.

5.4.2. Space

It is sometimes become difficult to connect the wires and cables. Wireless technology allows a specific space to the user through which it will be able to communicate with the other device.

5.1.3.2. Wireless Home Automation protocols

It is used to control the remote control of light, home appliances, and gadgets. Two basic protocols for home automation are z-wave and Zigbee. They have low data rates and support the low energy consumption for home automation .0.25Mbps for Zigbee and 0.01Mbps for Z-wave.

5.1.3.3. Ultra Wide Band

It is UWB and also called digital pulse wireless. It used for short distance and have a wide frequency band with lower power. And take that kind of data which was bend due to some obstacle or due to some higher power.

5.2. Factor affecting the performance of the wireless network

Ÿ Physical obstructionŸ The range of the network and distance between the devicesŸ Sharing of signalŸ Usage of network and load on the networkŸ Poor antennasŸ Reflection back of the signalŸ Spectrum channel limitationŸ Restriction of the wireless signalŸ The polarization of the signalŸ Speed loss due to wireless overheadŸ Lower performance

5.3. Uses of wireless network

Ÿ We can easily use the wireless network [19] in medical science without any kind of issue or danger. It is used in the remote monitoring of the patient, biometric data of wireless network and dispensers application.Ÿ We can also use the wireless technology while traveling with the help of an airline. Now we can travel from place to another without any ticket. Because all documentation is available on your mobile and can be accessed by using the hotspot wireless.Ÿ We can also use the wireless network on hotels. [20]It enhances the business of the hotel. The management uses the wireless internet as wireless network checking of hotel .we can check the guest list through mobile or tablet. Opening and closing of the door are also because of the wireless network.


wireless becomes difficult.Ÿ Speed is effective and slower while sending some sort of file.Ÿ If you go far away from the router it becomes difficult for you to access the internet. The range is limited.Ÿ Less secure because anyone can steal your internet bandwidth if your password is not secure and not protected.Ÿ Easily hacked the information.Ÿ It depends upon the wave-like radio.

5.7. How Wi-Fi has changed the world

Before the internet, nobody familiar with each other but after the internet has arrived people started using it and it becomes so popular that it demands increases day by day. And Wi-Fi becomes the life of people more easily because anyone can access the internet through laptop, mobile and can be accessed by the nearest Wi-Fi, hotspot or booster. Now a day 88% people come online according to the research through Wi-Fi.it played a most important and significant role in human life. With the help of Wi-Fi, we can improve our society and as well as ourselves steadily and speedily. Through Wi-Fi two or more cities are now able to connect to each other .it provide the online shopping facility to the people and this type of shopping criteria is very common in Europe because with the help of this technology people compare the prizes and purchase the item. Wi-Fi [14] is also giving f ac i l i t y o f t he communica t ion and communication become easier through this technology. Healthcare center is also available for the people online; people can easily connect to the hospital through this technology because now a day's doctor carries Personal Digital Assistant (PDA), through which they can communicate with the people easily, no matter where are they, at home or outside. Wi-Fi is the best solution for a whole geometric location and they can share their data easily. It going forward and forward and day by day its technologies is increases and its speed also up to 866.7Mb/s because of 802.11ac and 802.11n.

Figure5.7. Wi-Fi advance technology

5.4.3. Wireless Network Element

It is a device used by a carrier to support for the back-end network and it also supports the mobile switching center. Wireless technologies depend upon the network element. It is used in a wireless network in the physical layer.

5.4.4. Performance

It enhances the performance of network from 2G to 3G and now mostly using the 4G network with high speed.

5.5. Advantages of wireless network

Ÿ Users are free to move with a wireless network and can easily access the internet anywhere with their laptop and other handsets devices.Ÿ The user can easily share the files with other devices without any connection of cables.Ÿ There is no need of cable connection. [5] So it is cheap and not a time consumer.Ÿ Easily connected to more than one PC or device at the same time.Ÿ They are convenient and easily accessible.Ÿ It handles a large number of users because it is an open source and unlimited to use.Ÿ By using the wireless network social media information becomes easy to access and become easy to transfer.Ÿ It is convincing because the user can access from any nearly located resources.Ÿ It is useful to enhance the productivity.Ÿ In wireless network number of user connect with each other easily but in wired they all need their wire to connect Ÿ It is cheap.Ÿ Network security is becoming good and stronger than the system cannot be easily hacked because they insert the strong password in hardware and in software.Ÿ Although it is slow in speed it fulfills the requirement of the user and the user easily gets the desired thing from the internet.Ÿ Healthy and safe.Ÿ Wi-Fi is cost effective.

5.6. Disadvantages of wireless networkŸ It can require extra cost and other equipment to set up.Ÿ The person who is not so much familiar with a computer for that type person setting up the


working to increase the working capacity of the wireless technology more and more enhanced with better security and high speed.

6. CONCLUSION

The wireless network is better than wired network. 80% of the world using the wireless network now a day. Its future is brighter than wired network according to research. Wireless gives the freedom of movement and sharing of files becomes easier, no matter of slower speed. But there have been made some changes in the wireless network properties related to the speed, cost, and security. It gives the flexibility as data is transfer from one medium to another through radio waves. But in the wired network, there is a concept of cables, which sometimes mess the working place and become dangerous also. Cables can be easily damaged. There is a single connection and no multiple connections can be made or accessible on a single cable network. It is time-consuming and costly as compared to the wireless network. The wireless network is opposite to the wired network and its protocol is much beneficial than that of a wired network. Wired technology does not provide the generations of the internet to the users. It is limited and gives the connection through the wire to wire that's why wireless is commonly use nowadays and further going on we will see the brighter and brighter future of Wi-Fi technology.

References

[1] Tim Zimmerman, Christian Canales, Bill Menezes "Magic Quadrant for the Wired and Wireless" 17 October 2017.

[2] Kirsner, Scott. "The Legend of Bob Metcalfe," Wired, 6.11 (November 1998), 182-186.

[3] T. Kiravuo, M. Sarela, and J. Manner, "A Survey of Ethernet LAN Security," in IEEE Communications Surveys & Tutorials, vol. 15, no. 3, pp. 1477-1491, Third Quarter 2013. doi: 10.1109/SURV.2012.121112.00190.

[4] R. B. Marks "Standards from IEEE 802 Unleash the Wireless Internet" Standards from IEEE 802 Unleash the Wireless Internet March 5, 2001.

5.8. Future of wireless technologies

Wireless technology has changed the mean of communication .business industry are running and highly progressed because of the wireless network. It is now more suitable for the business because of its awesome features like speed, security, mobility, and Wi-Fi hotspot. Voice application can be successfully running because of the wireless network. By using this you can easily access the internet with high speed and including text, audio, video messages, and many more things become easier due to the wireless network.

Figure 5.8 5G network latest technology

Wireless technology is going faster and faster each year and day by day. [9]In Europe large number of people are using 4G internet and it is not so much common yet in some countries but now they looking forward for a new and advance generation of network which is 5G.it is probably introduced into the market in 2020 and it provides the more services to the people and lots of data with extremely high speed of 10 Gigabit per second and best quality of data with response time below one millisecond which is most beneficial for the internet things. After 5 to 10 years, billions of billions of new devices will use the facility of the 5G [15] including car, machine to machine access, telemedicine, household medicine, no matter of bandwidth because it provides the bandwidth 24/7 to people. And we will use the biohazard sensor which is used to carry the bit to bit data each and every day. This is happening because of the rapid increase in the technology and daily growing of the traffic data. The other most important technology is Li-fi (light Fidelity), which is used to connect the things and data with the help of a light signal. It is faster than Wi-Fi and provides the speed of 224 gigabits per second. It uses the ultraviolet and infrared waves to transfer data and they carry more information than radio frequency waves and it is 10,000 times larger than the radio frequency. Engineers are still


[5] H. Huang, "Flexible Wireless Antenna Sensor: A Review," in IEEE Sensors Journal, vol. 13, no. 10, pp. 3865-3872, Oct. 2013.doi: 10.1109/JSEN.2013.2242464

[6] V. Potdar, A. Sharif and E. Chang, "Wireless Sensor Networks: A Survey," 2009 International Conference on Advanced Information Networking and Applications Workshops, B r a d f o r d , 2 0 0 9 , p p . 6 3 6 - 6 4 1 . d o i : 10.1109/WAINA.2009.192.

[7] I. M. Obeidat and S. Y. Berkovich, "Reliability of network connectivity," 2008 First International Conference on the Applications of Digital Information and Web Technologies (ICADIWT), Ostrava,2008,pp.435441. doi:10.1109/ICADIWT.2008.4664388.

[8] F. Yan, Y. Jian-Wen and C. Lin, "Computer Network Security and Technology Research," 2015 Seventh International Conference on Measuring Technology and Mechatronics Automation, Nanchang, 2015, pp. 293-296.doi: 10.1109/ICMTMA.2015.77.

[9] M. Hur, H. Lee and M. Kim, "A study identifying the connection type of an end-host to the network using Round-Trip-Time," 2011 13th Asia-Pacific Network Operations and Management Symposium, Taipei, 2011, pp. 1-4.doi: 10.1109/APNOMS.2011.6076974.

[10] J. Zhang and L. Zhou, "Research and Design on Network Topology Management System of EJB Clustering," 2010 International Conference on Computational Intelligence and Software Engineering, Wuhan, 2010, pp. 1-4.doi: 10.1109/CISE.2010.5677092.

[11] Robert Metcalfe 'The Year of The LAN' is a long-standing joke, and I freely admit to being the comedian that first declared it in 1982...", InfoWorld Dec 27, 1993.

[12] A. Patel, S. Ghaghda and P. Nagecha, "Model for security in wired and wireless network for education," 2014 International Conference on Computing for Sustainable Global Development (INDIACom), New Delhi, 2 0 1 4 , p p . 6 9 9 - 7 0 4 . d o i : 10.1109/IndiaCom.2014.6828051.

[13] A. Gohil, H. Modi, and S. K. Patel, "5G

technology of mobile communication: A

survey," 2013 International Conference on

Intelligent Systems and Signal Processing

(ISSP), Gujarat, 2013, pp. 288-292.doi:

10.1109/ISSP.2013.6526920.

[14] I. Crayford, ""Fast Ethernet" gets plug-and-

play," Proceedings of WESCON'95, San

Francisco, CA, USA, 1995, pp. 354-.doi:

10.1109/WESCON.1995.485302.

[15] Chaitali Chaudhari, Suprabha Hajare,

Yogita Patil, Kinjal Rana, Deepukumari Singh,

ICMTEST 2016 Track, "Overview of

AppleTalk", International Journal on Recent and

Innovation Trends in Computing and

Communication (IJRITCC), ISSN: 2321-8169,

PP: 704 - 707.

[16] D. Pitt, "Standards for the token ring," in

IEEE Network, vol. 1, no. 1, pp. 19-22, Jan.

1987.doi: 10.1109/MNET.1987.6434298.

[17] F. E. Ross, "Fiber distributed data interface-

an overview," [1989] Proceedings. 14th

Conference on Local Computer Networks,

Minneapolis, MN, USA, 1989, pp. 5-8. doi:

10.1109/LCN.1989.65236.

[18] K. V. S. S. S. S. Sairam, N. Gunasekaran and

S. R. Redd, "Bluetooth in wireless

communication," in IEEE Communications

Magazine, vol. 40, no. 6, pp. 90-96, June

2002.doi: 10.1109/MCOM.2002.1007414.

[19] A. Karnik and K. Passerini, "Wireless

network security - a discussion from a business

perspective," Symposium, 2005 Wireless

Telecommunications, Pomona, CA, 2005, pp.

261-267.doi: 10.1109/WTS.2005.1524796

[20] J. Pang, B. Greenstein, M. Kaminsky, D.

McCoy, and S. Seshan, "Wifi-Reports:

Improving Wireless Network Selection with


Collaboration," in IEEE Transactions on Mobile

Computing, vol. 9, no. 12, pp. 1713-1731, Dec.

2010. doi: 10.1109/TMC.2010.151.


1. INTRODUCTION

Skin cancer is the main source of human casualty these days with 9,000 passing's every year. In United States more than 5 million individuals are determined to have skin cancer consistently. Numerous scientists are working here for robotized examination yet so far, nothing near or unified framework is being executed up until now. For better order of Skin sore for cancer discovery, scientists are utilized dermoscopic pictures. Dermoscopic pictures are gotten through imaging system in which the impressions of the skin surface are dispensed. Because of which skin further area are appropriately imagined and improved for better characterization. Then again dermatologist additionally preferred dermocospic pictures when contrasted with the standard photos for indicative exactness and furthermore extensive number of dermoscopic connections which are going to achieve the market for advanced cell clients and that would prompt expand the impact better patient care.


In 2009, G. Capdehourat et al. introduced a machine adapting approach to manage and

arrange melanocytic sores in unsafe form of dermatoscopic pictures. The photo database is made out of 433 generous wounds and 80 injuries. After a photo handling stage that fuses hair departure separating, each photo is thus divided using understood picture division calculations. On that point, each sore is portrayed by a part of that vector which contains shape, shading and association data, and besides neighborhood and overall parameters that undertake to reflect structures utilized as a bit of restorative conclusion. The learning and portrayal organizing is performed to utilizing AdaBoost.M1 with C4.5 decision trees. For the subsequently separated database, then arranged and passed on a false positive rate of 8.75% for an affectability of 95%. A comparative portrayal philosophy related with physically separated pictures by a specialist dermatologist yielded a false positive rate of 4.62% for an affectability of 95%.

These results are confirmed and seems more better than which are discussed in this literature. Execution appraisal is sensitive since each and every revealed outcome were gotten using unmistakable databases. Starting at this moment, improvement of a tremendous database of dermatoscopic pictures that could be

A Review on Skin Cancer Data Using Image ProcessingMuhammad Arslan Tariq, Rafaqat Alam Khan

Department of Computer Science Lahore Garrison University (LGU), Pakistan

Abstract

The wireless industry is going very fast nowadays. We can easily see the evolution from 2G to 3G and now advance to the 4G and 5G network. Before wireless networks, wired networks were commonly used in every field. But there were some disadvantages regarding mobility, quality of service and connectivity. Wired network bounded the region of the working area for the internet and it requires multiple wires to connect computer from one device to another. While on the other hand wireless network is an open source for everyone to use the internet. There is no limitation of the region and no issue regarding connectivity because data is transfer through signal which includes frequency in the form of waves. But there are also some disadvantages of wireless network regarding cost, speed, coverage, bandwidth etc. If we talk about the better network so it depends on the situation and problem.Keywords: Wired, Wireless, Network, Security, Internet



Tariq et al LGURJCSIT 2018



29

2.1. L o c a l i z a t i o n a n d D e t e c t i o n o f Dermoscopic Visual Patterns/ Features

In this part of Skin lesion, on the basis of super pixel tile automated prediction of dermoscipic features would be obtained. The feature data would compromise of original lesion image, Super-pixel mask and super-pixel mapped annotation consists of the following different features.

Ÿ StreaksŸ Negative NetworkŸ Milia-like CystŸ Network

Streaks

Pigment Network

The SLICO algorithms would be used to subdivide the lesion images in to super-pixels to minimize the variability and dimensionality. The training data consists of 2000 lesion images in JPEG format and the 2000 super-pixel masks are in PNG format. The ground truth annotation consisting of 2000 dermoscopic features are in JSON format. The JSON number would be either "0" or "1", where "0" would represent absence of dermoscopic feature while "1" would represent presence of dermoscopic feature. The different metrics that would be used for evaluation purpose are given below.

Ÿ AccuracyŸ SpecificityŸ SensitivityŸ ROC curveŸ Average Precision


used as reference test bed that shows up of being a basic issue

The point of this research is to grow such a better picture robotized determination device of melanoma from dermoscopic pictures. The undertaking is divided in two 3 noteworthy stages and are clarified as below.

2.1. Segmentation of Lesion area

The main objective of this task is, the automatic segmentation of dermoscopic images boundaries would be obtained.

Original Image Segment on Mask

The data consists of original image along with segmented lesion area and the doctor report related to the patient diagnosis would be e x t r a c t e d u s i n g n a t u r a l l a n g u a g e processing.There are 2000 images which are available in JPEG format. These images data would be further provided to classifiers for classification. The different classifiers used for classification are SVM, Deep Learning and Neural Network. The different metrics that would be obtained and on the basis of which our results would support scientific completion are as follows.Ÿ AccuracyŸ Dice CoefficientŸ Jaccard IndexŸ SpecificityŸ SensitivityŸ ROC CurveŸ Average precision

The training would be done on binary mask lesion images having PNG format. In these binary masks lesion images “0” would represent the background area i.e. the area outside the lesion area and “255” would represent the area inside the lesion.

30

3. CLASSIFICATION OF DISEASES

In this part the classification of two different binary classifications would be performed for diagnosis of skin lesion images i.e. The binary classification would be done between Melanoma (1) and Nevus, Seborrheic Keratosis (2, 3), and Melanoma, Nevis (1,2) and Seborrheic Keratosis (3) respectively.

3.1. Melanoma (Malignant Skin Tumor)

Melanoma is the deadliest form or last stage of the skin cancer if it is not diagnosed in his early stages. They appear as a mole on the skin and they make pigments inside skin cells known as melanocytes.

Melanomma

3.2. Nevus (Benign Skin Tumor)

Nevus

3. Seborrheic Keratosis (Benign Skin Tumor)

Seborrheic eratosis

The training data consists of 2000 images out of which melanoma cases are 374, Seborrheic Keratosis are 254 and the remaining 1372 are nevi cases. The data is in CSV format consists of three columns i.e. image id, age approximate and sex. The ground truth data is in CSV format compromise of three columns in which first column contains images id and in second column consists of first binary classification task in which "1" is used for lesion melanoma and "0" is used for lesion Nevi or Seborrheic Keratosis, while in third column of CSV the second binary classification in which "1" is used lesion Seborrheic Keratosis and "0" for melanoma or Nevi.

The binary classification metrics used for evaluation purpose are as given below.

Ÿ Accuracy at 0.5 thresholdŸ Sensitivity at 0.5 thresholdŸ Specificity at 0.5 thresholdŸ Average precisionŸ AUCŸ ROC curve

For all the three tasks discussed above the same dataset would be applied to get the desired results.

4. RESULTS AND DISCUSSION

T h e m e d i c a l r e p o r t i n g s u c h a s dermatologist report consists of specific wording (features) related to the particular disease organized into report related section. For machine learning, it requires to organize unstructured text into structured text information for understanding and learning the machine. In this regard, NLP based methods are applied to extract information from human text report automatically and machine learning approaches are used for association of extracted text information with images automatically. In this paper we explore how to extract the features from human written dermascopic reports using NLP based methods? How to automatically associate dermascopic images with report? How to automatically classify and predict skin cancer based on extracted information.

5. RELATED WORK

Several works have been reported regarding


saliency maps. arXiv preprint arXiv:1312.6034.[7] H.-C. Shin, L. Lu, L. Kim, A. Seff, J. Yao, and R. M. Summers. Interleaved text/image deep mining on a very large-scale radiology database. In CVPR, 2015. [8] H.-C. Shin, K. Roberts, L. Lu, D. Demner-Fushman, J. Yao, R.M. Summers, learning to read chest X-rays: recurrent neural cascade model for automated image annotation, in: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2016.[9] Shin, H. C., Roberts, K., Lu, L., Demner-Fushman, D., Yao, J., & Summers, R. M. (2016). Learning to read chest x-rays: Recurrent neural cascade model for automated image annotation. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (pp. 2497-2506).[10] Mariam A.Sheha,MaiS.Mabrouk," Automatic Detection of Melanoma Skin Cancer",[2012],International Journalof Computer Applications (0975 -8887) March,2012.[11] Alexandros Karargyris, Orestis Karargyris, AlexandrosPantelopoulos, "An Advanced image-Processing Mobile Application for Monitoring Skin Cancer",2012,Conferene peper of IEEE 24th International Conference.[12] CharalamposDoukas, Paris Stagkopoulos, C h r i s T . K i r a n o u d i s , a n d IliasMaglogiannis"Automated Skin Lesion Assessment", 2012,34th Annual International Conference of the IEEE EMBS San Diego, California USA.[13] E. Barati, M.Saraee, A.Mohammadi, N. Adibi and M. R. Ahamadzadeh,"A Survey on Utilization of Data Mining Approaches for D e r m a t o l o g i c a l ( S k i n ) D i s e a s e s Prediction",2011,Multidisciplinary Journals in Science and Technology, Journal of Selected Areas in Health Informatics (JSHI).[14] Suhail M. Odeh,"Automatic Diagnosis of S k i n C a n c e r " , 2 0 1 0 , J o u r n a l o f Communication and Computer 8 (2011) 751-755 at David Publishing.[15] Radu Dobrescu, Mateidobrescu, Stefan Mocnu, Danpopescu,"Medical images c l a s s i f i c a t i o n f o r s k i n c a n c e r " , 2010,WseasTransactionson Biologyand Biomedicine;ISSN-1109-9518[16] Gabriella Fabbrocini, Giovanni Betta, Giuseppe Di Leo, Consolatina Liguori, Alfredo Paolillo,AntonioPietrosanto, Paolo Sommella, Orsola Rescigno, Sara Cacciapuoti, Francesco Pastore,Valerio De Vi t a , I n e s M o r d e n t e a n d F a b i o Ayala,"Epiluminescence Image Processing for Melanocytic Skin Lesion Diagnosis Based on 7-P o i n t C h e c k - L i s t " , 2 0 1 0 , T h e O p e n Dermatology Journal, 2010,Vol 4.

segmentation, localization and classification of medical images. [3] Kisilev, P et al. proposed a semi-automated system that helps radiologists in diagnosing the medical images. They extract features from textual description and mapped them to the images. The work is done based on diagnosis descriptions (features) of radiology images. In another work Kisilev, P et al. [3] constructed a feature-based classifier for medical images. This classifier uses features obtained from image segmentation to classify images into categories. Zhang et al. [5] proposed MdNet for automatic report generation by mapping radiology images and textual diagnosis reports. Shin, H. et al. [9] proposed method that used disease name instead features for labeling the images. In this field of study most of the images in a dataset consists of diseased images. However, all cases are not diseased cases. Therefore, diagnosing the disease in an unbalanced dataset in which most images are normal is more challengeable.

References

[1] Bird, Steven, and Edward Loper. "NLTK: the natural language toolkit." Proceedings of the ACL 2004 on Interactive poster and demonstration sessions. Association for Computational Linguistics, 2004.[2] German Capdehourat, Andres Corez, Anabella Bazzano, and Pablo Mus, Pigmented S k i n L e s i o n s C l a s s i f i c a t i o n U s i n g Dermatoscopic Images, Springer Verlag Berlin Heidelberg, 2009, pp. 537-544[3] Kisilev, P., Walach, E., Hashoul, S. Y., Barkan, E., Ophir, B., & Alpert, S. (2015). Semantic description of medical image findings: structured learning approach. In BMVC (pp. 171-1).[4] Kisilev, P., Walach, E., Barkan, E., Ophir, B., Alpert, S., & Hashoul, S. Y. (2015). From medical image to automatic medical report generation. IBM Journal of Research and Development, 59(2/3), 2-1.[5] Zhang, Z., Xie, Y., Xing, F., McGough, M., & Yang, L. (2017, July). Mdnet: A semantically and visually interpretable medical image diagnosis network. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (pp. 6428-6436).[6] Simonyan, K., Vedaldi, A., & Zisserman, A. (2013). Deep inside convolutional networks: Visualising image classification models and


Cloud Based Intelligent Decision Support System for Disaster Management Using Fuzzy Logic

Areej Fatima1, [email protected] 2Sagheer Abbas , [email protected], Muhammad Asif , [email protected]

1Department of Computer Science Lahore Garrison University (LGU), Pakistan2Department of Computer Science NCBA&E Lahore, Pakistan

Abstract

Field of cloud computing is an emerging field in computer science. Computational intelligence and Decision support systems (DSS) have to gain concern as a computing solution to planned and unplanned problems of organizations in order to progress decision-making tasks in a better way. In today era, Disaster management is a big problem. To overcome this problem, a real time computation is required. Cloud computing is a tool to offer promising support to decision support system in a real time environment. In this paper, a fuzzy based decision support system is proposed to meet all the requirements using fuzzy logic inference system.

Keywords: Disaster management, decision support system, cloud computing.

1. INTRODUCTION

Cloud computing is composing of list of Information Technology services provided to users over network by the third-party vendors. Services of cloud computing are provided on network as "as a service" i.e. Software as a Service (SaaS), Platform as a Service (PaaS), Infrastructure as a Service (IaaS). Cloud computing is a very comprehensive perception of infrastructure convergence. [2] These huge services allow users such as Enterprises to get their application up and running from remote place in more convenient way and allow them to manipulate it in more efficient way and meet the growing demand of business. Concept of cloud computing can be understood easily with the example of Play Store for Android and iTunes for Apple. All the application of such devices like smart phones and tablets, can be stored and managed on one place. [1]

Cloud environment have solved many problems like storage problem, problem of large

memory space at user-end is solved due to the cloud services. Users can store his data over cloud storage in case of data loss or damage device he/she can easily get that data. Computing tendencies like grid computing, "Pay as you go use" distributed computing; virtualization show that cloud computing is a multi-disciplinary research field. Actually, lattice computing has evolved into business-oriented form as cloud computing. Cloud computing has brought a marked change in business and IT infrastructure, everywhere computing power, large data storage and other services are subcontracted to third. Cloud computing is a technology and computing model. It provides timely, on-demand network access to the common pool of configurable computing properties. [5]

In 2012 IBM have provided a global private cloud for French Open Tennis tournament. This cloud service was developed to fulfill demands, like real-time score, statistics as well as video clips of tournament, of increasing number of



Fatima et al LGURJCSIT 2018



33

Ÿ Pitch-perfect for distributed business operations.Ÿ Decrease capital overheadsŸ Reduce operational cost.Ÿ Have the capacity to support new decision system.Ÿ Have access from all over the world.Ÿ Available on multiple devices at a time.Ÿ Has the capacity to expansion and minimization.Ÿ Has the capacity to take backup remotely.

For example, support systems can be considered better. Consider that a user goes to the train. As soon as you are a train board, you can change the conditions of environmental computing. If the tunnel enters the entrance, the user's connectivity may be at the moment; the train leaves the tunnel when the connection restarts. But with a Decision Support System (DSS) user can adjust their relationship with the DSS as soon as he knows the tuning is approaching or to use the system in such a way. It plans to prevent it from being completely o b s t r u c t e d . D o w n l o a d i n g d a t a a n d communicating locally. It may be possible that time and distance of intervention cannot be accurately calculated, but at times it may be possible for a period of time and possession. The UN has described destruction as a society or an error in the society. Disasters include human, material, economic or environmental effects, to cope with their own resources more than the impact of the affected community or society.

Red Cross society and Red Crescent Society described demolition management as resource management and responsibilit ies and responsibilities, such as solutions to all human aspects of emergency, especially reducing the effects of preparation, response and maintenance.

There is no disaster in the country that is causing trouble. There are four main types of disasters.

Natural disasters: The effects of flood, storm, earthquakes and erosion have direct effect on human effects and minor effects that cause decease and trouble over flood, earthquake, fire and tsunami.

Environmental emergency conditions: including technical or industrialized accidents,

fans. These services can be accessed through it. Cloud computing is a dispersed computing platform. It needs a medium through which it may share the resources, these resources comprise the data that play a vital role for (DSS) Decision support systems. [6]

Cloud computation have four major designs which are:Ÿ Private Cloud.Ÿ Hosted Private Cloud.Ÿ Public Cloud.Ÿ Hybrid Cloud.

Cloud application can be accessed by using web browsers, desktop and mobile application. Organizations using mobile applications, big data or with multiple locations can benefit from a cloud at a large scale by deploying cloud services.

Irvine (2012) elucidate a private cloud as technology "hosted and managed on-premises by the client, usually behind a firewall, and access to cloud services is exclusive to that client." The major change for a hosted private cloud is that the technology is "hosted off-premises and managed by a cloud service provider, but access to cloud services is exclusive to one client". A public external cloud is hosted and also managed by a cloud service provider, such that multiple clients have access. [13]

The National Institute of Standards and Technology connote Web, smart phones and tablets such as the I Pad." An excellent design cloud that uses unusual, may be suitable for reliable, secure and commercial applications. Effective use of cloud to provide support can not only reduce winter costs, but also practical costs (Ari, 2002). According to Jackson (2011), "proper implementation can provide adequate savings, IT services and advanced credentials". It is claimed that other benefits are agility, software and device accessibility, as well as freedom of space. In short, benefits include:

1.1 Greater reliability: A DSS is always available where it is required.

More secure: Appropriate usage of shields brings about a framework that is less inclined to physical and digital (cyber) assaults


1.4. Disaster relief

It is a coordinated response by many agencies to decrease the impact of a disaster and its long-term consequences. Respite actions comprise of saving, transfer, diet and water supply, prevention of diseases and disabilities, key facilities such as communications and transportation, housing supplies and extra medical care.

1.5. Disaster Maintenance

Once emergency needs can be met and the initial crisis is over, people are affected and the support of those communities remains weak. Maintenance activities include construction of basic infrastructure, medical care and maintenance. These development activities, such as the creation of human resources for health and future development, should be combined with the development of policies and practices.

Disaster management is associated with sustainable development, especially with dangerous people. Foreign volunteers abroad talked about common misconceptions about disaster management.

Recent disasters have crashed destruction on human population and infrastructure, the cost of maintenance is much higher after such a disaster. Storm Ocean and Tsunami in the Ocean are examples of such disasters. Such disasters affect life, wound and spread of waste in large areas. In some cases, it was possible to see the events that occurred in real time. Satellite photos taken from these sites are used to use digital maps to monitor harmful reviews and monitor the guidance of efforts to restore digital land applications. However, research needs to be made to improve many key issues in this regard and these tools and tools of tools and geospatial have certainly produced some results, but they do not agree with their expectations. The media plays an important role in mobilizing relief efforts by informing people related to the affected area.

Fuzzy Logic provides resources through the audible to test the environment in the cloud. The fuzzy logic parameter eliminated by the user, which was introduced for RAM, disk storage, and system traffic. Cloud performance analysis

which may include the production, practice or transport of dangerous materials, and in the event where these materials are manufactured, cast-off or copied and the jungle fire triggered by humans.

Complex disorders: including conflicts and war conditions, authority on strategic facilities, looting and eliminating attacks.

Pandemic emergencies: A sudden appearance of a deadly disease that affects health, services and business intervention, they have a large economic and social value.

Any adversity can impose on basic managements, for example, community protection, supremacy, water, manure/junk removal, transportation and interchanges. The intrusion can seriously influence the well being, social and financial network of local groups and nations. Disasters have a major and long-lasting impact on people long after the immediate effect has been mitigated.

1.2. Disaster prevention

These activities are planned and offer ever lasting security against the adversities. Some natural calamities can be clogged, but the threat to life and wood defeat can be reduced with decent transition planning, eco friendly formation and scheme values, but not all the disasters. In Jan 2005, the Administrations provided a 10-year worldwide plan to 168 Hugo Framework to eliminate the natural hazard's risk reduction. It offers priorities for recipe rules, actions and practical resources to get the flexibility for the disaster for a dangerous community.

1.3. Disaster preparation

This type of activities is designed to design that minimize the damage by eliminating people and loss of life and property and timely and effective rescue, relief and healing from the threatened place. An important way of preparing to reduce the effects of disasters. For the management process, the preparation and managemen t o f peop le ' s g roups in physiotherapy should be high priority.


a need for exchange of information during the emergency; however, it may be very diverse and complicated "I study. It is believed that there are some communities that do not have resources, personal and skill, to help manage activities to develop a specific need to help them from this disaster.

Due to the decent and extraordinary nature of destructive situations, various emergency approaches are not enough. The right information and intellectual resources are accurate in accurate time. Many ICT-based mutual-mutual-co-ordination systems have been developed to share e-mail [12, 13 & 14]. Such measures are instantly administering any destructive situation but cannot meet the demands. Someone may have reason to take real-time situations for any project or alternative decision, as well as concrete and compact real-time information can be obtained [5 & 6]. Pinokova and Atal presented a view of integrated data storage using the operational control of the technical sector and environmental items [7]. Lee et al. reviewed the system of information management in mass destruction for the public. It has suggested that the proposed techniques are not applicable to a very unstable complex disaster environment, which requires immediate answers to a large number of large organizations [11]. Yang Lu explained how to participate in the exchange of information in the social winter virtual community. It analyzes online social networks and proposes social technical design rules to meet communication challenges under uncertain emergency.

3. METHODOLOGY

Here we have proposed a cloud-based model that you set up intelligent decision-makers system to manage destruction by using fuzzy logic based on fuzzy logic decisions (between 0 and 1), no decision between 0 and 1 Should be done. Emergency or destructive hours. All actions or ingredients are connected to the cloud

ST(DSS). 1 components are emergency professionals and local assistant agents on the site. The second internal administration and the government's decision-making. Third resource information all that depends on the cloud-based (DSS).

is calculated using logical and fuzzy rules. It explains how different values of the given parameters can develop cloud performance or drop-down.

In many cases, people living far away from the affected area have been better informed by media compared to those who manage and manage relief efforts. The affected area deprives the power, Internet connectivity and computer skills that prevent information flow. A recent report of the US National Research Council. One Yu (NCRC) 2007 stabilized the issues, based on broad discussion with respondents and emergency managers, improving the situation and making a series of recommendations for necessary research.


After taking into the support framework for the consumers, new technical tips should be developed. In recent days, cloud computing has become a paragraph which can be used for emergency professionals [2]. The main objective of improving this technology is to increase the capacity of individual computer technology, which is useful for urgent emergency support. Support system for decision making can be extended to train and train emergency professionals, health officials and workers, natural disaster workers and other relevant professionals [1, 8, 9 & 10].

United Nations report about the end of the disaster resulting in the destruction of Japan in Japan, which was claimed that 85,541 people per year affected 230 million people per year. For the 2005-2015 action, the Virtual Study of The Hague Framework is an annual climate related to climate, about 50% more than 50% over 2015, about 2015 Is affected with the passing of government and non-government levels, demand for working on joint-day disasters is increasing to meet together [3 & 4]. Accidents related to seasonal changes require immediate response of officials. Public and public organizations need joint efforts to form health teams, civil security, fire and saving services, basic health amenities to protect the human community and their structure. A real-time analysis of the situation requires to make and operate effective decisions.

Judgement making is not compulsory to pre-post and post emergency. Wherever there is


3.1. Proposed model

The proposed decision system is based on system cloud technology. The proposed decision support system (DSS) is an instrument that has managed to attract the attention of many current research studies, which applies by the various artificial techniques aimed at goals. DSS is based on cloud computing using Fuzzy Logic. There are three dimensions to form and distribute the decision-making system. Regarding the cloud-based (DSS) IP and LSA emergency situation, for example, identify the need for cooperation (BOM). The right-side box defines the involvement of admission in various operations, for example, and identifies local audience or death if they access the basic system of instant messages or warning signals. Official manager's box is right to manage the management. This means dimension resources to establish and make complicated decisions for planning plans.

Finally, emergency professionals on the left are important agents who need access to all the information supporting various information. This decision-making service is necessary for relevant stock holder groups, because they are directly related to action-making activities. The internal management of the information can be accessible with emergency updates for emergency professionals, with the latest updates for the latest information and process decision.

The main use of fuzzy suggestion that is for problem solving, provide us relationship for measurement and logics. Fuzzy suggestion has ability to hold and in perfect inputs. This module has three major components.a) Inference Engine: By the inference Engine that technically handles the unclearness of the program.B) Membership functions: Using this function of fuzzy deals that belongs to the analogous it extends fuzzy elements.b) Rule Base: It is a set of all rules which can be made on the basis of interface model theory also define this that depends upon the structure of the rules "if parameter of cloud is found then at what place does it exist”The inference process typically involves these five main steps:Fuzzification: This step real fuzzy scalar value

converts into fuzzy value. This is completed with different membership functions.Applying fuzzy operations: By using AND, OR operator it tries to achieve different relations.Implication: firstly, define the implication operator then obtain the set using fuzzy set by implementing all rules.Aggregation: By using aggregation operator give output of fuzzy sets and all typical rules.Defuzzification: Using defuzzification algorithm convert all aggregate fuzzy sets into cloud value to rank.

Figure 1: Proposed Fuzzy Based System

Proposed Fuzzy based Disaster Management System:

There are three different parameters in the Decision Support System for a cloud service operator management through different resources. There are three other DSL parameters and loca l ass is tant agents , in ternal administration and government decision makers and information resources based on the local emergency cloud-based local site. The following tables show the range of functions and areas in three stages: Low, Medium and High.


Table 1. EP and LSA

Table 2. IM and GDM

Table 3. Information resources (BOM)

Table 4. Fuzzifier output and linguistic values

Table 5. Input and Output Variables Membership Functions used in proposed model


Figure3 represents the comparison of two input variables Emergency Professional Onsite & Local Supporting Agent and Internal Management & Decision Support Makers if the Probability value of Emergency Professional Onsite & Local Supporting Agent up to 45% and Internal Management & Decision Support Makers is 1% then the Disaster Management is 0%. By gradually increasing Internal Management & Decision Support Makers up to 50% and Emergency Professional Onsite & Local Supporting Agents 45% then then the Disaster Management is also increasing p to 50%.

After increasing both variable probability from 50% to 100% the probability of Disaster Management is gradually increasing from 50% to 90%. Figure 4 shows the comparison between Emergency Professional Onsite & Local Supporting Agent and Information Resources by increasing both variable probability the Disaster Management is also increasing.

Figure:3. Rule Surface for IM&GDM and EP&LSA


Fiure5a shows that if the Internal Management & Decision Support Makers is 0.238, Emergency Professional Onsite & Local Supporting Agent is 0.47 and Information Resources is 0.27 then the Disaster Management is 0.25 which is Low.

Fiure5b shows that if the Internal Management & Decision Support Makers is 0.238, Emergency Professional Onsite & Local Supporting Agent is 0.7 and Information Resources is 0.49 then the Disaster Management is 0.49 which is Medium.

Fiure5c shows that if the Internal Management & Decision Support Makers is 0.49, Emergency Professional Onsite & Local Supporting Agent is 0.75 and Information Resources is 0.77 then the Disaster Management is 0.74 which is high.



European Union's disaster response capacity',. p. 14.[9] Hyogo Framework for Action Building the Resilience of Nations and Communities to Disasters', World Conference on Disaster Reduction,. (n.d.). 2005-2015.[10] Lee, J., Bharosa, N. Yang, J. Janssen, M. & Rao, H.R.. Group value and intention to use - A study of multi-agency disaster management information systems for public safety, Decision Support Systems 50, 404-414. (2011).[11] Otten, J., Heijningen, B & Lafortune, J.F. ( 3-4 May 2004,). An ICT implementation to canal i se informat ion! ' , In ternat ional Community on Information Systems for Crisis R e s p o n s e ( I S C R A M 2 0 0 4 ) Conference,.Brussel: The virtual crisis management centre.[12] Howard, R., Kiviniemi, A. & Samuelson, O. The latest developments in communications and e-commerce IT barometer in 3 Nordic Countries', CIB w87 Conference, , Aarhus School of Architecture. International Council for Research and Inno. ( 2002, June 12-14).[13] R. H. Sakr, F. Omara, O. Nomir "An Optimized Technique for Secure Data Over Cloud OS" International Journal of Emerging Trends & Technology in Computer Science (IJETTCS) Volume 3, ISSN 2278-6856 . ( 2014 , May-June 3,).[14] Sakr, R., Omara, F., & Nomir , O. (May-June 2014 ). An Optimized Technique for Secure Data Over Cloud OS" . International Journal of Emerging Trends & Technology in Computer Science (IJETTCS) Volume 3, Issue 3.

4. CONCLUSION

In this paper, an intelligent decision system has been proposed to overcome the disaster management in a real time using cloud computing platform. This system has the capacity for a real time computation using fuzzy inference system. This system utilizes three parameters: emergency professional onsite and local supporting agent, internal management and decision support makers and Information resources at a real time.

References[1] Carle, B., Vermeersch, F. & Palma, C.R. (2004) Systems improving communication in case of a nuclear emergency, International Community on Information Systems for Crisis Response Management (ISCRAM2004) Conference.Brussels, Belgium. p. 15 (2004, May 3-4).[2] Abha Sachdev, Mohi t Bhansa l i "Enhancing Cloud Computing Security using AES Algorithm" International Journal of Computer Applications (0975 - 8887) Volume 67- No.9,. (2011, April ). [3] Asimakopoulou, E., Bessis, N., Varaganti, R. & Norrington, P.kopoulou, E. and Bessis, N. (Eds.): Advanced ICTs for Disaster Management and Threat Dete. A personalized forest fire evacuation data grid push service - the FFED-GPS approach.[4] Bessis, N., Asimakopoulou, E., & Xhafa, FInternational Journal of Space-Based and Situational Computing, 1(1), 76-85. (2011). A next generation emerging technologies roadmap for enabling collective computational intelligence in disaster management,. p. 11[5] Bui, T. &. (1999). An agent-based framework for building decision support systems', Decision Support Systems, The International Journal , Elsevier Science BV., Holland. p. 25.[6] Graves, R.J. (2004) Key technologies for emergency response', International Community on Information Systems for Crisis Response (ICSCRAM2004) Conference2004, Brussels, Belgium. (2004, May 3-4). p. 19.[7] Hernandez, J.Z. & Serrano, J. M. Knowledge-based models for emergency management systems, Expert Systems with Applications, 20, 173-186. (2001).[8]http://ec.europa.eu/governance/impact/planned_ia/docs/28_echo_eu_disaster_response_capacity_en.pdf. ( (2010). Reinforcing the

42

Statistical Power Profiling of Various Network Switches

1 1Waqas Ahmad , [email protected], Awais Salman Qazi , [email protected] Department of Computer Science& IT, Lahore Garrison University, Lahore, 54000, Pakistan

Abstract In the present century there is an expansion in the use of telecommunication network services, because many organizations and business firms do not compromise on networking services, therefore this efficiency of networking services can be achieved only if there are reliable and proficient networking components. To provide exemplary networking services it is necessary to know what are the parameters for networking components. A standout amongst the most vital parameters is the vitality utilization or energy utilization of networking components. Subsequently, there is a need for power profiling of the network switches. Power usage of different kinds of switches, under various load scenarios are tested and recorded. The tool that we have used in this research paper to measure the power usage factor of various switches is emonpi.

Keywords-Power Profiling, Switches, Network Components, Power Usage

1. INTRODUCTION Recent ly, ne twork services in a telecommunication sector have great importance. Companies, organizations, business firms, institutions whether small or large do not prefer slow and outdated network service. Everyone is looking for a nonstop and exceptional network service. Due to this increasing demand of nonstop network service, the need of huge access/data rates came into being. Nowadays fastest data rates are in use. With these faster data rates, the internet speed and performance is rapidly improving day by day. High data rate is directly proportional to the network service. Data rate is the ability to transfer bits per second from one place to another. Therefore, high data rate helps in making network performance better. Network performance can also be maximized if we use more number of switches in a network. Switch is also a network component. Every network component in a communication network has its own duties. The responsibility of switch is to pass the traffic from one node to another in a telecommunication network. Every device or component has its specific power consumption, because it works with the input of electricity.

Some components consume more power as compared to the others. Network companies must be aware of the power usage factor of various networking components. Because if they ignore and do not give attention to the power usage of components, then they have to indulge themselves in submitting huge electricity bills of their organizations. So it is necessary to have a proper energy management of networking components. As there are many developments in the field of power sector, it principally brings growth in the use of computational resources as well. The more we will use these resources, power usage will be increased. Many organizations prefer to make large data centers in which maximum data can be stored. Organizations like Google, Microsoft, IBM, Siemens, Nokia, Samsung, HTC, LG, Huawei etc. have big data centers with large number of networking componentsintegrated with various servers, that makes possible of storing tremendous amount of data. These data centers are consuming huge power everyday just for the sake of their customer needs and to process the information on daily basis. Data center is a heart of an IT industry. Power usage must be taken into account and its profiling must



Ahmad et al LGURJCSIT 2018



43

power consumption of switches can help in maintaining the efficiency of communication network. Authors did not rely only on a single network switch, in fact they have studied the power consumption of various switches of different brands, so as to achieve a reliable understanding. They found out that if a network switch has a higher processing ability, it will consume more energy as compared to a switch that is a bit weak in processing. Nowadays Cisco switches are considered more safe and reliable and with more powerful processing capacity. In the research paper number [5], authors have put their effort on a sensor network to determine the energy consumption. An external traffic load has been inserted on to this sensor network and power consumption of network elements has been measured by using the tool named EPUMS along with the bit rate. They have done research to see the power consumption of routers and found that the router operating at maximum clock rate consumes more power than the others. In the reference number [8], power management of various network devices in local area network are considered for research. During the normal communication activity, network devices for power management cannot be examined, therefore during the low traffic the components that are being attached with the switches are turned off for a while and then examined. Authors obtained reasonable result and saved reasonable amount of energy after performing this activity. This clearly shows that turning off network components may save power consumption of any network. In [9] reference, authors did their research on network devices like hubs, switches (edge & core), routers and wireless access points and calculated the energy consumption in two different scenarios i.e. standalone and data center. The different thing in this research was a proportionality index proposed by authors that can be used to measure the power consumption of different network components. Similarly, in reference [10], authors used power estimator tool named BITWATTS that is used to measure the energy consumption at virtual machine level.

3. METHODOLOGY USED

We have to calculate the real power of any component or system of network. Real power's formula is little bit different as compared to the simple power formula. Real power is calculated as follows:

be done so as to control the budgeting of IT industry [1, 2]. Power consumption of data centers calculated in year 2012 was approximately about 270 TW/h, if compare this with the world's power consumption, we find out that it is only 2% consumption. New York times has also published a report in year 2012 which stated that the power consumption is equal to 29.9 GW/h. This consumption is almost equal to the power generation of 30 nuclear plants. But the power consumption of data centers is increasing at a high rate per year. It is increasing 4.3% annually, which is an alarming situation for all industries and organizations.Therefore, power profiling has been one of the latest research areas for many scientists. Researchers are continuously doing research on finding a better solution to overcome the access of power consumption. Power profiling is human friendly and system friendly technique that monitors power consumption properties of computational resources [3]. In this research paper we have performed power profiling by putting load on every important part of the computer and this load is being monitored by power monitoring tool, then all the data which has been collected by the monitoring tool is made available for analysis. The main aim in power profiling is to keep the quality of service intact, while monitoring the energy consumption of network components. There are two types of power monitoring tools i.e. Software tools and Hardware tools. Most common tools used to check the behavior of real world process or environment are the software tools. These tools are also known as simulators [6].


Already done work relating to this topic is introduced. Below in the references section, if we take a look on the reference number [7], those authors also worked on power profiling of various network components in an operational phase. They did power profil ing of heterogeneous network both in residential and commercial networks, keeping in mind the more network components. They found out that the energy consumption of network components has an increment of 5% from year 2009 to 2017. In another research paper i.e. [4], the authors have done research on specifically network switches, finding the power consumption of network switches. Network switches are the integral part of the whole communication network. These switches are busy all the time, so knowing the


Real Power = Apparent Power × Cosθ (1) Real power is the sum of instantaneous powers calculated at different samples divided by the number of samples. Therefore, mathematically real power can be calculated as follows:N = no. of samplesI-power = I-voltage × I-currentSum-I-power + = I-powerReal Power = {Sum-I-power} / {no. of samples} (2) Similarly, the experimental approach that we have used in our research can be better explained with the help of flow chart. Because flow chart describes the process step wise and it is easy to understand. Following flow chart shows what we have done for this research paper.

Figure 1. Experimental Procedures

3.1 Mathematical Equations

In this research paper the main objective of our research is to create a power profile for the network components, therefore we must know how to calculate the power. The mathematical formula for this is shown below:Power {P} = Voltage {V} × Current {I}P = V×I (3)Similarly, formula for energy will be:Energy (E) = Power (P) * Time (T)E = P×T (4) The standard international (SI) units for power and energy are Watts and Joules respectively. Power is defined as the rate at which work is done, while energy is defined as a capacity of physical system to perform work.

4. S TAT I S T I C A L A N A LY S I S O F SWITCHES

At the start, experiments have been conducted on three different switches i.e. Cisco-2950, Cisco-3560 and Netgear GS-724T. Different time intervals were taken, starting from 30 seconds to 210 seconds with a gap of 30 seconds each. Initially after some experiments we have seen an interval at which there was very little power variation. So we decided to use that interval and put traffic load onto three different switches and then tried to get the power consumption of these switches. Traffic load must be constant for all these switches. Data packets used for these switches were changed from 50,000 to 350,000 and time interval varied from 30 seconds to 210 seconds. Results obtained from these switches are summarized in the table below. Table 1 shows the power variations of different switches, and it has been observed that latest time has the minimum value of power variation in all of the switches. Therefore, maximum time value is the most suitable time period at which traffic load can be generated in order to observe the power usage of three switches.

Table 1. Statistics of Switchesat Different Time Intervals


Table 2 shows the power profiling of three switches when maximum load with size of 1470 bytes have applied on them. The link speed on each switch is constant i.e. 100Mbps.

Table 2.Power Profiles of Switches with Maximum Load Size 1470 bytes

When we take the maximum load of size 1280 bytes and then observe the power usage of different switches. That response has also been recorded in the form of table. Table 3 shows the statistical data when we have load of approximately 1280 bytes.

Table 3.Power Profiles of Switches with Maximum Load Size 1280 Bytes

4.1. Proposed Power Consumption Model Statistical data shows the impact of parameter i.e. traffic load on three different switches relating to power consumption. We have collected the overall insight from previous results and proposed a theoretical model for power consumption under different scenarios. Our proposed model is versatile as compared to other previous works done in [11] [12]. So, power consumed by any switch is equal to the sum of Pbaseand Pparameters which can be written as follows.Pswitch = Pbase + Pparameters (5) Where Pbase represents the power consumed by the switch when none of its ports is active. On the other hand, Pparameters represents power consumed by switch when different parameters have been checked under a load environment. Above equation can be written in other form as well.Pswitch = Pbase + Pports (6) Pports means power consumed by switch using active static ports and dynamic ports. Therefore, above equation can be rewritten as follows.P s w i t c h = P b a s e + ∑ P i p o r t s ( 7 ) As it has been observed that power consumption by switch does not only depend on


of Integrated Circuits and Systems, vol. 31, pp. 1465-1484, Oct. 2012.[3] H. Kameda, E. Z. Fathy, I. Ryu, and J. Li, “A Performance comparison of dynamic vs. static load balancing policies in a mainframe-personal computer network model,” in Decision and Control, 2000. Proceedings of the 39th IEEE Conference on, vol. 2, pp. 1415 – 1420, IEEE, 2000.[4] E. Granell, S. Andrade-Morelli, E. Ruiz-Sanchez and J. Lloret,“Energy consumption study of network access switches toEnhance energy distribution,” in 2012 IEEE GlobecomWorkshops, pp. 1496-1501, Dec. 2012.

[5] R. Lent, “A sensor network to profile the electrical power consumption of computer networks,” in 2010 IEEE Globecom Workshops, pp. 1433–1437, Dec. 2010.[6] E. Jagroep, J.M.E.M. van der Werf, S. Jansen, M. Ferreira, and J. Visser, “Profiling energy profilers,” pp. 2198-2203, ACM Press, 2015.[7] C. Lange, D. Kosiankowski, R. Weidmann, and A. Gladisch, “Energy Consumption of Telecommunication Networks and Related Improvement Options,” IEEE Journal of Selected Topics in Quantum Electronics, vol. 17, pp. 285-295, Mar. 2011.[8] M. Gupta, S. Grover and S. Singh, “A feasibility study for power management in LAN switches,” in Network Protocols, 2004. ICNP 2004. Proceedings of the 12th IEEE International Conference on, pp. 361-371, IEEE, 2004.[9] P. Mahadevan, P. Sharma, S. Banerjee, and P. Ranganathan, “A power benchmarking framework for network devices,” in International Conference on Research in Networking, pp. 795-808, Springer, 2009.[10] M. Colmant, M. Kurpicz, P. Felber, L. Huertas, R. Rouvoy, and A. Sobe, “BitWatts: a process-level power monitoring middleware,” in Proceedings of the Posters & Demos Session, pp. 41-42, ACM, 2014.[11] M.A. Marsan, A.F. Anta, V. Mancuso, B. Rengarajan, P.R. Vasallo and G. Rizzo, “A Simple Analytical Model for Energy Efficient Ethernet,” IEEE Communications Letters, vol. 15, pp. 773-775, July 2011.[12] S. Herreria-Alonso, M. Rodriguez-Perez, M. Fernandez-Veiga, and C. Lopez-Garcia, “How efficient is energy-efficient ethernet?” in 2011 3rd International Congress on Ultra-Modern Telecommunications and Control Systems, pp. 1-7, Oct 2011.

ports but on packet rate as well. Suppose, Packet Transmit Time = L/C and Packet Rate = 1/(L/C). Lastly, important parameters for power consumption of switches are taken into account and final formula is generated, where f{Packetrate × L × 8} is a function that generates load when we use various parameters in different scenarios.Pswitch = Pbase + Pparametersf{Packetrate × L × 8} (8)5. CONCLUSION Experiments performed initially have shown that there will be less power variation at the maximum time interval. In our research paper, clearly it can be seen that power variation of different switches is minimum when the time interval is 210 seconds. Another experiment also conducted with respect to a unit called packet rate to find the power consumption of different switches. These experiments have shown that power utilization increases with the increase in load value and packet rate value. From all experiments it is observed that the switch named Netgear GS 724T has minimum power usage under all scenarios and it outclasses other two switches.6. FUTURE WORKIn this research paper we have tried our best to profile the power usage of different network switches by conducting various experiments with different scenarios. We have proposed a different and unique theoretical model for finding the power consumption of different network components. Moreover, our model will help to save energy in a communication network. Because power/energy saving will be a great achievement in any case. In future, this model can be enhanced to a later version with more amendments and can be used on complex scenarios having variable traffic patterns. Our model will help to have some research on the upper layer traffic properties as well. Also, this theoretical model will open a way to do power profiling of routers and other computer components/devices.

References

[1] J. ArjonaAroca, A. Chatzipapas, A. Fernández Anta, and V. Mancuso, “A measurement-based analysis of theEnergy consumption of data center servers,” pp. 63 – 74, ACM Press, 2014.[2] M. Pedram, “Energy-Efficient Datacenters,” IEEE Transactions on Computer-Aided Design


cs. 3 newlgu.edu.pk/research/images/pdf/computer-science-and-information-technology/volume-2...aftab...

Documents