Mapping Using Multithreading in Graph Search

International Journal in Foundations of Computer Science & Technology (IJFCST), Vol.4, No.3, May 2014
DOI:10.5121/ijfcst.2014.4306 67
MAPPING USING MULTITHREADING IN GRAPH
SEARCH
Pooja Devi1
, Shalini Gupta2
and Sunita Choudhary3
1,2
M.Tech Scholar, Banasthali Vidyapith, Rajasthan, India
3
Associate Professor, Banasthali Vidyapith, Rajasthan, India
ABSTRACT
Cloud computing is a way of delivery any or all information technology from computing power to
computing infrastructure, application, business processes and personal collaboration to an user as a
service wherever and whenever they need it. The cloud in cloud computing is set of hardware, network,
software, storage, service and interfaces that combine to deliver aspects of computing as a service. Shared
resource, software and information are providing to computers and other devices on demand basis. It
allows people to do things, they want to on a computer without the need for them to build an IT
infrastructure or to understand the underline technology. Cloud computing refers to application and
services that run on distributed network using virtualized resources and access by common internet
protocols and network standards. It is a moving computing and storage from the user desktop or laptop to
remote location where as huge collection of server storage system and network equipment from a seamless
infrastructure for an application and storage. Online file storage, social networking sites, webmail and
online business application are the example of cloud services. Now a day many people are connected to
internet and Social networking sites. Social network have become a powerful platform for sharing and
communication that focus on real world relationships. Social networking plays a major role in everyday
lives of many people. Facebook is one of the best examples of Social networking sites where more than 400
million active users are connected. Thus Social cloud is a scalable computing model where in virtualized
resource provided by users dynamically. In this paper we used concept of MapReduce with Multithreading.
MapReduce is a paradigm that allows for massive scalability across hundreds or thousands of servers in a
cluster. MapReduce job usually split the input data into independent chunks which are processed by the
map tasks in completely parallel manner. It sorts the output of the map which are than input to the reduce
task. Using mapping techniques is to find out a good performance in terms of cost and time.
KEYWORDS
Map Reduce, Cloud Computing, Social Graph, Social Network, MySQL .
1. INTRODUCTION
1.1 Cloud Computing
Cloud Computing is the next generation in computation. Maybe clouds can save the worlds;
possibly people can have everything they need on the cloud. Cloud Computing is the next natural
step in the evolution of on-demand information technology services and products. This
Computing is latest trend in IT world. It is internet based computing where resources, software
and information are provided to computers and other devices on demand basis. This technology
has the capacity to admittance a common collection of resources on requirements. It is proving
extremely striking to cash-strapped IT departments that are wanted to deliver services under
pressure [2, 4].

68
1.1.1. Characteristics of Cloud computing
 Self-service on-demand basis
 Wide area for network access
 Pool of resources
 Highly Elastic
 Payment on use basis
1.1.2. Challenges of Cloud Computing
 Lack of Security and Privacy
 Lack of Standards
 Continuously Evolving New Technology
1.2. Social Cloud
Social Cloud is a scalable computing model where more number of resources contributed by
users are dynamically provisioned amongst a group of friends. These resources are freely used
and shared by the users. A large number of commercial cloud providers like Microsoft Azure,
Amazon EC2/S3 and Google App Engine provides access to scalable resources [6, 9].
1.2.1. Applications of Social Cloud
 Government
 Education
 Finance application
 Medical and health application
1.3. Social Network
Social network is a structure of interconnected entities that shows relationship between each
other. These entities are referred to as “users”. The relationship between the users has different
friends and followers. With these relationship users can share message and media amongst them.
Some popular Social networking sites like- Facebook, Twitter has million active users. A social
network is also called a set of people or groups of people with some pattern of contacts or
interactions between them [1].Cloud computing and social networking has intermingled in a
variety of ways. Cloud computing and social networks have numerous example of being hosted
on cloud. Most social networks can be hosted on cloud platforms or have a scalable applications
within the social networks. Benefits of Social cloud over a social network are flexibility, disaster
recovery, automatic software updates, increase collaboration, document control, security etc.
Uses of Social Network
 Purely Personal Reasons
 Business
 Marketing
 Entertainment

69
Social Graph
Social graph is a diagram that depicts the relationship among people and groups in Social
network. In this graph Individuals and organizations called actors. Interdependencies called ties
that can be multiple and diverse including such characteristics or concepts as age, gender, ideas,
financial transactions, trade relationships, political affiliations, club memberships, occupation,
education and economic status. The social graph in the Internet context is a graph that depicts
personal relations of internet users [3, 5]. Services such as Facebook allow users to exchange
information, news and other things among users. The social graph for a particular user consists of
the set of nodes and ties connected directly or indirectly to that actor. To search information in
short time in Social Graph, the concept of Parallel Computing is widely result.
Figure 1. Social Graph
1.4. Parallel Computing
In parallel computing a task is divided into multiple subtasks using MapReduce concept and each
one of them are processed on different CPUs. Programming on multiprocessor system using
MapReduce concept is called parallel programming Mostly Parallel computing is used to reduce
the execution time and to utilize larger memory/storage resources. The essence of parallel
computing is to partition and distribute the entire computational work among processor and each
processor can exchange information [15].
Why Do Parallel Computing?
Many applications today require more computing power than a traditional sequential computer.
Parallel computing provides s cost-effective solution to problems by increasing the number of
CPUs in a computer and by adding an efficient communication between them.
The development of parallel processing is being influenced by many factors.
 Overcome limits to serial computing.
 Limits to increase transistor density.
 Limits to data transmission speed.
 Prohibitive cost of supercomputer.
 Commodity components to achieve high performance.
 Faster turn-around time.
 Solve larger problems.

70
2. MODULES TO IMPLEMENT MAPPING
2.1.1. Search Graph
This is the first module to implement Mapping process. In this module we make a graph by using
graph simulator. This graph simulator provide some facilities like-
 How to create nodes
 How to connect nodes.
This module show that how to connect nodes with single node and after that connected nodes
connect with other nodes and finally connect initial node and final node at third level. In other
way we can say make a graph where one node connected to other node and other node is
connected to more number of other nodes and we find out the common interest between one node
and other more nodes in very efficient and good manner. Here node act as a people and
connection between nodes make an edge. Figure 2 shows that initially one person (N1) have two
friends.
Initial level 1st
level
Figure 2
Now one people (N1) has two friends (N2 and N3) but node N2 and N3 also have two friends
(N4, N5 and N6, N7). This is shown by figure 3.

71
Initial level 2nd
level Third level
Figure 3
Above figure shows that these are the very small graph but now a day’s social networking sites
frequently used by the many other people, so number of contacts will increase day by day and
also accessing will become very difficult. Then we assume 20 nodes at initial level and each have
2 friends and at 3rd
level each have 2 friends so we have 400 total number of friends to search
this is very difficult and searching process become very complex and take more time. To solve
this problem clustering techniques has used. In this module we also focus on database that have
all information about the people (Nodes) like- Name, Job, Gender, organization etc and show all
details on the screen like- Name, city, Gender, Job, Organization and relationship on a node by
fetching the information from the database and also show wanted node information when we
select that particular node.
2.1.2. Cluster Making
This is the second module to implement Mapping Process. In this we are considering 20 nodes
initially. Suppose 20 nodes have 20 connected nodes and these 20 nodes also have 20 connected
nodes so there are total 400 nodes to find out a good performance result with less cost and time is
very difficult. To solve this problem we making cluster of these huge data so that graph searching
become easy by cluster making.
In this module we use a concept of MapReduce with Multithreading on particular level to solve
search graph problem. In this concept we take different no of threads to solve above problem.
When we apply no of threads on nodes then we find a cluster wise result so that graph searching
can easily do. This can be explained by the following example
In this paper we use only mapping with multithreading we can explain by an example-
Suppose initially we take 7 nodes. In this case at 3rd
level total four numbers of nodes are
available and at that level we perform mapping using processor. At that level only two threads are
required. Firstly we map all four nodes information onto one thread and after that we map four
nodes information onto two threads that divide mapping time but can increase overhead time.

72
Seven Nodes with One Thread:
Figure 4
Seven Nodes with Two Threads:
Figure 5
Now second example of 15 nodes in this case we use four threads. In this case at 5th
level eight
nodes are available and that level we want to perform mapping with different no of threads.
Firstly we map information of all eight nodes onto one thread and after that we map information
of eight nodes by dividing two clusters onto two threads and after that we map information of
eight nodes by dividing four clusters onto four threads.

73
Fifteen Nodes with One Thread:
Figure 6
Fifteen Nodes with Two Threads:
Figure 7

74
Fifteen Nodes with Three Threads:
Figure 8
With this overall processing we further calculate parallel efficiency in terms of time and cost of
number of threads and find out which threads have good efficient.
2.1.3. Processing
In this phase mapper and reducer will work accordingly means mapper maps the nodes
information and then reducer will reduce or combine extracted information which is the output of
mapper in to one. The MapReduce programming model generalizes the computational structure
of the above example. Each map operation consists of transforming one set of key-value pairs to
another:
Map: (K1, v1) → [(K2, v2)]
The reduce operation groups the results of the map step using the same key k2 and performs a
function f on the list of values that correspond to each key value:
Reduce: (K2, [v2]) → (K2, f ([v2]))
The implementation also generalizes. Each mapper is assigned an input-key range (set of values
for K1) on which map operations need to be performed. The mapper writes results of its map
operations to its local disk in R partitions, each corresponding to the output-key range (values of
K2) assigned to a particular reducer and informs the master of these locations. Next each reducer
fetches these pairs from the respective mappers and performs reduce operations for each key K2
assigned to it [8]. But in this paper we use only mapping process with multithreading Reducing
are not required means we done mapping of all information’s of nodes.
CONCLUSION
From all experiment we can come to the following conclusions on performing parallel computing
using cloud technologies. Cloud technologies work well for problems because Cloud computing
support large data sets. In this paper we are trying to simulate the concept of mapping on number
of nodes of a Graph. Using mapping concept we find out Parallel efficiency ratio with the use of
different number of Mappers and conclude how many numbers of Mappers are required to

75
calculate good Parallel efficiency ratio. In this paper we observe that this work may become
helpful to find out people throughout world with their more or less information.
FUTURE WORK
We can enhance this work by adding more and more searching constraint and number of nodes
for extraction. As we know that Social Network site like- Facebook one can add personal profile
with their information like- name, city, job, relationship and any other information so on. In this
paper we consider only two or three constraint (city, organisation) for extraction of people
information among large amount of information. But in future we can search or extract
information of any person with many attributes. MapReduce concept has two phase, mapper and
reducer. In this paper we perform searching only on 2o nodes that mapping phase is efficient, but
if we perform searching on more number of nodes then reducing phase must be required for
getting efficient result. In future we can implement a reducing phase in MapReduce concept for
frequent search means we can get optimized result so that we obtain better search with reduced
cost and time.
ACKNOWLEDGEMENTS
We thanks to the Dr. Sunita Choudhary mam who have contributed towards development of the
paper.
REFERENCES
[1] A. Fortino, N.A.,” architecture for applying social networking to business,” Applications and
Technology Conference (LISAT), 2010 Long Island Systems, vol., no., pp.1-6, 7-7 May 2010 doi:
10.1109/LISAT.2010.5478285.
[2] A.K. Hussein, Research Agenda in Cloud Technologies, Sriram Department of Computer Science,
University of Bristol, UK 2009.
[3] D. Moskovitz, Facebook Meets the Virtualized Enterprise, Washington, DC, USA, 2008. IEEE
Computer Society.
[4] F Fabris, Introduction to Cloud Computing Architecture, Sun Microsystems, Inc.
[5] K. Chard, S. Caton, y. Omer Rana, z. Kris Bubendorfer, Social Cloud: Cloud Computing in Social
Networks School of Engineering and Computer Science, Victoria University of Wellington, New
Zealand.
[6] M. M. Alabbadi, "Cloud Computing for Education and learning as a service", 14th International
Conference on Interactive Collaborative Learning (ICL)-11th International conference Virtual
University (vu'11), Slovakia, 2011.
[7] M. Taber, MySQL 5.0 Reference Manual, www.gen.tcd.ie/molevol/pdfs/mysql_tut.pdf.
[8] P. Seshadri, M. Sridhar an, A. Maharaja, S. Faloutsos, “Eigen Spokes: Surprising Patterns and
Scalable Community Chipping in Large Graphs,”MapReduce and Hadoop workshop, 2009. ICDMW
‘09. IEEE International Conference on, Dec. 2009 doi: 10.1109/ICDMW.2009.103.
[9] R. A. Stoica, I. Zaharia (2010). “A View of Cloud Computing”, Communications of the ACM, 53(4).
[10] S. Meyer, eclipse Juno, www.eclipse.org/whitepapers/eclipse-overview.pdf.
[11] T. Dillon, W. Chen, and E. Chang, Cloud Computing: challenges, Proc 24th IEEE International
Conference on Advanced Information Networking and Applications (AINA), Perth, Australia, April
2010.
[12] U. K. Wiil, J. Gniadek, N. Memon; Measuring Link Importance in Terrorist Networks. Social
Network Analysis, 225-232. Ed. Nasrullah Meson, Reda Alhajj (Eds.): International Conference On
Advances in Social Networks Analysis and Mining, ASONAM 2010, Odense, Denmark, August 9,
11, 2010. IEEE Computer Society 2010, ISBN 978-0-7695-4138-9.

76
[13] W.Lewis, T.Chu, W.Salehi-Abari, “Media Monitoring Using Social Networks,” Social Computing
(Social Com), 2010 IEEE Second International Conference on, vol., no., pp.661-668, 20-22 Aug.
2010 doi: 10.1109/SocialCom.2010.102.
[14] Y. u. Zhang, C. Xia, “Identifying Key Users for Targeted Marketing by Mining Online Social
Network,” Advanced Information Networking and Applications Workshops (WAINA), 2010 IEEE
24th International Conference on, vol., no., pp.644-649, 20-23 April 2010
doi:10.1109/WAINA.2010.137.
Books:
[15] Gautam Shroff, ENTERPRISE CLOUD COMPUTING TECHNOLOGY, ARCHITECTURE,
APPLICATIONS. Published in the United States of America by Cambridge University Press, New
York.
Authors
Pooja Devi is an active researcher in the field of Cloud Computing, currently studying in
M.Tech (IT) from Banasthali University.
Shalini Gupta is an active researcher in the field of Cloud Computing, currently studying
in M.Tech (IT) from Banasthali University.
Sunita Choudhary is an Associate Professor in Department of Computer Science at
Banasthali University (Rajasthan), India. She has done PhD from Banasthali University
(Rajasthan), India.

Mapping Using Multithreading in Graph Search

Recommended

Recommended

More Related Content

What's hot

What's hot (17)

Viewers also liked

Viewers also liked (18)

Similar to Mapping Using Multithreading in Graph Search

Similar to Mapping Using Multithreading in Graph Search (20)

More from ijfcstjournal

More from ijfcstjournal (20)

Recently uploaded

Recently uploaded (20)

Mapping Using Multithreading in Graph Search