Unit 4

• Ubiquitous Clouds and the Internet of Things:

– Cloud Trends in Supporting Ubiquitous Computing
– Performance of Distributed Systems and the Cloud
– Enabling Technologies for the Internet of Things
– Innovative Applications of the Internet of Things
– Online Social and Professional Networking
 Cloud Trends in supporting Ubiquitous
Computing
1) Use of Clouds for HPC/HTC and Ubiquitous Computing
• Ubiquitous computing is a concept in software engineering and
computer science where computing is made to appear anytime
and everywhere.
• In contrast to desktop computing, ubiquitous computing can
occur using any device, in any location, and in any format.
• A user interacts with the computer, which can exist in many
different forms, including laptop computers, tablets and
terminals in everyday objects such as a refrigerator or a pair of
glasses.
• Ubiquitous cloud computing refers to the use of Internet
resources at any place and any time for any objectives.
a)Herbert Simon on Information Consumption
He says, ”What information consumes is rather
obvious: it consumes the attention of its
recipients. Hence a wealth of information
creates a poverty of attention and a need to
allocate that attention efficiently among the
overabundance of information sources that
might consume it.”
This clearly outlines a supply and demand
model in the information age.
b)Vannevar Bush on Predicting Memex
He predicted a dream device called Memex that
stores all books, records, communications of
an individual, and is mechanized so that it may
be consulted with exceeding speed and
flexibility. It is an enlarged intimate
supplement to his memory.
This Memex is nowadays is iPad, smartphone,
cloud, IoT etc.
c)Cloud Service Trends
• Many Companies like Amazon, Enomaly,
Google, GoGrid, Microsoft, NetSuite,
Rackspace and Saleforce.com offer various
financial services, communications and media,
energy and health care.
d) IBM Cloud Projects

• IBM has gradually changed from a

hardware/software and system company to a
computing services company.
• The IBM cloud platforms are mostly built with
IBM server clusters supported by IBM
Websphere and accumulated software assets
over the years. Both the z-series and p-series
servers at IBM are being upgraded with
virtualization capabilities.
e)IBM RC2 Cloud

• The computer and IT infrastructure of eight IBM

Research Centers are now strongly connected to
form a private cloud, called the Research Compute
Cloud(R-C2).
• RC2 is web-based, allowing more than 3,000 IBM
researchers worldwide to share computing resources.
• It is also IBM’S internal testbed for promotion of
cloud technology. The system supports machine
virtualization ,service life-cycle management and
performance monitoring.
f)Cloud System from SGI

• In February 2010, SGI announced the Cyclone,

a large-scale on-demand cloud computing
service for high-performance computing (HPC)
applications. Both SaaS and IaaS models are
offered on the Cyclone.
g) Salesforce.com’s Force.com Cloud

• Salesforce.com was established in 1999 to

provide online solutions to SaaS, mainly in
CRM applications. Initially, it used third-party
cloud platforms to promote its software
services.
• Gradually, Salesforce.com launched Force.com
as a PaaS platform. By 2010, Salesforce.com
subscribers exceeded 2.1 million.
2)Large scale Private Clouds at NASA and CERN

• In 2010, two large scale private clouds were under

construction in United States and European Union
• The U.S. cloud called, Nebula is developed by NASA and is
designed for NASA scientists to run climate models on
remote systems provided by NASA. This can save thousands
of NASA users from acquiring super computers at their local
sites.
• The EU cloud is built by CERN in Geneva. This is a large
private cloud for distributing data, applications and
computing resources to thousands of scientists around the
world. CERN deals with large data sets and through put with
a global work force and a finite budget.
a)NASA’s Nebula Cloud

• Nebula is a cloud computing pilot under development

at NASA Ames Research Center.
• It integrates a set of open source components into a
seamless, self-service platform, providing high capacity
computing, storage and network connectivity using a
virtualized ,scalable approach to achieve cost and
energy efficiencies.
• Nebula enhances NASA’S ability to collaborate with
external scientists and researchers by providing high-
speed data connections and consistent tool sets and
open data APIs used by commercial cloud providers.
b)Nebula Development Status

• Nebula is available to NASA’s internal project

groups and its research and academic
partners. Nebula is not available for use by
private industry or the general public. Nebula
is also currently acting as a test bed for federal
cloud computing technology for the Office of
the Management and Budget.
c)CERN Super Cloud

• CERN has long been an influential leader in solving

complex IT problems.
• In fact, the World Wide Web owes a lot to CERN for its
usefulness in getting some early work done there.
• Now the focus is on cloud computing, CERN scientists plan
to smash two particle beams together in an attempt to re-
create the big bang event to uncover the enigma of many
fundamental physics phenomena.
• The CERN scientists realize the use of cloud computing
over their large data centers may be more cost-effective in
performing the required simulation experiments.
d) CERN VM Projects

• Scientists working on the ALICE (A Large Ion

Collider Experiment)collaboration are conducting
heavy ion simulations at CERN.
• They have been developing and debugging
compute jobs on a collection of internationally
distributed resources, managed by a scheduler
called AliEn, to integrate a cloud’s dynamically
provisioned resources into an existing
infrastructure such as the ALICE pool of
computers.
3)Cloud Mashups for Agility and Scalability
a)The idea of a Cloud Mashup
• In web application development, a mashup is
a web page or application that combines
data,presentations, or functionality from two
or more sources to create a new service.
• The main characteristics of the mashup are a
combination of virtualization and aggregation.
• The mashup design leverages Google’s web
agility with the scaling power of AWS EC2.This
mashup helps a user to write agile software on
GAE and use the user inputs to perform
parallel computing operations on AWS.
b)Advantages of Cloud Mashup

• Scalability on the EC2 virtual clusters Scaling the

implementation of MapReduce implementation can be
scaled in terms of problem size and virtual cluster size.
• Agility in using GAE interfaces
A mashup platform is build by combining GAE and
AWS,essentially using Google’s web interfaces and
Amazon’s EC2 computing power. We use GAE to create
a web page that is capable of uploading documents to
GAE and running MapReduce on EC2. The mashup
then sends the results back to Google users.
4)Cloudlets for Mobile Cloud Computing

a)The Cloudlet Idea

• Mobile handsets face problems of limited CPU power,storage
capacity, and network bandwidth on smartphones and tablet
computers.
• Mobile devices cannot be used to handle large data sets.
However,the distant cloud on the Internet faces the WAN-latency
problem.
• Clouds have to solve this collision problem of too many clients
logging in to the cloud simultaneously.
• To solve this two-sided problem cloudlets are being deployed at
convenient sites such as coffee shops and bookstors, similar to
access points for WiFi services for connecting to the Internet.
Differences between Local Cloudlets and
Distant Clouds
 Performance of Distributed systems and
the Cloud
1)Review of Science and Research Clouds
a)Future Grid
Future Grid is a NSF project funded as an experimental component of TeraGrid
and designed around the concept of Grid5000.FutureGrid is led by Indiana
University in partnership with Purdue University ,the San Diego Supercomputer
center at the university of California and some other universities.

b)Grid’5000
The Grid’5000 distributed system links nine sites in France and has more than
5000 cores spread over 1500 nodes. It has been used in a variety of computer
science research projects including cloud and grid computing and green IT with
both software systems and performance goals.
c)Magellan
Magallan is the largest of the research clouds and is situated at
two U.S. Department of Energy sites. Argonne and the
National Energy Research Scientific Computing
Center(NERSC) in California. The NERSC system has more
than 700 nodes and about 6000 cores, while Argonne has more
than 500nodes and 4000 cores
d)Open Cirrus
This is a broad consortium of some 14 partners exploring
cloud computing. The quite powerful test-bed consists of more
than 600 nodes at four major sites. The Open Cirrus testbed is
designed to support research into the design, provisioning and
management of services at a global, multidata-center scale.
e)Open cloud Testbed
• The Open Cloud Consortium(OCC) is a member-driven
organization that supports the development of standards for
cloud computing and frameworks for interoperating between
clouds, develops benchmarks for cloud computing and
supports reference implementations for cloud computing.
• The OCC also manages a testbed for cloud computing, the
Open Cloud Testbed and operates cloud computing
infrastructure to support scientific research, called Open
Science Data Cloud with an emphasis on data-intensive
computing.
f)Science Clouds
• Science clouds is an open cloud federation providing
compute cycles in the cloud for scientific communities
exploring the use of Nimbus. The infrastructure federates
sites across Europe and United States.
g)Sky Computing
This project is aligned with Science Clouds and ,like the
European Reservoir project, is designed to federate
multiple clouds into a single resource. Linking separate
clouds together can help scalability, but also improves
fault tolerance. The federation also supports the special
case of a hybrid cloud-two clouds, one private and one
public.
h)Venus-C
Venus-C is an European project exploring the
applicability of cloud computing in areas such
as building structure analysis,3D architecture
rendering ,prediction of marine species
population, risk prediction for wildfires,
metagenomics(study of genetic material
recovered from environmental
samples),system biology, and drug
development.
2)Data-Intensive Scalable Computing

a)Big Data Sources

In today’s data-driven world, huge databases are generated
from science in astronomy, genomics, natural languages and
seismic modeling among other applications. In addition
scanned books, newspapers and historical documents are
creating a data ocean. In commerce, corporate sales, stock
market transactions, census data, and airline traffic data all
demand the handling of large data sets on a daily basis.
b)DISC Systems versus Supercomputers
3)Performance Metrics for HPC/HTC Systems

a)System Throughput and Efficiency

• System Throughput of a distributed system or a cloud platform
measures the number of jobs that can be done per unit of time.
• The throughput measure is attributed to several key factors that
affect the total execution (Ttotal ) of all jobs processed in a given time
window.
• The overhead can be attributed to five components
• 1)Infrastructure initiation delay,
• 2)Resource Provisioning delay,
• 3) Interjob communication delay,
• 4)OS overhead,
• 5)loading application software overhead.
• The OS and software overheads remain constant in
running different jobs. The other three time factors
may vary with problem size and system
management policies.
• For simplicity in analysis, we lump all five time
factors into a single overhead time kT0 ,where T0 is
their sum averaged over many jobs running in a
fixed time interval. The constant k is a platform-
dependent coefficient that varies with specific
system configurations.
• We compute the total job execution time with
Ttotal =kT0 +Te (n,m).
• Then system throughput is given by ∏=n/Ttotal =n/[kT0
+Te (n,m)]
Where Te (n,m) be the effective execution time to
complete n independent jobs in a cloud platform, where
m accounts for the number of machine instances in a
given system configuration.
The system efficiency(α) is defined by normalized
throughput as follows:
α= ∏/[n/Te (n,m)]=Te (n,m)/[kT0 +Te (n,m)]
b)Multitasking scalability(B)
• Multitasking implies the use of a system to handle many jobs
simultaneously or concurrently.
• By horizontal scalability, we mean to increase cloud resources
of the same type. For example, the number of VM instances
offers such a measure in a cloud.
• Vertical scalability means the increase in performance due to
adding more resources along the service layers.
• We define multitasking scalability(B) as
B=(n/m) α=(n/m) Te (n,m)/[kT0 +Te (n,m)]
Where m is the number of machine instances
and n is the number of user jobs being processed and α is system
efficiency.
c)System Availability(Y)
• System availability(Y) refers to the percentage
of time the system is up and running normally.
This percentage reflects the effects of
downtime after unexpected failures and
scheduled maintenance for software upgrades.
• System availability is defined by the ratio of
MTTF(mean time to failure) to the sum of
MTTF and MTTR(mean time to repair)
• High availability (HA) specifies the performance level of a
system with controlled downtime through quick recovery
from failure.
• The term “continuous availability” (CA) refers an
operational level with almost no failure. Generally
speaking, system availability should be the responsibility
of the providers.
• Usually, availability is maintained sufficiently high with at
least five 9s, meaning that with 99.999 percent of uptime
(at most, five minutes of downtime per year) the system
can provide services.
d)Security Index(δ)

• Google has hundreds of data centers and more

than half a million servers. Cloud security is
attributed to user confidentiality, data integrity,
access control, firewalls, IDSes, defense
capability against viruses or worm attacks,
reputation systems, data lock-in.
• The resource site security index and user
access records should be maintained through
periodic auditing.
e)Cost Effectiveness(µ)

• The cost model considers cost by both the cloud service provider and
the data-center owner. The provider must pay the data-center owner
for the resources used even if they belong to the same company.
• We denote c as the hourly charge to the user by the provider and h as
the number of service hours provided to end users.The total income of
the provider from user payment of the service is calculated by CostP
=hc.
• The total cost the provider paid to the data center is estimated by Costd
=h(d/u) where d is the hourly charge by the data center and u is the
utilization rate (oftern between 0.6 to 0.8)of the data-center resources.
• The cost-effectiveness(µ) of the cloud service is a measure for assessing
the profit margin by the service provider.
• µ=(Costp - Costd )/Costd =(hc-hd/u)/(hd/u)=cu/d-1
4)Quality of Service in Cloud Computing

• We calculate the cumulative performance of a cloud platform

by a compound metric, called quality of cloud services(QoCS)
and denoted by (θ).This metric measures the overall quality of
the cloud service, regardless of the particular service models
applied
• We use a five-dimensional (System efficiency(α), Multitasking
scalability(B), System Availability(Y), Security Index(δ), Cost
Effectiveness(µ)) Kiviat graph to plot the aggregated cloud
performance or to quantify the QoCS graphically.
• θ= Ashaded /Apentagon where Ashaded and Apentagon are the
areas of shaded and outer pentagons as shown in the below
figure.
 ENABLING TECHNOLOGIES FOR THE INTERNET OF
THINGS

• In ubiquitous computing, the Internet of Things(IOT)

provides a network of sensor-or radio connected
devices that can be uniquely identified and located in
the cyber-physical space.
• The term IOT combines RFID technology with today’s
IPv6-based Internet technology. All things(objects)
have IP addresses, which can be uniquely identified.
• The IP-identifiable objects are readable, recognizable,
locatable, addressable and controllable via the
Internet, aided by RFID, WiFi, ZigBee, mobile
networks and GPS.
1)The Internet of Things for Ubiquitous
Computing
• The IoT is a natural extension of the Internet.
• The foundation of the IOT is radio-frequency
identification(RFID).
• This enables the discovery of tagged objects
and mobile devices searching by browsing an
IP address or searching for a database entry.
a)Ubiquitous Computing

• Ubiquitous computing is a post-desktop model of

human-computer interaction in which information
processing is integrated into everyday objects and
activities.
b)Development of Internet of Things
• In 2005,the concept of IoT entered the limelight. The IoT
should be designed to connect the world’s objects in a
sensory manner. The approach is to tag things through
RFID, feel things through sensors and wireless networks,
and think things by building embedded systems that
interact with human activities.
c)Enabling and Synergistic Technologies
Supporting technologies are divided into two
categories.
• 1)Enabling Technologies build up the
foundations of the IoT. Examples are RFID,
sensor networks and GPS are critical.
• 2)Synergistic technologies play supporting
roles. Examples are Biometrics, Augmented
Reality, Robotics, Machine vision.
d)Architecture of the Internet of Things
• The IoT system is likely to have an event-
driven architecture. IoT development is shown
with a three–layer architecture.
2)Radio-Frequency Identification(RFID)

• RFID is applied with electronic labels or RFID tags

on any objects being monitored or tracked.
• The tagging may be applied to any objects, such as
merchandise, tools, smartphones ,computers, animals
or people.
• The purpose is to identify and track the objects using
radio waves or sensing signals.
• Some tags can be read from tens or hundreds of
meters away via a wireless reader.
a)RFID tags and Device components

• In terms of functionality, there are three major

components of RFID hardware:
• i)RFID tag : A tiny silicon chip attached to a small
antenna
• ii)Reader Antenna : Used to radiate the energy and
then capture the return signal sent back from the tag.
It can be integrated with a handheld reader device or
connected to the reader by cable.
• iii)Reader: The device station that talks with the
tags. A reader may support one or more antennae.
b)How RFID works
• There are two types of RFID tags;1)Active RFID
tags 2)Passive RFID tags
• Active RFID tags contain a battery and transmit
signals autonomously.
• Passive RFID tags have no battery and require an
external source to provoke signal transmission.
• To read the information encoded on a tag, a two-
way radio transmitter-receiver called an interrogator
or reader emits a signal to the tag using an antenna.
The tag responds with the information written in its
memory bank. The interrogator will then transmit
the read results to an RFID computer program.
3)Sensor Networks and ZigBee Technology
• Today’s sensor networks are mostly wireless, and are
known as wireless sensor networks(WSNs).
• A typical WSN consists of spatially distributed
autonomous sensors to cooperatively monitor physical
or environmental conditions such as temperature,
sound, vibration, pressure, motion or pollutants.
• Applications are industrial and civilian application
areas, process monitoring and control, machine health
monitoring, environment and habitat monitoring,
healthcare and home automation and Intelligent traffic
control.
a)Wireless Sensor Networks
• The Sensor network consists of multiple detection stations
called sensor nodes, each of which is small, lightweight, and
portable.
• Every sensor node is equipped with a transducer,
microcomputer, transceiver and power source.
• The transducer generates electrical signals based on sensed
data.
• The microcomputer processes and stores the sensor output.
• The transceiver, which can be hardwired or wireless, receives
commands from a central computer and trasmits data to that
computer.
• The power for each sensor node is derived from the electric
utility or from a battery.
• A sensor node may vary in size from that of a shoebox down
to a dust grain.
b)Wireless Network Support for Ubiquitous
Computing
• Here we discuss four classes of wireless
networks for ubiquitous computing. They are
• i) ZigBee 802.15.4
• ii) GSM/GPRS/CDMA/1XRTT
• iii) WIFI 802.11g
• iv) Bluetooth 802.15.1
• The ZigBee networks are mainly used in low-cost and
low-speed monitoring and control applications such as
those used in a wireless home are network(WHAN).
• The GSM/GPRS/CDMA networks are cellular mobile
networks that can cover a wide area for voice and data
telecommunications.
• The WiFi 802.11g is used for wirelessly accessing the
Internet, reading e-mails and for web search.
• Bluetooth is mainly used for short distance wireless
connection of computer peripherals.
• Of all above four standards, ZigBee has the slowest rate of
20-250kbps.However ZigBee has high reliability and low
power/cost advantages.
c)Three generations of Sensor Networks
• i)First Generation(1990’s)
• ii)Second Generation(2000’s)
• iii)Third Generation(2010’s)
• In the first generation, the sensors were mainly vehicle-placed or
air-dropped single sensors. They were quite bulky, like a shoebox
and weighed in the kilograms and used in point to point and star
topologies.
• In the second generation, the sensors became smaller, like a pack
of playing cards, and weighed in grams and used in client/server
or P2P systems.
• In the third generation, sensors are as large as dust particles and
are used in P2P networks for embedded and remote applications.
d) ZigBee Network

• ZigBee is a high-level communication

protocol using small, low-power, radio-based
sensors as specified by the IEEE 802.15.4
standard.
• ZigBee devices are applied in WHANs in
wireless switches and consumer electronics
controlled by short-range radio.
• Zigbee device is simpler to use and less
expensive than Bluetooth or WiFi.
There are three different types of ZigBee devices.
i)ZigBee Coordinator(ZC)
ii)ZigBee Router(ZR)
iii)ZigBee End Device(ZED)
• ZigBee Coordinator is unique to a network which starts the
network and acts as trust center and repository of security
keys.
• ZigBee router acts as a intermediate router passing on data
from end device to end device.
• ZigBee End device contains just enough functionality to
talk to the parent node(either the coordinator or the router)
and cannot relay data from other devices and requires least
amount of memory.
Structure of Typical ZigBee network
4)Global Positioning System(GPS)
Location-based service(LBS) helps people and machines find things and
determine their physical whereabouts.
• a)How GPS Works
GPS consists of three parts.
i)Space segment
ii)Control segment
iii)User Segment
• The U.S AirForce develops, maintains and operates the space and control
segments. There are 24 satellites deployed around the Earth in fixed orbits
at an altitude of approximately 20,200KM.
• GPS satellites broadcast signals from space by which each GPS receiver
calculates its 3D location(Longitude, Latitude and Altitude) plus the current
time.
• The Control Segment is composed of a master control station and a host
of dedicated and shared ground antennas and monitor stations.
• The User segment is composed of hundreds of thousands of U.S. and allied
military users of the secure GPS precise positioning services.
b)Passive versus Active GPSes
• In passive tracking, the GPS is just a receiver, not
a tansmitter. They lack a transmission capability
to send the GPS data from the vehicle and so are
used as loggers and recording devices.
• Active GPS tracking units incorporate a method to
transmit user information from a vehicle.
• Passive GPS tracking devices store GPS location
data in their internal memory which can then be
downloaded to a computer for viewing at a later
time,while active GPS tracking systems send the
data at regular intervals to be viewed in real time.
 Innovative applications of the Internet of
Things
1) Specific Wireless Sensor Applications
2) Retailing and Supply-Chain Management
3) Smart Power Grid and Smart Buildings
4) Cyber-Physical System(CPS)
 1)Specific Wireless Sensor Applications

• Military networks to detect and gain as much

information as possible about enemy movements,
explosions and other phenomena of interest.
• Sensor networks to detect and monitor
environmental changes in plains, forests, oceans and
so on.
• Wireless traffic sensor networks to monitor vehicle
traffic on highways or in congested parts of a city
• Wireless parking sensor network to determine
whether the lot is occupied or not.
2)Retailing and Supply-Chain Management

• IoT applications will promote business, improve our

society, and enhance economical growth.
• These include specific IoT applications in railing and
logistics services and supply chain management.
• Many industrial, government, and community
services can benefit from these applications.
• These include activities or initiatives to promote the
development of better and more efficient
societies,cities, and governments.
 a) Retailing and Logistics Services

• Emergence of RFID applications depends strongly on

adoption by retailers, logistics organizations and
package-delivery companies.
• Retailers may tag individual objects in order to solve
a number of problems at once: Accurate
inventorying, loss control, and ability to support
unattended walk-through point of sale terminals
which promise to speed checkout while reducing
both shoplifting and labor costs.
 Supply Chain Management
• Supply chain management can be aided by an IoT system. The
idea is to manage a whole network of related businesses or
partners involved in product manufacturing, delivery, and
services as required by end customers.
• A supply line combines the processes, methodologies, tools,
and delivery options to guide collaborative partners to work
in a sequence to conduct business with high efficiency and
delivery speed.
• A supply chain is an efficient network of facilities that
procures materials, transforms these materials to finished
products, and finally distributes the finished products to
customers.
Supply chain management in multipartner
business area
3)Smart Power Grid and Smart Buildings

• Various sensors at individual homes (smart

thermostats) can collect information that is sent via a
network to main stations that can apply complex
power management and send control signals back to
the grid to save energy.
• Smart buildings may include alarm systems, access
controls, indoor climate controls, elevators and so
forth.
 4)Cyber Physical system(CPS)
• A cyber-physical system(CPS) is an embedded system
which integrates the computing process with the
physical world as an interactive and intelligent system.
• Cursor based CPSes can be found in the automotive,
aerospace, health care, robotics, manufacturing,
battlefield training and consumer appliance industries.
• It is desired to design CPSes to replace humans in
dangerous or inaccessible environments such as
battlefields, rescue from earthquake rubble, and
deep-sea exploration
 Online Social and Professional Networking
Online social networking enables many social and professional functions over the
Internet.
1)Online Social Network Characteristics
• Online social networks(OSNs) are formed with Individual or organizational
entities are related, connected, or associated special interests or specific
interdependencies.
• These include personal friendships, kinships, professional orientations,
common interests, financial exchange, like or dislike relationships, community
or racial groups, religious or political beliefs, knowledge or prestige, and
celebrity fans.
• In general, social networks are formed by complex sets of interpersonal to
international relationships among members of our societies or communities
on Earth.
• OSN services are built to reflect the social relationships among people. Social
networking services are a communication tool among people.
2)Online Social Network Services
• An OSN offers special social or professional services through
Internet access and web services. These social networking
services are linked by Internet communication among people.
Here are some ideas for providing OSN services:
a)Personal page or profiles for each user linked by social
connections
b)Social graph traversal along specific social links or networks
c)Ability to share music, photos and videos with friends or
professional groups.
d)Operation of a community in special niche areas like
healthcare, sports and hobbies
3) Social Networking Properties
• In most social networks, each user has a personal page (or profile),
which is linked via the user’s social relationships. This is why
modern social networking web sites are individual-oriented.
• Once a user accepts other users as “friends,” he will be able to
access all permitted contents on those friends’ profile pages. Also,
users can access their friends’ lists of friends, and view their profile
pages. This is an example of social graph traversal.
• Commonly, online social networking web sites provide two basic
communication methods: internal e-mail-like direct messages and
public message boards.
• Also, users can share information, such as photos, videos, web links,
and articles with friends.
• All the popular online social networking web sites have these four
essential items, but the way they implement these features is
different based on the different service models and market
orientations.
4)Social Networks Benefits
a)High Return Visit Rate :Users return to the social network
community frequently. This opens up the opportunity for
great page impressions and huge advertising inventory.
b)User loyalty : Users connect to their friends and will not
abandon them easily.They do not move over to a competing
social network.Instead,they show high customer loyalty.
c)Virtual Growth:Members invite their friends to the social
network community.This is effective marketing at a low cost,and
the OSN grows by itself.
d)Business model: With a social network community you can
earn revenues through subscriptions to premium content in
addition to advertising revenues.
 5) Graph-Theoretic Analysis of Social Networks
• Social network analysis has emerged as a key technique in modern
sociology.
• It has also gained a lot of attention in anthropology, biology, communication,
economics, geography, information science, organizational studies, social
psychology, and sociolinguistics.
• Characterizing the existing relationship among a person’s social group is the
main task in social network analysis.
• Social relationships are often mapped into directed or undirected graphs,
sometimes called acquaintance graphs or simply social connection graphs.
• The nodes in a social graph correspond to the users or actors and the graph
edges or links refer to the ties or relationships among the nodes.
• The graphs can be complex and hierarchically structured to reflect
relationships at all levels. There can be many kinds of ties between the
nodes. Social networks operate from the family level up to national and
global levels.
• There are pros and cons of social networks. Most free societies welcome
social networks. For political or religious reasons, some countries block the
use of social networks to prevent possible abuses.
6) Communities in Social Networks
Six specialized social network communities are pointing to different sectors of
our society. This list only offers a few examples. One can form many other
communities from online social networking.
• Industry communities -Special industrial workers or professionals are often
connected with one another. They share knowledge and work experience.
• Artist communities -These network communities are specifically composed
to enable artists, musicians, or celebrities to personalize and intensify their
contact with their existing fans as well as enabling contact among community
members.
• Sport communities- These are network communities for special interests
and activities of athletes and sport fans. People can find friends, celebrate
their passion, and exchange ideas.
• Health communities -These are dedicated to the needs of actors concerned
about health issues.
• Congresses and event communities -These are customized to support all
preparations necessary for congresses and events, as well as all processes
thereafter.
• Alumni communities- After completing their studies, alumni can find fellow
students, stay in touch, and foster friendships.
7) Social Network Application Domains
There are an infinite number of applications for
social networks. We classify those applications into
five specific domains.
- Business
- Education
- Government
- Medical and Healthcare
- Online Dating