You are on page 1of 59

FOG

COMPUTING
B A S E D O N PA P E R S BY
I VA N A N D F L AV I O

BIRTH OF THE CLOUD

Virtualization Technologies
High Bandwidth Availability Improvements in
Physical Layer
Universal Software Interoperability Standards

A BRIEF HISTORY OF CLOUD


99 Salesforce delivers enterprise applications
through a website.
2002 AWS launches. on demand workforce
through mechanical turk.
2006 AWS offers small companies cloud
computing ahead of Microsoft.
2009 Google offers Google Apps.

WHY IS CLOUD COMPUTING POPUL AR?


Economy at scale.
Pay as you go.
Globalize workforce.
Accessibility and flexibility.
Minimize licensing.

INTERNET OF THINGS

INTERNET OF THINGS

Desire to connect all devices.


Increase machine to machine communication.
Integrating sensors to the network.
Smart appliances, smart homes, smart
vehicles etc.

DEPLOYMENT PROBLEMS
Lack of shared infrastructure.
Lack of standards.
A missing piece of technology to help
deploy IoT.

THE MISSING PIECE

FOG COMPUTING THE MISSING PIECE.


Extending the cloud to the edge of the
network.
Not a trivial extension!

WHAT IS THE FOG?


Computing capabilities where they
matter.
Take analytics, processing and storage to
the edge of the network.
Impractical to send all data from all
devices to the cloud for processing.

DEFINING CHARACTERISTICS OF THE FOG


Location Awareness
Low latency.
Geographical distribution. Not centralized like
the cloud!
Can leverage large scale sensor networks.
Large, large number of nodes.

DEFINING CHARACTERISTICS OF THE FOG


Support for mobility.
Real time performance, analytics and
interactions.
Predominantly wireless access.
Seamless operation. Example: streaming.

Requires interoperability.

Federation.

INTERPL AY OF FOG AND CLOUD COMPUTING


Analytics - center of our applications.
We need fog for context aware, local analytics.
We need the cloud for centralized, heavy duty,
global analytics.
Protection and control information requires real
time processing.
Several tiers of hierarchical filtering of data.

INTERPL AY OF FOG AND CLOUD COMPUTING


Highest tier involves human interaction- like
visualization.
From seconds in lowest tier to days in highest tier.

SUMMARIZING CLOUD AND FOG COMPUTING


Cloud

Fog

Data and applications


processed in the cloud.

Processing takes place at


the edge of the network.

Bandwidth limitation, since


data is sent through cloud
channels.

Bandwidth not a limitation.

Centralized

Distributed

Slow response(high latency) Low latency applications.


Scalability issues.

Scalable as the number of


nodes increases.

APPLICATIONS OF FOG COMPUTING


Smart Grids.
Smart Traffic Lights - Ambulance Sensors for
example.
Connected Vehicles.
Wireless sensors and actuator networks.
Smart Building.
IoT.
SDN(Software Defined Networking)

IOT AND CYBER PHYSICAL SYSTEMS


Tight coupling of engineered systems and physical
reality.

Embedded Systems with networking capabilities.

IOT AND CYBER PHYSICAL SYSTEMS


Physical systems are noisy, dynamic and
uncertain.
Software components are precise.
Can we give intelligence to these physical
systems?

Or derive intelligence from it?

IOT AND CYBER PHYSICAL SYSTEMS


Examples:

Intelligent medical devices.

Smart highways.

Smart buildings.

Smart factories.

Smart agriculture.

Robotics.

SOFTWARE DEFINED NETWORKING


Traditional networks are:

Stagnant and difficult to perform experiments on.

Closed systems that are vendor specific.

Meaningful vendor collaboration is difficult.

HUGE barrier for ideas in networking.

SDN is:

A software architecture that decouples the control and data


plane.

Programmatic interface into network equipment.

Centralized controller can operate an entire network.

SOFTWARE DEFINED NETWORKING

SOFTWARE DEFINED NETWORKING


What does SDN with Fog Computing Provide?
Specifically solve vehicular communication
problems:

Intermittent connectivity.

Collisions.

High packet loss rate.

How?

Better vehicle to vehicle communication.

Better vehicle to infrastructure communication.

Fog Computing
Applications

Implemented at Network Edge


Low Latency
Location Awareness
Improved QOS
Supports Heterogeneity
End-User devices
Access Points
Edge Routers and Switches

SMART GRID
Load Balancing Devices
Help to switch to alternative energy
sources
Fog Collectors
Process data sent by sensors
Filter data locally and send rest
to higher Tiers
Fog supports ephemeral storage

SMART TRAFFIC LIGHTS


Video Cameras can automatically
change lights depending on situations
Interact locally with sensors to detect
traffic
Measure Speed and distance of
vehicles
Send warning signals to
approaching vehicles

CCONNECTED VEHICLES
Enable Real-Time interaction
Cars
Access Points
Traffic Lights
FOG Clusters analyze data locally to reroute traffic
and maintain flow

WIRELESS SENSOR &


ACTUATOR NETWORKS
Traditional sensor networks need actuators to exert
physical actions
Fog Devices can control measurement by creating
closed-loop
Sensors can monitor heat levels on trains ball
bearing, to stop the train in case of emergencies
Sensors on air vents regulate the flow of air

DECENTRALIZED SMART BUILDING


CONTROL
Wireless sensors deployed in buildings to measure
temperature and humidity levels
Sensors combined to form better measurements
Increase or decrease the temperature depending
on reading
Can be used to conserve energy, water and other
resources

WIND FARMS
Fog can be used to regulate the windmills based on
weather data
Better control of turbines
Based on Elevation
Topography of the terrain
Can be used to collect and use long term data

CCONTENT DELIVERY AND CACHING


Traditional Web content are not optimized for
user-side requests.
Fog enables dyanamic customizable optimization.
Using caching techniques latency is further
reduced and can result in bandwidth usage
reduction

HEALTHCARE AND OIL/GAS


Healthcare
Patient monitoring system in Real Time in
critical care units
Reduction in latency might result in saving
lives
Oil & Gas
Pipeline monitoring for leaks, fire, theft etc.

AGRICULTURE & RETAIL


Agriculture
Smart forms with crop monitoring and irrigation
control systems.
Retail
Tracking of shopping carts and automatic billing
systems.
Results in time saving measures

FOG COMPUTING
APPLICATIONS

MOBILE FOG: A PROGRAMMING MODEL FOR


LARGESCALE APPLICATIONS ON THE INTERNET
OF THINGS

As a High level programming model for geospatially distributed, large-scale and latency
sensitive Internet applications.
Mobile Fog consists of a set of event handles
and functions that an application can call.
Not a generic model but built for particular
application, while leaving out functions that
deal with technical challenges of involved
image processing primitives.

MOBILE FOG: A PROGRAMMING MODEL FOR


LARGESCALE APPLICATIONS ON THE INTERNET
OF THINGS

Vehicle tracking using Cameras


Camera processes are the leaves of the
tree and are responsible for sensing the
environment and delivering parent
processes.

Traffic monitoring using MCEP

MOBILE FOG: A PROGRAMMING MODEL FOR


LARGESCALE APPLICATIONS ON THE INTERNET
OF THINGS

MOBILE FOG: A PROGRAMMING MODEL FOR


LARGESCALE APPLICATIONS ON THE INTERNET
OF THINGS

MIGCEP: OPERATOR MIGRATION FOR


MOBILITY DRIVEN DISTRIBUTED COMPLEX
EVENT PROCESSING
Increasing deployment of powerful mobile
sensors and large scale sensor networks. E.g.
Smartphones, CCTV
Complex Event Processing (CEP) is a key
paradigm to realize such applications.
Mobile CEP, consumers and sensors are
mobile.
Cost associated with each migration.
Propagation of state across the network is
expensive.

MIGCEP: OPERATOR MIGRATION FOR


MOBILITY DRIVEN DISTRIBUTED COMPLEX
EVENT PROCESSING

Exploit application knowledge of the


MCEP system and predicted mobility
patterns to plan the migration ahead of
time.
Amortize migration costs.
Reduces Bandwidth.

Ensures application-defined end-to-end


latency restrictions.

IMPROVING WEB SITES PERFORMANCE USING


EDGE SERVERS IN FOG COMPUTING
ARCHITECTURE
Users are connected to Internet via edge servers. All
web requests that the user makes first goes through
the edge servers.
Fog server can optimize the incoming portions of the
webpage based on portions that have been
examined already.
Fog server has the distinct advantage of knowing
the network conditions local to an end user.
E.g. The type of device, Congestion in network

Per User Optimization for Inline or External Scripts


Advantage of an inline JS or CSS is that HTTP
requests are minimized.
Fog server can observe each user based on their
MAC addresses or local IP addresses and keep track
of each users website requests.

SERVICE- ORIENTED HETEROGENEOUS RESOURCE


SHARING FOR OPTIMIZING SERVICE LATENCY IN
MOBILE CLOUD

Pervasive mobile devices share their


heterogenous resources and support
services.
Neighboring nodes in a local network
form a group called a local Cloud.
A local resource coordinator (LRC)
serving as Fog device is elected from
nodes in each local Cloud.

SECURITY AND PRIVACY


CONCERNS

SECURITY ISSUES
Main security issues are authentication at
different levels of gateways as well as devices
installed at the consumers end.
Each smart meter and smart appliance has an
IP address. A malicious user can either tamper
with its own smart meter, report false
readings, or spoof IP addresses.

SECURITY ISSUES SOLUTIONS


Solution for Authentication
Public Key Infrastructure (PKI) based
solutions which involve multicast
authentication.
Diffie Hellman key exchange based
authentication techniques.
Intrusion Detection Techniques can also be
applied in Fog computing.
Intrusion can be captured by using an anomalybased method where an observed behavior is
compared with expected behavior to check if
there is a deviation.

MAN-IN-THE-MIDDLE ATTACK
A man-in-the-middle attack is one in which
the attacker secretly intercepts and relays
messages between two parties who believe
they are communicating directly with each
other.
MITM attacks pose a serious threat to online
security because they give the attacker the
ability to capture and manipulate sensitive
information in real-time while posing as a
trusted
party
during
transactions,
conversations, and the transfer of data.

MITM IN THE CONTEXT OF FOG


Man-in-the-middle
attack has potential to
become
a
typical
attack
in
Fog
computing.
Gateways serving as
Fog devices may be
compromised
or
replaced by fake ones.
E.g. KFC or Star Bar
customers
connecting
to
malicious
access
points
which

STEALTH TEST
MITM only consume a small amount of resources in Fog
devices, such as negligible CPU utilization and memory
consumption.

In this scenario, a 3G user sends a video call to a WLAN


user. Since the man-in-the-middle attack requires to
control the communication between the 3G user and
the WLAN user, the key of this attack is to compromise
the gateway which serves as the Fog device.

STEALTH TEST
In order to hijack and
replay victims video
communication,
a
hook
program
is
inserted
into
the
TCP/IP stack of the
compromised system.
Hook is a technique of
inserting code into a
system call in order to
alter it.

WORK FLOW OF MITM


Communication between 3G and WLAN needs a
gateway to translate the data of different protocols
into the suitable formats.
MITM is divided into four steps
Hook process redirects data from 3G user to attacker.
Attacker replays or modifies the data of the
communication locally.
Attacker sends the data back to gateway.
Gateway forwards the data from the attacker to the
WLAN user.

The attacker can monitor and modify the data sent


from the 3G user to the WLAN user in the middle of
the communication.

RESULT OF STEALTH TEST


Memory consumption
and the CPU utilization
of gateway during the
attack, is measured
and compared to
normal utilization for
anomaly detection.
If MITM does not
greatly change the
features of the
communication it can
be proofed to be a
stealthy attack.

MITM & FOG


MITM is simple to launch but difficult
to be addressed.
Encryption may also not protect users
as attackers can set up a legitimate
terminal and replay the
communication without decryption.
Complex encryption and decryption
not suitable for all scenarios.

PRIVACY ISSUES

EPPA
An Efficient and Privacy Preserving Aggregation
Scheme for secure smart grid communications
Super increasing sequence to structure multidimensional data and encrypt the structured data
by the homomorphic cryptogram technique.
Ensures privacy of data collected by smart meters
but does not guarantee that the device transmits
the correct report to other gateways.
Data communications from user to smart grid
operation center, data aggregation is performed
directly on cipher-text at local gateways without
decryption.

REFERENCES

THANK YOU FOR YOUR ATTENTION

You might also like