Cloud Computing

Module 1
 Issues/Challenges
• The distance between the cloud and the data
• Network Bandwidth latency and cost
• Response time
• Processing speed
• Operational cost
• Security
 Layers of Computing

The terms “cloud,” “edge,” and “fog” represent three layers of

computing:
• Cloud Layer: Industrial big data, business logic and analytics
databases and data “warehousing”
• Fog Layer: Local network assets, micro-data centers
• Edge Layer: Real-time data processing on industrial PCs,
process-specific applications and autonomous equipment
✔ Edge computing allows processing to be performed locally at multiple
decision points for the purpose of reducing network traffic.
Contd…
 Cloud Computing
• Cloud Computing ≈ Grid Computing + Utility
Computing
• Outsource utilities (storage, computation etc.) to a
third party, called Cloud Service Provider
• Managed by the service provider
• Illusion of unlimited resources
• Pay-as-you-go model
• Accessible through a network
 Evolution of Applications

Stand Alone Web Apps Cloud Apps

• Resides on • Resides on • Multi-tenancy

local system remote system • Elasticity
• Local • Client - Server • Heterogeneity
resources Model • Measured use
• Self • Network • On-demand
Sustaining Dependent • Network
• Not shareable • QoS depends dependent
• Prohibitive on number of
costs users
• Frequent • Inflexible
updates usage model
A Simple Architecture
Cloud Computing
 Is Cloud Computing the Best Choice
Always?

CLOUD
SERVER

The car in front of me is

braking! What should I
do?

Consider the scenario of driverless cars

Is Cloud Computing the Best Choice
Always?

BRAKE!!! CLOUD
Sorry for the delay ☹ SERVER

Consider the scenario of driverless cars

Problem: Self Driving Cars
• In real time systems,
would there be enough
time for data to be
processed by a cloud
server?
• In such cases, wouldn’t
it be better if data was
processed closer to the
source?
Edge Computing
 Features of Edge Computing
• Advantages
– Low Latency
– Faster Decisions
– Privacy
– Edge computing also improves security by encrypting data
closer to the network core
• Disadvantages
– Edge devices will have lower computational power
– Lack of a global (or network level) view
 Automated Museum Tour Guides

More and more museums are now relying on automated robots to replace
museum tour guides. For example, the Smithsonian Museum launched a
robot named “Pepper” in 2018
 An Issue with Edge Computing
• Consider the case of the automated museum
tour guides:
– All information that the robot needs is localized –
need not use the “cloud” realistically
– Interaction with tourists – need for local
processing power
– What if some exhibits are closed? Or the order
changes?
– There is a need for a broader, museum-level view
for efficient functioning
Fog Computing
 Fog Computing
• Fog and edge computing are both extensions of cloud
networks.
• Move computation to a location between the nodes and the
cloud.
• Mobile users have predictable service demands subject to their
locations.
• Fog Servers can periodically connect with cloud servers and
cache location specific information.
• Information from edge devices can be processed by fog servers
to provide an interactive experience.
 Features of Fog Computing
• Contextual location awareness + low latency
• Geographical distribution
• Heterogeneity
• Interoperability
• Real-time interactions
• Scalability
 How do Fog and Edge Computing Work?
• Fog computing is useful when the Internet connection isn’t
always stable.
• Fog computing allows to implement data processing at the
local networks, especially if it has to be processed in real
time.
• On the cloud, data is distributed to dozens of servers, whereas
edge computing uses hundreds, possibly thousands of local
nodes.
• Each device can act as a server in the edge network. To break
into, hackers would need access to thousands of distributed
devices, which is practically impossible.
 Benefits of Fog Computing
• Low latency. The fog network can process large volumes of
data with little-to-no delay. Because a lot of data is stored
locally, the computing is performed faster.
• Better data control. In cloud computing, third-party servers
are fully disconnected from local networks, leaving little to no
control over data. In fog computing, users can manage a lot of
information locally and rely on their security measures.
• A flexible storage system. Fog computing doesn’t require
constant online access. The data can be stored locally or pulled
up from local drives — such storage combines online and
offline access.
• Connecting centralized and decentralized storage. Fog
computing builds a bridge between local drives and third-party
cloud services, allowing a smooth transition to fully
decentralized data storage.
 Benefits of Edge Computing
• No delays in data processing. The data stays on the “edges”
of the IoT network and can be acted on immediately.
• Real-time data analysis. Works great when large amounts of
data have to be processed and immediately.
• Low network traffic. The data is first processed locally, and
only then sent to the main storage.
• Reduced operating costs. Data management takes less time
and computing power because the operation has a single
destination, instead of circling from the center to local drives.

IoT edge computing is an optimal solution for small immediate

operations that have to be processed with millisecond rates. When many
small operations are happening simultaneously, performing them locally
is faster and cheaper.
Hierarchy of Edge, Fog and Cloud
 Role of IoT and Cloud Computing
• Increased scalability. IoT systems produce and exchange a
lot of data and require a lot of storage space to seamlessly
function. IoT services on Cloud platforms like Microsoft Azure
IoT Suite, Google Cloud’s IoT service, and IBM’s IoT platform
provide access to powerful cloud services able to handle all this
IoT data.
• Improved safety. IoT services should rely on safe data storage
able to prevent hackers from trying to access and jeopardize the
system. Decentralized storage represents one of the ways to
secure sensitive resources.
• Enable IoT. Decentralized data storage corresponds with the
main IoT needs, such as accessibility, safety, mobility, and
scalability. Cloud services process massive amounts of data and
quickly distribute the information among multiple servers,
where it can be later accessed by a connected device anytime.
 The future with IoT, Edge, Fog and
Cloud Computing?
• IoT sector is rapidly growing, and we are likely heading into a
future where every device is connected.
• With smart homes, cars, equipment, and everything else, vast
masses of data are generated every second.
• Devices will continue to require increases in computer power,
and cloud computing offers decentralized storage solutions for
faster and cheaper deployments.
• Developers only have to connect their systems to IoT cloud
platform existing infrastructure to benefit from third-party
computing power.
 Benefits
1. Smart analytics: Internet of Things systems produce a lot of data.
Developers and businesses can use it to know their users better.
Cloud services provide a safe environment where this data could
be analyzed, managed, and stored.
2. Better security: A security breach in IoT networks could mean
compromising entire businesses and industries, affecting millions
of connected devices and people who use them. Cloud storages are
more difficult to target because of their remote position and
security practices.
3. Inter-device interactions: The cloud improves communication
between devices and applications, quickly sending data between
data centers and local nodes. Fog and edge computing can be
useful for offline communication and micro-operations, reducing
operation costs and increasing speed.
 Mist Computing
• Mist computing is a lightweight and rudimentary
form of fog computing that resides directly within
the network fabric at the edge of the network
fabric
• Uses microcomputers and microcontrollers to feed
into fog computing nodes and cloud computing
services.

[Widely varying definitions of fog, mist and edge computing can

be seen in literature. The definitions used here are adopted from
NIST specifications]
Architecture of Mist Computing
 Dew Computing
• What happens to cloud computing when you lose
network access?
• Dew Computing is a framework wherein a version
of the cloud architecture is replicated on a local
system – immune to loss of connectivity
• Applicable to systems with direct human
interaction, such as laptops, desktops and mobile
devices
• Eg: Dropbox
Dew vs Cloud Computing