Cloud Computing

Lecture 2
7-8-2024
Mr. Ajay B. Kapase
 Defining Cloud Computing
• General term for the delivery of hosted computing resources and IT
services over the internet with Pay as you Go pricing.
• Computing Resources
• Compute- Processing Power (Processors)
• Memory- RAM
• Storage – Secondary Storage Space
• IT Services
• Database Service
• Development tools
• Storage Service
• Other Applications (softwares)
• Network Protocols
Historical Developments in Computing
Distributed Computing (1950)
• Distributed computing is the method of making multiple computers
work together to solve a common problem.
• System where processing and data storage is distributed across
multiple devices or systems, rather than being handled by a single
central device.
• Each device has its own compute (processing power) and storage.
• Work together to perform single task distributing among them and
share resources.
Distributed Computing
• Each device has its own local memory
and processing power.

• Communication across units is done using

message passing concept (Over
Network).

• Communication within individual unit is

done via bus.

• Decentralized management
 Advantages of Distributed Computing
• Robustness:
• Work well in Node failure.
• No single point of failure.
• can continue to operate even if one device or system fails.
• Performance
• Processing and Memory is shared => Large data / Large Computing Operations
• Load balancing.
• Scalability:
• can easily add new devices or systems to the network to increase processing and
storage capacity.
 Disadvantages of Distributed Computing
• Computational Speed is slow as compared to monolithic system.

• In comparison to a monolithic, the database associated with distributed

systems is much more complex and challenging to manage.
Mainframe Computing (1951)
• Mainframe: highly powerful, large and reliable computing machine
typically used for complex calculations and data processing tasks.

• Responsible for handling large data such as massive io operations.

• It can connect to multiple end clients simultaneously.

• Several users can access different applications and runs processes

concurrently without impacting performance or security.
Mainframe Computing
• Stands between Servers and
Supercomputers.

• Advanced than servers but

lower than supercomputers.

• Usually used for massive

transactional computing and
I/O operations.
 Advantages of Mainframe Computing
• The processing speed is very fast.

• Can handle multiple inputs/clients at same time.

• Can withstand failure of a part without affecting the function of rest of

the computer.
Disadvantages of Mainframe Computing
• It requires a special OS to run.

• Are very expensive.

Cluster Computing (1980)
• Cluster, is a group of interconnected computers forming a single
system together with the components cooperating in providing high
availability, load balancing, and parallel processing.

• Units, or nodes are usually in the same physical space and are
interoperable.
Cluster Computing
• Consists multiple cooperating
homogeneous nodes giving impression
of single standalone system.

• Consists one head node responsible for

distributing of work, load balancing,
combining results.

• Centralized management, nodes usually

in the same location
Advantages of Cluster Computing
• High Performance
• Better and enhanced performance than that of mainframe computer
networks.
• Easy to manage
• Centralized management.
• Scalable
• Add new nodes as per requirement.
• Robustness
• Work even if any node fails.
Disadvantage of Cluster Computing
• High cost:
• not so much cost-effective due to its high hardware requirements and
complex design.

• Need of Homogeneous Nodes.

 Feature Distributed Systems Clustered Systems

Centralized management, nodes usually in the same

Architecture Decentralized, nodes can be geographically dispersed location

Load Balancing Varies, depends on implementation Integral feature, evenly distributes tasks

Management
No single point of control, often requires complex Centralized control, simpler management
management

Geographical Distribution Can span across different geographic locations Typically confined to a single physical location

Resource Sharing Independent resources across nodes Shared resources within the cluster

Communication Often relies on message passing and network protocols Generally, uses high-speed local networks

Latency Can have higher latency due to network communication Generally low latency due to local network

System Administration More complex due to the distributed nature of the nodes Simplified due to centralized management
Grid Commuting (1990)
• Different systems placed at entirely different geographical locations
and these all are connected via the internet.
• Grid consisted of heterogeneous nodes (Different OS and Hardware)
• For controlling the network and its resources a software/networking
protocol is used called as middleware.
• Nodes are heterogeneous and geographically spread at different
locations like distributed computing.
• But controlled by a single control node over internet just like cluster
computing.
Grid Computing • Consists 3 types of Nodes

• Control Node: Server / group of server

which administrates the grid nodes.

• Provider(Grid) Node: The nodes which

provides/contributes its resources in
grid network.

• User Node: The node which uses the

resources in the grid network
Advantages of Grid Computing
• Centralized Management
• Proper utilization of resources in the grid due to control node.
• Scalable
• Add new nodes as per requirement.
• Robust
• Work even if any node fails.
• Flexible
• Can work with heterogeneous nodes.
Disadvantage of Grid Computing
• High Complexity:
• Design and implementation of grid system is complex and requires
skills.

• Dependence of Network Speed:

• Since nodes in grid located in geographically different locations they need to
have high speed interconnect among them.
Virtualization
• Virtualization uses software to create an abstraction layer over computer
hardware.

• Enables division of a single computer's hardware components—such as

processors, memory and storage—into multiple virtual machines (VMs).

• Abstraction Layer is called as Hypervisor.

• Each VM runs its own operating system (OS) and behaves like an independent
computer, even though it is running on a single hardware platform.
Virtualization Scenario
Web 2.0
• Web 1.0 made up of static pages connected to a system via hyperlinks.

• Web 1.0 content comes from the server's file system, not a relational
database management system.

• Web 2.0 contains dynamic content that responds to the user’s input.

• Web 2.0 offers free information sorting, allowing users to retrieve and
classify data collectively with use of RDBMS
Web 2.0
• Allows user interactions with the online content like
• Podcasting
• Social media
• Tagging
• Blogging
• Commenting
• Curating with RSS
• Social networking
• Web content voting
Service Orientation
• Service-oriented architecture (SOA) is a software development model
that makes services reusable.

• Rather than developing and hosting software -> develop and host
service.

• These services can be reused in order to develop a software

• Integration of services in software can be done via API

Service Orientation (Example)
Food Delivery App Case 1
Coding from Scratch

Login Hotel Listing

Payment Order Tracking
Service Orientation (Example)
Food Delivery App Case 2

Authentication Coding from Payment Gateway Maps Service

Service Scratch Service

Login Hotel Listing Payment Order Tracking

Service Oriented Architecture
Building blocks of SOA
• Service:
• Actual functionality hosted on the internet.
• Example: Google Maps

• Service Provider
• Entity which is responsible for developing and hosting specific service.
• Example: Maps service provided by Google

• Service Consumer
• Entity which uses the service or embed the service in the software.
• Example: Swiggy uses maps service by google for tracking orders.
Utility Computing
• Utility computing allows users to scale up and down based on their
needs.

• Clients, users, or businesses can rent data storage space, computing

power, or even hardware such as CPUs, monitors, etc.

• Works on model Pay as per usage / Pay as you go