The cloud computing reference model is an abstract model Layers of the Cloud:

Fractures of Cloud Programming:

that divides a cloud computing environment into abstracti on The Cloud Reference Model defines three main cloud service layers:
There's no single, universally accepted definition of "fractures" in
layers and cross-layer functions to characteriz e and 1. Infrastructure as a Service (IaaS): This is the lowest layer, providing the
cloud programming. It might refer to:
fundamental building blocks like physical servers, storage, and networking
standardiz e its functions. This reference model divides cloud • The distributed nature of cloud applications,
computing activities and functions into three cross-layer where components run on different servers and resources. Users have full control over the underlying infrastructure but
functions and five logical layers. manage it themselves.
need to communicate effectively.
Each of these layers describes different things that might be 2. Platform as a Service (PaaS): This layer sits on top of IaaS, offering a
• The potential for failure points in cloud
present in a cloud computing environment, such as computing platform for developing, deploying, and managing applications. PaaS
deployments, requiring careful design for fault
systems, networking , storage equipment, virtualizati on removes the need for managing the underlying infrastructure and focuses on
tolerance and redundancy.
software, security measures, control and management application development.
• The challenges of managing and debugging cloud 3. Software as a Service (SaaS): The highest layer delivers ready-to-use
software, and so forth. It also explains the connecti ons applications due to their distributed nature.
between these organizations. The five layers are the Physical software applications accessible over the internet. Users access and utilize
layer, virtual layer, control layer, service orchestrati on layer, the software functionality without managing the infrastructure or platform.
and service layer. Types of Clouds (Public, Private, Hybrid):
Cloud Computing reference model is divided into 3 major Cloud deployments can be categorized into three main types:
service models:
1. Software as a Service (SaaS) • Public Cloud: Public clouds are owned and operated by cloud service
2. Platform as a Service (PaaS) providers (CSPs) and offer services to the general public over the internet.
They are highly scalable and cost-effective, but users have less control over
3. Infrastructure as a Service (IaaS)
the underlying infrastructure.

• Private Cloud: Private clouds are dedicated to a single organization and

Data Center Design: offer greater control, security, and customization. However, they can be more

•
expensive to manage compared to public clouds.

•
Physical Infrastructure:
o Ensures efficient power delivery, cooling systems, and
Hybrid Cloud: Hybrid clouds combine aspects of both public and private
clouds. An organization might use a public cloud for non-critical workloads
physical security for housing IT equipment like servers, and a private cloud for sensitive data or applications requiring stricter control.
storage, and networking devices.
o Optimizes airflow and heat management to prevent
equipment overheating. Cloud Service Models:

• Logical Design:
Cloud service models define the type of services offered by cloud providers and the level of control and
responsibility allocated to users. Here are the three main models:
o Defines how IT resources are organized and 1. Infrastructure as a Service (IaaS):
interconnected within the data center. o What it offers: Provides the most basic layer, including virtual
o Considers factors like scalability, redundancy, and
machines (VMs), storage, networking resources, and firewalls.
performance optimization. o Control/Responsibility: Users have full control over the underlying
o Virtualization technologies play a crucial role in logical infrastructure, similar to managing their own data center. They are
responsible for patching operating systems, installing software, and
design, allowing for efficient resource sharing and
utilization. configuring security.
2. Platform as a Service (PaaS):
Interconnection Network:

• Connects various components within the data center. This includes

o What it offers: Delivers a platform for developing, deploying, and
managing applications. PaaS services typically include pre-configured
servers, storage devices, and network switches. environments, databases, development tools, and middleware.

• High bandwidth and low latency are critical for efficient data transfer o Control/Responsibility: Users focus on developing and deploying
between IT resources and minimizing communication delays. applications on the platform. They manage their applications but don't

o Latency refers to the time it takes for data to travel 3.

have control over the underlying infrastructure (OS, servers, etc.).
Software as a Service (SaaS):
between points in the network.
o What it offers: Provides ready-to-use software applications accessible
over the internet. Users access the software functionality through a
web browser or API without managing the infrastructure or platform.
Parallel and Distributed Programming Paradigms:
o Control/Responsibility: Users have the least control in this model.
• MapReduce: A programming paradigm for processing large datasets in They can typically configure settings within the application but don't
manage the underlying infrastructure or platform.
parallel across clusters of computers. It involves two key functions:
o Map: Processes data records in parallel, transforming
each record into an intermediate key-value pair.
Programming of Google App Engine:
o
•
Reduce: Aggregates intermediate key-value pairs based
on the key, producing the final result. Google App Engine (GAE) is a Platform-as-a-Service

•
(PaaS) offering from Google. It provides a managed
Hadoop: An open-source framework that implements the MapReduce environment for developing and deploying web applications
paradigm and offers distributed storage (HDFS) for large datasets. at scale.

• GAE supports several high-level languages like Python,

Java, and Go. Developers can focus on application logic
Architectural Design of Compute and Storage Clouds in cloud computing: without managing underlying infrastructure.
Layered Architecture:
Cloud computing follows a layered architectural approach, with each layer offering specific functionalities:
1. Infrastructure Layer (IaaS):
-- Foundation of the cloud, providing the physical infrastructure like servers, storage, and networking equipment.
-- Often virtualized using technologies like Hypervisors, allowing for efficient resource sharing and creation of virtual
machines (VMs). High-Level Languages for Cloud:
2. Platform Layer (PaaS):
-- Built on top of the IaaS layer, offering a platform for developing, deploying, and managing applications.
-- Provides pre-configured environments, databases, development tools, and middleware to streamline application • Cloud platforms often provide SDKs (Software
development. Development Kits) with high-level languages like
3. Software Layer (SaaS): Python, Java, or .NET that simplify cloud
-- Topmost layer, delivering ready-to-use software applications accessible over the internet. application development. These SDKs offer
-- Users access the software functionality through a web browser or API without managing the underlying infrastructure or abstractions for interacting with cloud services like
platform. storage, databases, and compute resources.
Compute Cloud Design:

• Resource Pooling: Multiple physical servers are pooled and virtualized to create a pool of compute resources.

• Virtualization: Hypervisors enable creation of virtual machines (VMs) that share the underlying physical resources. VMs offer
isolation and flexibility for running diverse applications.

• Dynamic Provisioning: Cloud providers can dynamically allocate and release compute resources (VMs) based on user demand,
improving resource utilization and cost efficiency.

• Fault Tolerance: Mechanisms like VM migration and redundancy ensure service availability even if individual physical servers fail.
Storage Cloud Design:

• Scalable Storage: Cloud storage offers scalable storage solutions based on user needs. Users can easily increase or decrease
storage capacity as required.

• Data Replication: Data is replicated across geographically distributed storage systems for redundancy and disaster recovery
purposes.

• Data Security: Encryption and access control mechanisms ensure data security and privacy within the cloud.

• Storage Tiers: Cloud storage can be categorized into different tiers based on performance and cost. For example, frequently
accessed data might be stored on high-performance SSDs, while less frequently accessed data could reside on lower-cost HDDs.
Benefits of Layered Architecture:

• Flexibility: Users can choose the service level (IaaS, PaaS, SaaS) that aligns with their needs.

• Scalability: Cloud resources can be easily scaled up or down to meet changing demands.

• Cost-efficiency: Users typically pay only for the resources they use.

CLOUD