UNIT -3

1. Analytical Services
The cloud based analytical services allow to analyse the large amount of data stored in cloud
database or Cloud storage using a programming model called MapReduce. This service helps
to perform various task such as data mining, log file analysis, machine learning and web
indexing etc.

The following are various Cloud providing Analytical services.

Amazon Elastic MapReduce (EMR):
It is a service provided by Amazon based on Hadoop Framework. The EMR support various
Job types.
 Custom JAR: It is a job that enable to run a Java program uploaded in Amazon S3
 Hive Program: It is a job that provide a SQL like Interface called Hive script for
querying and analysing large set of data.
 Streaming Job: It is a job that the enable to analyse the data using MapReduce. It uses
Java Programming. Instead other programming language such as python, Ruby, R,
Pearl, Bash or C++ can be used.
 Pig Program: It is a program written in Pig Latin launched Pig Platform. Here pig
script are used to analyse large data set.
 HBase: It is distributed, scalable, NoSQL database built on top of Hadoop Distributed
File System (HDFS). It provide fast and random access to the data set.
Refer Book for Diagram
To create a MapReduce job, one should enter the Job name, streaming option, location of
input, output, mapper, reducer program and number of nodes to be used in Hadoop cluster.
Google MapReduce Service:
 It is a feature of Google App Engine that helps developers build and run web applications.
The service can be used through the MapReduce API.
A MapReduce pipeline is like a program flow where you define:
 Mapper → processes input data and breaks it into smaller pieces.
 Reducer → combines the processed pieces into a final result.
 Input reader → reads the data to be processed.
 Output writer → stores the final results.

Google Big Query:

It is a service for analysing massive datasets using Big Query console. Data can be either in
CSV or JSON format.
 Window Azure HDInsight:
It is a service from Microsoft. It use windows Azure Blob storage as a default file system. It
support both Java Script and Hive.

2. Deployment and Management services

This service help to easily deploy and manage applications in the cloud. It also helps to
manage task such as capacity provisioning, load balancing, auto scaling and application
health monitoring.
Amazon Elastic Beanstalk
The elastic Beanstalk is a service provided by Amazon that help to deploy and manage
applications in AWS. It supports languages like Java, PHP, .NET, Node Js, Python and Ruby.
To access Elastic beanstalk one can just upload the application and specify the configuration
settings and the service automatically handle all deployment task. The launch wizard we need
to specify the URL, application file, container ad instance type. If the load increases or
decreases the auto scaling launces new application to handle it.
Amazon Cloud Formation

It is a service provided by Amazon. Amazon Cloud Formation is an Infrastructure as Code

(IaC) service. It is a tool from AWS that helps you set up cloud resources automatically.
Instead of manually configuring everything, you write instructions in a file called a template.
This template tells Cloud Formation what to build, such as servers, databases, or networks. It
then creates all these resources together as one unit called a stack. The main benefit is that
you can easily create the same setup multiple times, for example, in testing, staging, or
production environments. If something goes wrong during the setup, Cloud Formation
automatically rolls back the changes so nothing is left incomplete.
3. Identity and Access Management services (IDAM)
Identity and Access Management (IAM) services are tools that control who can use your
system and what they are allowed to do. It is mainly used in organization where multiple user
who access to cloud resources.
Amazon Identity and Access Management (IAM)
Amazon IAM is a security service in AWS that helps you control who can access your AWS
resources and what actions they can perform. With IAM, you can create users (people),
groups (teams), and roles (temporary permissions for apps or services), then assign them
policies that define what they are allowed or not allowed to do.
Refer Book for Diagram
Window Azure Active Directory
Azure Active Directory (Azure AD) is Microsoft’s cloud-based identity and access
management service. It helps organizations control who can sign in (identity) and what they
can access (resources like apps, files, or services).
Refer Book for Diagram
 4. Open Source Private Cloud Software
Open-source private cloud software is software that lets organizations build and manage their
own private cloud environment using open-source technologies.
Cloud stack
Apache CloudStack is an open-source software for building private clouds. It has two parts:
the Management Server, which controls everything and the Cloud Infrastructure that controls
servers, storage, and network. A cloud can be as small as one server or as large as hundreds.
CloudStack organizes resources as Zones called as data centres, which contain Pods which
are racks of hardware. Pods have Clusters the groups of servers/hosts with primary storage.
Each Host runs virtual machines. Primary Storage stores VM data, while the Secondary
Storage stores templates, ISOs, and backups.

Eucalyptus
Eucalyptus is an open-source software platform used to build private and hybrid clouds that
are compatible with Amazon Web Services (AWS) APIs. Its architecture is divided into
different components. At the lower level, the Node Controller (NC) runs on physical servers
to launch, manage, and monitor virtual machines and their networking.
At the cluster level, there are three main components. The Cluster Controller (CC), which
acts as the front-end for the cluster and manages the virtual machines. The Storage Controller
(SC) manages block storage volumes and snapshots for virtual machines, similar to AWS
EBS and the optional VMware Broker, which provides an AWS-compatible interface for
VMware-based environments. At the cloud level, the Cloud Controller (CLC) manages
scheduling, authentication, quotas, and accounting, while Walrus provides storage for VM
images, snapshots, and data, similar to AWS S3.
Open Stack
OpenStack is an open-source cloud system that manages computing, storage, and
networking through different services: Nova provides virtual servers, Neutron handles
networking, Cinder manages storage volumes, Swift stores files, Keystone controls
authentication, Glance stores VM images, the scheduler decides VM placement, RabbitMQ
manages communication, nova-api processes requests, and Horizon gives a web interface.

5. Cloud Application Design:

Introduction:
Due to the Dynamic nature of modern Web Applications as it allows the user to interact,
collaborate, includes user generates content such as comment, discussion, integrated social
networks and other multimodal content like text, image, video and presentation etc. Hence it
is necessary for consideration of cloud applications using various types of application and
design methodologies such as SOA(Service Oriented Architecture), CCM (Cloud Computing
Model) and MVC (Model View Controller).
Design Consideration for Cloud Applications:
The following are the parameter should be consider while designing a Cloud Applications.
Scalability:
It helps to build applications that can serve millions of user without getting delay on their
performance. The other consideration to increase the scalability of a cloud applications are
 Loose Coupling of Components:
Traditional method use Tight Coupling in which hard wired links make difficult to scale
application component independently. But by designing loosely coupled components, it is
possible to scale each component independently.
 Asynchronous Communication:
By allowing Asynchronous Communication between components, it is possible to
increase capacity by adding additional servers, when application load increases.
 Stateless Design:
This design enable to store the state outside of the component in a separate database allow
scaling the application components independently.
 Database Choice Design:
Decision such as to choose a traditional database (SQL) with strict schema or a non-
traditional database (No- SQL) with schema less approach will affect the application
scalability.
Reliability and Availability
The important consideration while developing highly reliable and available applications are
 No Single Point of failure:
No single point of failure means your app is designed with redundancy, load balancing,
and failover, so one failure won’t take everything down.
 Triggered automated actions on failures:
If an applications server experiences high CPU usage and it is unable to serve new
request, a new server is launched automatically to improve application reliability and
availability.
 Graceful Degradation:
If some components of an application is unavailable, it continue to serve the user with
limited functionality. Eg: Ecommerce websites.
Logging:
 Logging all events improves reliability by recording errors and performance issues for
quick diagnosis. It also supports availability by enabling real-time monitoring, alerts, and
faster recovery during failures.
 Replication:
It enable to create and maintain multiple copied of data in cloud. Hence any data loss at
primary location, organization can continue to operate their application from secondary
data source.
Security
The security of an applications must be ensured in domain such as health care etc. The
security aspects of cloud applications are
 Securing Data at rest
 Securing Data in motion
 Authentication
 Authorization
 Identity and Access management
 Key Management
 Data Integrity
 Auditing
Maintenance and Upgradation
It enable in keeping systems healthy and current. Business applications are launched with
different features. The applications must be adapted with user feedback to main the user.
Loosely coupled components will help in reducing application maintenance and upgradation.
Logging, triggering automated actions also lower the maintenance cost.
Performance
Performance should be kept in mind while designing applications. There are various matrices
used to evaluate the performance such as response time (less than 4 sec) and throughput.