Cloud Computing

SEWP ZG527

BITS Pilani
Motivation

Powerful
multi-core 1. Web Scale
processors Problems
General
Explosion of
domain
purpose 2. Web 2.0 and
graphic
applications Social
processors
Networking
Superior
Proliferation 3. Information
software
of devices
methodologies
Explosion
Virtualization 4. Mobile Web
Wider bandwidth leveraging the
for communication powerful
hardware
BITS Pilani, Deemed to be University under Section 3 of UGC Act, 1956
Technology Advances
Cloud applications: data-intensive,
compute-intensive, storage-intensive

Bandwidth
WS
Services interface

Web-services, SOA, WS standards

VM0 VM1 VMn

Storage Virtualization: bare metal, hypervisor. …

Models: S3,
BigTable,
BlobStore, ... Multi-core architectures

64-bit
processor

BITS Pilani, Deemed to be University under Section 3 of UGC Act, 1956

What is Cloud Computing?

Cloud Computing is a general term used to describe a new class of

network based computing that takes place over the Internet,
 basically a step on from Utility Computing
 a collection/group of integrated and networked hardware,
software and Internet infrastructure (called a platform).
 Using the Internet for communication and transport provides
hardware, software and networking services to clients
These platforms hide the complexity and details of the underlying
infrastructure from users and applications by providing very simple
graphical interface or API (Applications Programming Interface).

BITS Pilani, Deemed to be University under Section 3 of UGC Act, 1956

What is Cloud Computing cont.…

In addition, the platform provides on demand services, that are always

on, anywhere, anytime and any place.
Pay for use and as needed, elastic
 scale up and down in capacity and functionalities
The hardware and software services are available to
 general public, enterprises, corporations and businesses markets

BITS Pilani, Deemed to be University under Section 3 of UGC Act, 1956

Drivers for the new Platform

http://blogs.technet.com/b/yungchou/archive/2011/03/03/chou-s-theories-of-cloud-computing-the-5-3-2-principle.aspx

BITS Pilani, Deemed to be University under Section 3 of UGC Act, 1956

Cloud Summary

• Shared pool of
configurable
computing
resources
• On-demand
network access
• Provisioned by
the Service
Provider

BITS Pilani, Deemed to be University under Section 3 of UGC Act, 1956

Cloud Computing: Definition

The US National Institute of Standards (NIST) defines

cloud computing as follows:
Cloud computing is a model for enabling ubiquitous,
convenient, on-demand network access to a shared pool
of configurable computing resources (e.g., networks,
servers, storage, applications, and services) that can be
rapidly provisioned and released with minimal
management effort or service provider interaction.
BITS Pilani, Deemed to be University under Section 3 of UGC Act, 1956
3-4-5 rule of Cloud Computing

NIST specifies 3-4-5 rule of Cloud Computing

3 cloud service models or service types for any cloud platform

4 deployment models
5 essential characteristics of cloud computing infrastructure

BITS Pilani, Deemed to be University under Section 3 of UGC Act, 1956

Characteristics of Cloud Computing

 On demand self-
service
 Broad network
access
 Resource
pooling
 Rapid elasticity
 Measured
service

BITS Pilani, Deemed to be University under Section 3 of UGC Act, 1956

4 Deployment Models

1. Public Cloud Mega-scale cloud infrastructure is

made available to the general public
or a large industry group and is
owned by an organization selling
cloud services.

BITS Pilani, Deemed to be University under Section 3 of UGC Act, 1956

4 Deployment Models

2. Private Cloud
The cloud infrastructure is operated
solely for an organization. It may be
managed by the organization or a
third party and may exist on
premise or off premise.

BITS Pilani, Deemed to be University under Section 3 of UGC Act, 1956

4 Deployment Models

3. Hybrid Cloud The cloud infrastructure is a

composition of two or more clouds
(private or public) that remain
unique entities but are bound
together by standardized or
proprietary technology that enables
data and application portability

BITS Pilani, Deemed to be University under Section 3 of UGC Act, 1956

4 Deployment Models

4. Community Cloud Community Clouds are when an

‘infrastructure is shared by several
organizations and supports a specific
community that has shared concerns
(e.g., mission, security requirements,
policy, and compliance considerations).
It may be managed by the organizations
or a third party and may exist on premise
or off premise’ according to NIST.
A community cloud is a cloud service
shared between multiple organizations
with a common tie/goal/objective.
E.g. OpenCirrus

BITS Pilani, Deemed to be University under Section 3 of UGC Act, 1956

3 Cloud Service Models
Software as a Platform as a Infrastructure as a
Service (SaaS) Service (PaaS) Service (IaaS)

SalesForce CRM

LotusLive

Google
App
Engine

BITS Pilani, Deemed to be University under Section 3 of UGC Act, 1956

Software as a Service (SaaS)

Software as a service features a complete

application
offered as a service on demand.
A single instance of the software runs on the cloud
and services multiple end users or client
organizations.
E.g. salesforce.com , Google Apps
BITS Pilani, Deemed to be University under Section 3 of UGC Act, 1956
Platform as a Service
Platform as a service encapsulates a layer of software and provides it as a
service that can be used to build higher-level services.
2 Perspectives for PaaS :-
1. Producer:- Someone producing PaaS might produce a platform by
integrating an OS, middleware, application software, and even a
development environment that is then provided to a customer as a service.
2. Consumer:-Someone using PaaS would see an encapsulated service that
is presented to them through an API. The customer interacts with the
platform through the API, and the platform does what is necessary to
manage and scale itself to provide a given level of service.
Virtual appliances can be classified as instances of PaaS.

BITS Pilani, Deemed to be University under Section 3 of UGC Act, 1956

Infrastructure as a Service

Infrastructure as a service delivers basic storage and

computing capabilities as standardized services over the
network.
Servers, storage systems, switches, routers , and other
systems are pooled and made available to handle
workloads that range from application components to
high-performance computing applications.

BITS Pilani, Deemed to be University under Section 3 of UGC Act, 1956

Cloud Infrastructures

Key Technology is Virtualization

App App App

App App App OS OS OS

Operating System Hypervisor

Hardware Hardware

Traditional Stack Virtualized Stack

Virtualization plays an important role as an enabling technology for

datacentre implementation by abstracting compute, network, and storage
service platforms from the underlying physical hardware
BITS Pilani, Deemed to be University under Section 3 of UGC Act, 1956
Cloud Providers Characteristics

• Provide on-demand provisioning of computational resources

• Use virtualization technologies to lease these resources
• Provide public and simple remote interfaces to manage those
resources
• Use a pay-as-you-go cost model, typically charging by the
hour
• Operate data centers large enough to provide a seemingly
unlimited amount of resources to their clients

BITS Pilani, Deemed to be University under Section 3 of UGC Act, 1956

Management of Virtualized Resources

Distributed Management of Virtual Machines

Reservation-Based Provisioning of Virtualized Resources

Provisioning to Meet SLA Commitments

BITS Pilani, Deemed to be University under Section 3 of UGC Act, 1956

 The cloud operating system is
responsible for:
1. managing the physical and
virtual infrastructure,
2. orchestrating and
commanding service
provisioning and deployment
3. providing federation
capabilities for accessing and
deploying virtual resources in
remote cloud infrastructures

BITS Pilani, Deemed to be University under Section 3 of UGC Act, 1956

Technology made cloud possible

Key Technology is Virtualization

App App App

App App App OS OS OS

Operating System Hypervisor

Hardware Hardware

Traditional Stack Virtualized Stack

Virtualization plays an important role as an enabling technology for

datacentre implementation by abstracting compute, network, and storage
service platforms from the underlying physical hardware

BITS Pilani, Deemed to be University under Section 3 of UGC Act, 1956

What is Virtualization

BITS Pilani
What does Virtualization do?
• Virtualization allows multiple operating system instances to run
concurrently on a single computer
• It is a means of separating hardware from a single operating
system.
• Each “guest” OS is managed by a Virtual Machine Monitor (VMM),
also known as a hypervisor.
• Because the virtualization system sits between the guest and the
hardware, it can control the guests’ use of CPU, memory, and
storage, even allowing a guest OS to migrate from one machine to
another.
• Instead of purchasing and maintaining an entire computer for one
application, each application can be given its own operating
system, and all those operating systems can reside on a single
piece of hardware.
• Virtualization allows an operator to control a guest operating
system’s use of CPU, memory, storage, and other resources, so
each guest receives only the resources that it needs. 6

BITS Pilani
Changes after Virtualization
Before Virtualization After Virtualization
• Single OS image per machine • Hardware-independence of
• Software and hardware tightly operating system and
coupled applications
• Running multiple applications on • Virtual machines can be
same machine often creates provisioned to any system
conflict
• Can manage OS and
• Underutilized resources
application as a single unit by
• Inflexible and costly encapsulating them into
infrastructure
virtual machines

BITS Pilani
 Virtualization Architecture
• OS assumes complete control of the underlying
hardware.
• Virtualization architecture provides this illusion through a
hypervisor/VMM.
• Hypervisor/VMM is a software layer which:
• Allows multiple Guest OS (Virtual Machines) to run
simultaneously on a single physical host
• Provides hardware abstraction to the running Guest
OSs and efficiently multiplexes underlying hardware
resources

BITS Pilani
Hypervisor

A thin layer of software that generally provides virtual

partitioning capabilities which runs directly on hardware,
but underneath higher-level virtualization services.
Sometimes referred to as a “bare metal” approach.

BITS Pilani
Hypervisor Design Goals

• Isolation
• Security isolation
• Fault isolation
• Resource isolation
• Reliability
• Minimal code base
• Strictly layered design
• Not extensible
• Scalability
• Scale to large number of cores
• Large memory systems
5

BITS Pilani
How Hypervisor goals are achieved?
• Partitioning Kernel
▪ “Partition” is isolation boundary
▪ Few virtualization functions; relies on virtualization stack
• Very thin layer of software
▪ Microkernel
▪ Highly reliable
▪ Basis for smaller Trusted Computing Base (TCB)
• No device drivers
▪ Drivers run in a partition
• Well-defined interface
▪ Allow others to create support for their OSes as guests

BITS Pilani
Hypervisor
Monolithic versus Microkernelized
• Monolithic hypervisor • Microkernelized hypervisor
• Simpler than a modern ▪ Simple partitioning functionality
kernel, but still complex ▪ Increase reliability and minimize
lowest level of the TCB
• Contains its own drivers
▪ No third-party code
model
▪ Drivers run within guests

BITS Pilani
CPU Sharing

• VMM or Hypervisor provides a virtual view of CPU to

VMs.
• In multi processing, CPU is alloted to the different
processes in form of time slices by the OS.
• Similarly VMM or Hypervisor allots CPU to different VMs.

BITS Pilani
Memory Sharing

• In Multiprogramming there is a single level of indirection

maintained by Kernel.
• In case of Virtual Machines there is one more level of
indirection maintained by VMM

10

BITS Pilani
IO Sharing

• Device needs to use Physical Memory location.

• In a virtualized environment, the kernel is running in a
hypervisor-provided virtual address space
• Allowing the guest kernel to convey an arbitrary location
to device for writing is a serious security hole
• Each device defines its own protocol for talking to drivers

11

BITS Pilani
Approaches for Virtualization

BITS Pilani
Full Virtualization

❑ Full virtualization
• In its basic form known as “full virtualization” the
hypervisor provides a fully emulated machine in which an
operating system can run. VMWare is a good example.
• The biggest advantage to this approach is its flexibility:
one could run a RISC-based OS as a guest on an Intel-
based host.
• While this is an obvious approach, there are significant
performance problems in trying to emulate a complete set
of hardware in software.

BITS Pilani
ParaVirtualization
❑ Paravirtualization
• “Paravirtualization,” found in the XenSource, open source Xen
product, attempts to reconcile these two approaches. Instead of
emulating hardware, paravirtualization uses slightly altered
versions of the operating system which allows access to the
hardware resources directly as managed by the hypervisor.
• This is known as hardware-assisted virtualization, and improves
performance significantly.
• In order to retain flexibility, the guest OS is not tied to its host OS.
Drastically different operating systems can be running in a
hypervisor at the same time, just as they can under full
virtualization.
• In this way, paravirtualization can be thought of as a low-
overhead full virtualization
5

BITS Pilani
SKI Virtualization

❑ Single Kernel Image (SKI),

• Single Kernel Image (SKI), in which the host OS spawns
additional copies of itself. This kind of virtualization can be found
in Swsoft Virtuozzo and Sun Solaris, Zones. SKI can be thought
of as “lightweight” virtualization.
• While this approach avoids the performance problems with pure
emulation, it does so at the expense of flexibility.
• It is not possible, for instance, to run different versions or even
different patch levels of a particular operating system on the
same machine.
• Whatever versions exist in the host, that same software will be
provided in the guest. SKI also sacrifices the security and
reliability provided by other virtualization methods.
6

BITS Pilani
x86 Hardware Virtualization
• For Industry-standard x86 systems, the two approaches typically
used with software-based partitioning are
• hosted and
• hypervisor architectures
• A hosted approach provides partitioning services on top of a
standard operating system and supports the broadest range of
hardware configurations.
• In contrast, a hypervisor architecture is the layer of software
installed on a clean x86-based system (hence it is often referred to
as a “bare metal” approach). Since it has direct access to the
hardware resources, a hypervisor is more efficient than hosted
architectures, enabling greater scalability, robustness and
performance

BITS Pilani
x86 Hardware Virtualization

BITS Pilani
Advantages of Virtualization

• Instant provisioning - fast scalability

• Live Migration is possible
• Load balancing and consolidation in a Data Center
is possible.
• Low downtime for maintenance
• Virtual hardware supports legacy operating
systems efficiently
• Security and fault isolation

11

BITS Pilani
Issues to be aware of
• Software licensing
One of the most significant virtualization-related issues to be aware of is software
licensing. Virtualization makes it easy to create new servers, but each VM requires its
own separate software license. Organizations using expensive licensed applications
could end up paying large amounts in license fees if they do not control their server
sprawl.
• IT training
IT staff used to dealing with physical systems will need a certain amount of training in
virtualization. Such training is essential to enable the staff to debug and troubleshoot
issues in the virtual environment, to secure and manage VMs, and to effectively plan
for capacity.
• Hardware investment
Server virtualization is most effective when powerful physical machines are used to
host several VMs. This means that organizations that have existing not-so-powerful
hardware might still need to make upfront investments in acquiring new physical
servers to harvest the benefits of virtualization

16

BITS Pilani
Issues to be aware of

• Performance can be a concern, especially for in-band

deployments, where the virtualization controller or
appliance can become a bandwidth bottleneck.

• Interoperability among vendor products is still evolving.

• Failure of the virtualization device, leading to loss of the

mapping table.

17

BITS Pilani
BITS Pilani
Current Problem the Industry is facing

BITS Pilani
BITS Pilani
 Dockers
• All applications have their own dependencies, which
include both software and hardware resources.
• Docker is a mechanism that helps in isolating the
dependencies per each application by packing them into
containers.
• In terms of technology, it provides cloud portability by
running the same applications in different virtual
environments.
• Containers are scalable and safer to use and deploy as
compared to regular approaches.

BITS Pilani
 Virtual Machines
• Virtual machines are used extensively in cloud
computing.
• Isolation and resource control have continually been
achieved through the use of virtual machines.
• Virtual machine loads a full OS with its own memory
management and enable applications to be more efficient
and secure while ensuring their high availability.
App App App

App App App OS OS OS

Operating System Hypervisor

Hardware Hardware

Traditional Stack Virtualized Stack

BITS Pilani
 How are Docker Containers different from a
Virtual Machine?
• Virtual machines have a full OS with its own
memory management installed with the
associated overhead of virtual device
drivers.
• Docker containers are executed with the
Docker engine rather than the hypervisor.
• Containers are therefore smaller than Virtual
Machines and enable faster start up with
better performance, less isolation and
greater compatibility possible due to sharing
of the host’s kernel.

BITS Pilani
How are Docker Containers different from a
Virtual Machine?

BITS Pilani
 Docker Container Lifecycle ……
• The Life of a Container
– Conception
• BUILD an Image from a Dockerfile
– Birth
• RUN (create+start) a container
– Reproduction
• COMMIT (persist) a container to a new
image
• RUN a new container from an image
– Sleep
• KILL a running container
– Wake
• START a stopped container
– Death
• RM (delete) a stopped container
• Extinction
– RMI a container image (delete image)
BITS Pilani
 Dockerfile ……
• Like a Makefile (shell script with keywords)
• Extends from a Base Image
• Results in a new Docker Image
• Imperative, not Declarative
• A Docker file lists the steps needed to build an images
• docker build is used to run a Docker file
• Can define default command for docker run, ports to expose, etc

BITS Pilani
https://docs.docker.com/engine/installation/windows/
Thank you

BITS Pilani
IaaS

Really, what is iaas???

BITS Pilani
heard of 3 models of Cloud
Computing?
Yes, Yes, IaaS, PaaS and SaaS

BITS Pilani
Key concepts of IaaS
• Cloudbursting: The process of off-loading tasks to the cloud during times
when the most compute resources are needed

• Multi-tenant computing

• Resource pooling: Pooling is a resource management term that refers to

the grouping together of resources (compute(cpu), network(bandwidth),
storage) for the purposes of maximizing advantage and/or minimizing
risk to the users
• Hypervisor

BITS Pilani
Two primary facets that make
IaaS special
Elasticity:
Wikipedia: “In cloud computing,
elasticity is defined as the degree
to which a system (or a particular
cloud layer) autonomously adapts
its capacity to workload over time”
OR simply put “Ability of a system to
expand or contract its dedicated
resources to meet the demand”

&
Virtualization

BITS Pilani
4 considerations:

• Developing for a specific vendor's proprietary IaaS could

prove to be a costly mistake
• The complexity of well-written resource allocation
software is significant and do not come cheap
• What will you be sending off to be processed in the
cloud? Sending data such as personal identities, financial
information, and health care data put an organization's
compliance at risk
• Understand the dangers of shipping off processes that
are critical to the day-to-day operation of the business.

• http://www.ibm.com/developerworks/cloud/library/cl-cloudservices1iaas/

BITS Pilani
IaaS

Really, what is iaas???

BITS Pilani
Amazon Web Services

Amazon Web Services Cloud

• Provides highly reliable and scalable infrastructure for
deploying web-scale solutions
• With minimal support and administration costs
• More flexibility than own infrastructure, either on-premise
or at a datacenter facility

BITS Pilani
AWS infrastructure services

BITS Pilani
Examples

AWS – EC2, EBS, S3, LB

BITS Pilani
IaaS for you

Thanks, I feel so “Clouded” now

BITS Pilani
heard of 3 models of Cloud
Computing?
Yes, Yes, IaaS, PaaS and SaaS

BITS Pilani
 Openstack overview
 OpenStack is a collection of open source technologies
delivering a massively scalable cloud operating system.

 OpenStack cloud operating system controls large pools of

compute, storage, and networking resources throughout a
datacenter, all managed through a dashboard that gives
administrators control while empowering their users to
provision resources through a web interface.

BITS Pilani
Openstack Components

BITS Pilani
OpenStack conceptual architecture

BITS Pilani
Virtual Machine Provisioning and Manageability Life
Cycle
• The cycle starts by a request delivered to the IT
department, stating the requirement for creating
a new server for a particular service.
• This request is being processed by the IT
administration to start seeing the servers’ Virtual Machine Life Cycle
resource pool, matching these resources with
requirements
•Starting the provision of the needed virtual
machine.
• Once it provisioned and started, it is ready to
provide the required service according to an
SLA(Service Level agreement ).
3
• Virtual is being released; and free resources.
BITS Pilani
 VM Provisioning Process
Steps to Provision VM -
• Select a server from a pool of available servers along with the appropriate OS template
you need to provision the virtual machine.
• Load the appropriate software.
• Customize and configure the machine (e.g., IP address, Gateway) to an associated
network and storage resources.
• Finally, the virtual server is ready to start with its newly loaded S/W.

BITS Pilani
 VM Provisioning
• Server provisioning is defining server’s configuration based on the organization
requirements, a H/W, and S/W component (processor, RAM, storage, networking,
operating system, applications, etc.).

VMs can be provisioned by

• Manually installing an OS,
• Using a preconfigured VM template,
• Cloning an existing VM, or importing a physical server or a
• Server from another hosting platform.
• Physical servers can also be virtualized and provisioned using P2V (Physical
to Virtual)

BITS Pilani
 VM Provisioning using templates

• Provisioning from a template reduces the time required to create a

new virtual machine.
• Administrators can create different templates for different
purposes.
For example –
• Vagrant provision tool using VagrantFile (template file) (demo)
• Heat – Orchestration Tool of openstack (Heat template in YAML format)
(demo – Instance creation in cloud, Load balancer in cloud)
This enables the administrator to quickly provision a correctly configured
virtual server on demand.
6

BITS Pilani
 Virtual Machine Migration Services
Migration service -
The process of moving a virtual machine from one host server or storage location to
another;
There are different techniques of VM migration-
- Hot/live migration,
- Cold/regular migration, and
- Live storage migration of a virtual machine.

In this process, all key machines’ components, such as CPU, storage disks,
networking, and memory, are completely virtualized, thereby facilitating the entire state
of a virtual machine to be captured by a set of easily moved data files.

BITS Pilani
 Cold/regular migration
Cold migration is the migration of a powered-off virtual machine and is done in
the following tasks:
• If the option to move to a different datastore was chosen, the configuration
files, including the NVRAM file (BIOS settings), and log files are moved from
the source host to the destination host’s associated storage area. If you
chose to move the virtual machine's disks, these are also moved.
• The virtual machine is registered with the new host.
• After the migration is completed, the old version of the virtual machine is
deleted from the source host if the option to move to a different datastore
was chosen.

BITS Pilani
 Live Migration Technique
Pre-assumption :
– We assume that all storage resources are
separated from computing resources.
– Storage devices of VMs are attached from
network :

• NAS: NFS, CIFS VM 1 VM 2

• SAN: Fibre Channel

• iSCSI, network block device
• drdb network RAID
– Require high quality network connection

• Common L2 network (LAN)

• L3 re-routing
VM 2 VM 1
Disk Disk

BITS Pilani
 Live Migration Technique
Challenges of live migration :
– VMs have lots of state in memory
– Some VMs have soft real-time
requirements :

• For examples, web servers,

databases and game servers, ...etc.
• Need to minimize down-time
Relocation strategy :
1. Pre-migration process
2. Reservation process
3. Iterative pre-copy
4. Stop and copy
5. Commitment

BITS Pilani
Live Migration Technique

BITS Pilani
 Live Migration Technique
Live migration process :

Host A Host B
BITS Pilani
 Live Migration Technique
Live migration process :

Host A Host B
BITS Pilani
 Live Migration Technique
Live migration process :

Host A Host B
BITS Pilani
 Live Migration Demo

• Using Proxmox deployment tool

BITS Pilani
Dependency on IaaS and PaaS

3
BITS Pilani
Introduction to PaaS

• Platform as a Service, referred to as PaaS, is a category

of cloud computing that provides a platform and
environment to allow developers to build applications and
services over the internet.
• Platform as a Service allows users to create software
applications using tools supplied by the provider.
• PaaS services are hosted in the cloud and accessed by
users simply via their web browser.
• PaaS services can consist of preconfigured features that
customers can subscribe to; they can choose to include
the features that meet their requirements while discarding
those that do not.

BITS Pilani
Building blocks of PaaS

• PaaS providers can assist developers from the conception of their

original ideas to the creation of applications, and through to testing
and deployment.
• Below are some of the features that can be included with a PaaS
offering:
❑ Operating system
❑ Server-side scripting environment
❑ Database management system
❑ Server Software
❑ Support
❑ Storage
❑ Network access
❑ Tools for design and development
❑ Hosting

BITS Pilani
Characteristics of PAAS
• Services to develop, test, deploy, host and maintain applications in the same
integrated development environment. All the varying services needed to fulfill the
application development process

• Web based user interface creation tools help to create, modify, test and deploy
different UI scenarios

• Multi-tenant architecture where multiple concurrent users utilize the same

development application

• Built in scalability of deployed software including load balancing and failover

• Integration with web services and databases via common standards

• Support for development team collaboration – some PaaS solutions include

project planning and communication tools

• Tools to handle billing and subscription management

BITS Pilani
Characteristics of PAAS
PaaS, which is similar in many ways to Infrastructure as a
Service, is differentiated from IaaS by the addition of value
added services and comes in two distinct flavours;
1. A collaborative platform for software development, focused on
workflow management regardless of the data source being
used for the application. An example of this approach would
be Heroku, a PaaS that utilizes the Ruby on Rails
development language.
2. A platform that allows for the creation of software utilizing
proprietary data from an application. This sort of PaaS can be
seen as a method to create applications with a common data
form or type. An example of this sort of platform would be the
Force.com. PaaS from Salesforce.com which is used almost
exclusively to develop applications that work with the
Salesforce.com CRM

BITS Pilani
Advantages and Risks
Advantages
• Users don’t have to invest in physical infrastructure
• PaaS allows developers to frequently change or upgrade
operating system features. It also helps development teams
collaborate on projects.
• Makes development possible for ‘non-experts’
• Teams in various locations can work together
• Security is provided, including data security and backup and
recovery.
• Adaptability; Features can be changed if circumstances dictate
that they should.
• Flexibility; customers can have control over the tools that are
installed within their platforms and can create a platform that suits
their specific requirements. They can ‘pick and choose’ the
features they feel are necessary.

BITS Pilani
Advantages and Risks

Risks
• Since users rely on a provider's infrastructure and
software, vendor lock-in can be an issue in PaaS
environments.
• Other risks associated with PaaS are provider downtime
or a provider changing its development roadmap.
• If a provider stops supporting a certain programming
language, users may be forced to change their
programming language, or the provider itself. Both are
difficult and disruptive steps.

BITS Pilani
Paas Example
• PaaS does not typically replace a business' entire
infrastructure. Instead, a business relies on PaaS providers for
key services, such as Java development or application
hosting.
• For example:
Deploying a typical business tool locally might require an IT
team to buy and install hardware, operating systems,
middleware (such as databases, Web servers and so on) the
actual application, define user access or security, and then
add the application to existing systems management or
application performance monitoring (APM) tools. IT teams
must then maintain all of these resources over time.
Paas solution: A PaaS provider, however, supports all the
underlying computing and software; users only need to log in
and start using the platform – usually through a Web browser
interface.

BITS Pilani
Paas Example: Windows Azure

• Windows Azure is Microsoft's operating system for cloud

computing.
• Windows Azure is intended to simplify IT management and
minimize up-front and ongoing expenses
• To this end, Azure was designed to facilitate the management
of scalable Web applications over the Internet.
• Windows Azure can be used to create, distribute and upgrade
Web applications without the need to maintain expensive,
often underutilized resources onsite.
• New Web services and applications can be written and
debugged with a minimum of overhead and personnel
expense.

BITS Pilani
Paas Example: Windows Azure

• The Azure operating system is the central component of the

company's Azure Services Platform, which also includes
separate application, security, storage and virtualization
service layers and a desktop development environment.
• Windows Azure supports a wide variety of Microsoft and third-
party standards, protocols, programming languages and
platforms. Examples include XML (Extensible Markup
Language), REST (representational state transfer), SOAP
(Simple Object Access Protocol), Eclipse, Ruby, PHP and
Python.
• Although it faces steep competition from Amazon Web
Services (AWS), Microsoft Azure has managed to hold a
strong second place among cloud hosting platform
providers. http://azure.microsoft.com/en-us/

BITS Pilani
Windows Azure Runtime Environment
• The Windows Azure runtime environment provides a
scalable compute and storage hosting environment along
with management capabilities. It has three major
components: Compute, Storage and the Fabric Controller

BITS Pilani
Windows Azure Runtime Environment
• The hosting environment of Azure is called the Fabric Controller.
It has a pool of individual systems connected on a network and
automatically manages resources by load balancing and geo-
replication. It manages the application lifecycle without requiring
the hosted apps to explicitly deal with the scalability and
availability requirements. Each physical machine hosts an Azure
agent that manages the machine.
• The Azure Compute Service provides a Windows-based
environment to run applications written in the various languages
and technologies supported on the Windows platform.
• The Windows Azure storage service provides scalable storage
for applications running on the Windows Azure in multiple forms.
It enables storage for binary and text data, messages and
structured data through support for features called Blobs,Tables,
Queues and Drives.

BITS Pilani
BITS Pilani
Server Rack 1 Server Rack 2

BITS Pilani
BITS Pilani
BITS Pilani
BITS Pilani
BITS Pilani
Paas Vendors

• Common PaaS vendors include Salesforce.com's

Force.com, which provides an enterprise customer
relationship management (CRM) platform. PaaS
platforms for software development and management
include Appear IQ, Mendix, Amazon Web Services
(AWS) Elastic Beanstalk, Google App Engine and
Heroku.

BITS Pilani
Dependency on IaaS and PaaS

BITS Pilani
What is SaaS?
• Software as a service is a software distribution model in which
applications are hosted by a vendor or service provider and made
available to customers over a network, typically the Internet.

• Shortly, in the SaaS model software is deployed as a hosted service

and accessed over the Internet, as opposed to “On Premise.”

• Software delivered to home consumers, small business, medium and

large business
– The traditional model of software distribution, in which software
is purchased for and installed on personal computers, is
sometimes referred to as software as a product.

BITS Pilani
Problems in traditional Model

• In the traditional model of software delivery, the customer

acquires a perpetual license and assumes responsibility
for managing the software.

• There is a high upfront cost associated with the purchase

of the license, as well as the burden of implementation
and ongoing maintenance.

• ROI is often delayed considerably, and, due to the rapid

pace of technological change, expensive software
solutions can quickly become obsolete.

BITS Pilani
Problems in traditional Model
Traditional Software On-Demand Utility

Plug In, Subscribe

Build Your Own
Pay-per-Use

BITS Pilani
SaaS – How is it delivered
• The web as a platform is the center point. The web as a platform is
the center point
• Network-based access to, and management of, commercially
available (i.e., not custom) software application delivery that typically
is closer to a one-to-many model (single instance, multi-tenant
architecture) than to a one-to-one model, including architecture,
pricing, partnering, and management characteristics

• Software delivered to home consumers, small business, medium and

large business
– The traditional model of software distribution, in which software
is purchased for and installed on personal computers, is
sometimes referred to as software as a product.

BITS Pilani
SaaS – Architecture

• Run by
• Bandwidth technologies
• The cost of a PC has been reduced
significantly with more powerful computing but
the cost of application software has not
followed
• Timely and expensive setup and maintenance
costs
• Licensing issues for business are contributing
significantly to the use of illegal software and
piracy. 3

BITS Pilani
SaaS Application Architecture

• Scalable
• Multitenant efficient
• Configurable

• Scaling the application - maximizing

concurrency, and using application resources
more efficiently
• i.e. optimizing locking duration, statelessness,
sharing pooled resources such as threads and
network connections, caching reference data,
and partitioning large databases. 4

BITS Pilani
SaaS Application Architecture

• Multi-tenancy – important architectural shift from

designing isolated, single-tenant applications
• One application instance must be able to
accommodate users from multiple other
companies at the same time
• All transparent to any of the users.
• This requires an architecture that maximizes the
sharing of resources across tenants
• is still able to differentiate data belonging to
different customers.
5

BITS Pilani
SaaS Application Architecture
• Configurable - a single application instance on a single
server has to accommodate users from several different
companies at once
• To customize the application for one customer will change
the application for other customers as well.
• Traditionally customizing an application would mean code
changes
• Each customer uses metadata to configure the way the
application appears and behaves for its users.
• Customers configuring applications must be simple and easy
without incurring extra development or operation costs

BITS Pilani
SaaS Models

BITS Pilani
Business Model comparisons
Traditional packaged software Software as a service

 Designed for  Designed from the

customers to install, outset up for delivery
manage and maintain. as Internet-based
services

 Architect solutions to  Designed to run

be run by an individual thousands of different
company in a customers on a single
dedicated instantiation code
of the software

BITS Pilani
Business Model comparisons
Traditional packaged Software Software as a serivce

 Infrequent, major  Frequent, "digestible"

upgrades every 18-24 upgrades every 3-6
months, sold individually months to minimize
to each installed base customer disruption and
customer. enhance satisfaction.
 Fixing a problem for one
 Version control
customer fixes it for
 Upgrade fee everyone

 Streamlined, repeatable  May use open APIs and Web

functionality via Web services to facilitate integration,
services, open APIs and but each customer must typically
standard connectors pay for one-off integration work.
BITS Pilani
Business Model comparisons

Hidden Cost

BITS Pilani
SaaS Advantages

BITS Pilani
BITS Pilani
Applicability – Scenario 1
Single-User software application
– Organize personal information
– Run on users’ own local computer
– Serve only one user at a time
– Inapplicable to Saas model
• Data security issue
• Network performance issue
– Example: Microsoft office suite

12

BITS Pilani
Applicability – Scenario 2
Infrastructure software
– Serve as the foundation for most other
enterprise software application
– Inapplicable to Saas model
• Installation locally is required
• Form the basis to run other application
– Example: Window XP, Oracle database

13

BITS Pilani
Applicability – Scenario 3
Embedded Software
– Software component for embedded system
– Support the functionality of the hardware device
– Inapplicable to Saas model
• Embedded software and hardware is
combined together and is inseparable
– Example: software embedded in ATM
machines, cell phones, routers, medical
equipment, etc

14

BITS Pilani
Applicability – Scenario 4
Enterprise Software Application
– Perform business functions
– Organize internal and external information
– Share data among internal and external users
– The most standard type of software applicable
to Saas model
– Example: Saleforce.com CRM application,
Siebel On-demand application

15

BITS Pilani
Cloud ? That looks easy!!!

Is it So?????

BITS Pilani
Introduction

So CLOUD requires managing……

But what will you manage

BITS Pilani
Introduction
Ok, ok got to know what to manage…..
Is it so………..????

BITS Pilani
• OMG!!!!

Therefore the key requirement

for cloud architecture is
“efficient management”
of resources at all the
three layers of cloud stack
BITS Pilani
• Cloud Distributed environment
• With large scale of systems to manage
• Support of multi-tenancy
• Management to maintain SLAs

• So there is need for automation to replace manual

operations and to reduce overall cost

BITS Pilani
Need of the hour???

Virtual Infrastructure Managers

BITS Pilani
Why a Virtual Infrastructure Manager?
 VMs are great!!...but something more is needed
 Where did/do I put my VM? (scheduling & monitoring)
 How do I provision a new cluster node? (clone)
 What IP addresses are available? (networking)
 Provide a uniform view of the resource pool
 Life-cycle management and monitoring of VM
 The VIM should integrate Image, Network and
Virtualization

VM VM VM VM

OpenNebula (VIM)
hypervisor hypervisor hypervisor hypervisor

BITS Pilani
Extending the Benefits of Virtualization to Clusters

• Dynamic deployment and re-placement of virtual machines on a

pool of physical resources
• Transform a rigid distributed physical infrastructure into a flexible
and agile virtual infrastructure
• Backend of Public Cloud:
Internal management of the
infrastructure
• Private Cloud: Virtualization
of cluster or data-center for
internal users
• Cloud Interoperation: On-
demand access to public
clouds
BITS Pilani
Virtual Machine Management Model
Distributed VM Management Model

BITS Pilani
What is OpenNebula?

BITS Pilani
 What is OpenNebula?
Enabling Technology to Build your Cloud
• Private Cloud to simplify and optimize internal operations
• Hybrid Cloud to supplement the capacity of the Private Cloud
• Public Cloud to expose your Private to external users

BITS Pilani
 The Benefits of OpenNebula
For the Infrastructure Manager
• Centralized management of VM workload and distributed
infrastructures
• Support for VM placement policies: balance of workload, server
consolidation…
• Dynamic resizing of the infrastructure
• Dynamic partition and isolation of clusters
• Dynamic scaling of private infrastructure to meet fluctuating
demands
• Lower infrastructure expenses combining local and remote Cloud
resources
For the Infrastructure User
• Faster delivery and scalability of services
• Support for heterogeneous execution environments
• Full control of the lifecycle of virtualized services management
BITS Pilani
Interoperability from the Cloud Provider
perspective
Interoperable (platform independent), innovative (feature-rich) and
proven (mature to run in production).

BITS Pilani
The main features of OpenNebula
Feature Function

Internal Interface • Unix-like CLI for fully management of VM life-cycle and physical boxes
• XML-RPC API and libvirt virtualization API

Scheduler • Requirement/rank matchmaker allowing the definition of workload and resource-aware allocation
policies
• Support for advance reservation of capacity through Haizea

Virtualization • Xen, KVM, and VMware

Management • Generic libvirt connector (VirtualBox planned for 1.4.2)

Image Management • General mechanisms to transfer and clone VM images

Network Management • Definition of isolated virtual networks to interconnect VMs

Service Management and • Support for multi-tier services consisting of groups of inter-connected VMs, and their auto-
Contextualization configuration at boot time

Security • Management of users by the infrastructure administrator

Fault Tolerance • Persistent database backend to store host and VM information

Scalability • Tested in the management of medium scale infrastructures with hundreds of servers and VMs (no
scalability issues has been reported)

Flexibility and • Open, flexible and extensible architecture, interfaces and components, allowing its integration with
Extensibility any product or tool

BITS Pilani
Inside OpenNebula

BITS Pilani
 OpenNebula Architecture
Tools

Command Line
Scheduler Other Tools
Interface

Request Manager
SQL Pool
(XML-RPC)
Core

VM Host VN
Manager Manager Manager
Drivers

Transfer Virtual Machine Information

Driver Driver Driver

BITS Pilani
 The Core
• Request manager: Provides a XML-RPC interface to manage
and get information about ONE entities.
• SQL Pool: Database that holds the state of ONE entities.
• VM Manager (virtual machine): Takes care of the VM life cycle.
• Host Manager: Holds handling information about hosts.
• VN Manager (virtual network): This component is in charge of
generating MAC and IP addresses.
Tools

Command Line
Scheduler Other Tools
Interface

Request Manager
SQL Pool
(XML-RPC)
Core

VM Host VN
Manager Manager Manager
Drivers

Transfer Virtual Machine Information

Driver Driver Driver
BITS Pilani
The tools layer
Scheduler:
– Searches for physical hosts to deploy newly defined
VMs
Command Line Interface:
– Commands to manage OpenNebula.
– onevm: Virtual Machines
• create, list, migrate…
– onehost: Hosts
Tools
Command Line
Scheduler Other Tools
Interface

• create, list, disable… SQL Pool

Request Manager
(XML-RPC)
Core

– onevnet: Virtual Networks VM Host VN

• create, list, delete… Manager Manager Manager

Drivers

Transfer Virtual Machine Information

Driver Driver Driver

BITS Pilani
The drivers layer
Transfer Driver: Takes care of the images.
– cloning, deleting, creating swap image…
Virtual Machine Driver: Manager of the lifecycle of a virtual
machine
– deploy, shutdown, poll, migrate…
Information Driver: Executes scripts in physical hosts to
gather information about them
– total memory, free memory,

Tools
Command Line
Scheduler Other Tools
Interface

total #cpus, cpu consumed… SQL Pool

Request Manager
(XML-RPC)
Core
VM Host VN
Manager Manager Manager
Drivers

Transfer Virtual Machine Information

Driver Driver Driver

BITS Pilani
 Process separation
Scheduler CLI

XML-RPC Interface

OpenNebula Core
Unix Pipes (diver message protocol)

Transfer Virtual Machine Information

Manager Driver Driver Driver

• Scheduler is a separated process, just like command line interface.

• Drivers are also separated processes using a simple text messaging protocol to
communicate with OpenNebula Core Daemon (oned)
BITS Pilani
Constructing a private cloud

BITS Pilani
 System Overview
 Executes the OpenNebula Services
 Usually acts as a classical cluster front-end
 Repository of VM images
 Multiple backends (LVM, iSCSI..)

 Modular components to
interact with the cluster
services
 Types: storage,
monitoring, virtualization
and network

 Provides physical resources to VMs

The same host can be can be a
the front-end and a node  Must have a hypervisor installed

BITS Pilani
Storage

Complex Storage behind OpenNebula

Virtual machines and

Transfer Manager Drivers their images are
Datastore
shared ssh iscsi qcow vmware represented as files
System OK OK Virtual machines and
File-System OK OK OK their images are
represented as block
iSCSI OK
devices (just like a disk)
VMware OK OK OK
BITS Pilani
Network

Networking for private clouds

 OpenNebula management operations use ssh
connections
 Image traffic, may require the movement of heavy
files (VM images, checkpoints). Dedicated storage
links may be a good idea
 VM demands, consider the typical requirements of
your VMs. Several NICs to support the VM traffic may
be a good idea
 OpenNebula relies on bridge networking for the VMs

BITS Pilani
Network

Example network setup in a private cloud

BITS Pilani
Virtual machines

BITS Pilani
VMs

VM Description
Option Description
NAME • Name that the VM will get for description purposes.
CPU • Percentage of CPU divided by 100 required for the Virtual
Machine.
OS (KERNEL, INITRD) • Path of the kernel and initrd files to boot from.
DISK (SOURCE, • Description of a disk image to attach to the VM.
TARGET, CLONE, TYPE)

NIC (NETWORK) • Definition of a virtual network the VM will be attached to.

Multiple disk an network interfaces can be specified just adding

more disk/nic statements.
To create swap images you can specify TYPE=swap, SIZE=<size
in MB>.
By default disk images are cloned, if you do not want that to
happen CLONE=no can be specified and the VM will attach
the original image.

BITS Pilani
VMs

VM States overview

HOLD MIGRATE

PENDING PROLOG BOOT RUNNING SHUTDOWN EPILOG DONE

SUSPENDED STOPPED

BITS Pilani
VMs

Pending state
After submitting a VM description to ONE it is added to the
database and its state is set to PENDING.
In this state IP and MAC addresses are also chosen if they
are not explicitly defined.
The scheduler awakes every 30 seconds and looks for VM
descriptions in PENDING state and searches for a
physical node that meets its requirements. Then a deploy
XML-RPC message is sent to oned to make it run in the
selected node.
Deployment can be also made manually using the
Command Line Interface:
⇒ onevm deploy <vmid> <hostid>

PENDING PROLOG BOOT RUNNING SHUTDOWN EPILOG DONE

BITS Pilani
VMs

Prolog state
In PROLOG state the Transfer Driver prepares the images
to be used by the VM.
Transfer actions:
– CLONE: Makes a copy of a disk image file to be used by the
VM. If Clone option for that file is set to false and the Transfer
Driver is configured for NFS then a symbolic link is created.
– MKSWAP: Creates a swap disk image on the fly to be used by
the VM if it is specified in the VM description.

PENDING PROLOG BOOT RUNNING SHUTDOWN EPILOG DONE

BITS Pilani
VMs

Boot state
In this state a deployment file specific for the virtualization
technology configured for the physical host is generated
using the information provided in the VM description file.
Then Virtual Machine Driver sends deploy command to
the virtual host to start the VM.
The VM will be in this state until deployment finishes or fails.

PENDING PROLOG BOOT RUNNING SHUTDOWN EPILOG DONE

BITS Pilani
VMs

Contextualization
The ISO image has the contextualization for that VM:
– context.sh: contains configuration variables
– init.sh: script called by VM at start to configure
specific services
– certificates: directory that contains certificates for
some service
– service.conf: service configuration

User provided, OpenNebula provided contextualization info

BITS Pilani
VMs

Running and Shutdown states

While the VM is in RUNNING state it will be periodically
polled to get its consumption and state.
In SHUTDOWN state Virtual Machine Driver will send the
shutdown command to the underlying virtual
infrastructure.

PENDING PROLOG BOOT RUNNING SHUTDOWN EPILOG DONE

BITS Pilani
VMs

Epilog state
In EPILOG state the Transfer Manager Driver is called
again to perform this actions:
– Copy back the images that have SAVE=yes option.
– Delete images that were cloned or generated by MKSWAP.

PENDING PROLOG BOOT RUNNING SHUTDOWN EPILOG DONE

BITS Pilani
Hybrid cloud

BITS Pilani
Overview

BITS Pilani
 Making an Amazon EC2 hybrid
Amazon EC2 cloud is managed by OpenNebula as any
other cluster node
– You can use several accounts by adding a driver for
each account (use the arguments attribute, -k and -c
options). Then create a host that uses the driver
– You can use multiple EC2 zones, add a driver for
each zone (use the arguments attribute, -u option),
and a host that uses that driver
– You can limit the use of EC2 instances by modifying
the IM file

BITS Pilani
 Using an EC2 hybrid cloud
Virtual Machines can be instantiated locally or in EC2
The VM template must provide a description for both
instantiation methods.
The EC2 counterpart of your VM (AMI_ID) must be
available for the driver account
The EC2 VM template attribute should describe not only the
VM’s properties but the contact details of the external
cloud provider

BITS Pilani
 Hybrid cloud Use Case
On-demand Scaling of Computing Clusters
• On-demand Scaling of Web Servers
• Elastic execution of the NGinx web
server
• The capacity of the elastic web
application can be dynamically
increased or decreased by adding
or removing NGinx instances

BITS Pilani
High Availability

BITS Pilani
High Availability

BITS Pilani
 High Availability

Steps to achieve high availabilty

Build for server failure
Build for zone failure
Build for Cloud failure
Automating and testing

BITS Pilani
 Key aspects of SLA
In the early days of web-application deployment,
performance of the application at peak load was
a single important criterion for provisioning server
resources.

Provisioning in those days involved deciding

hardware configuration, determining the number of
physical machines, and acquiring them upfront so
that the overall business objectives could be
achieved.

The web applications were hosted on these

dedicated individual servers within enterprises’ own
server rooms. These web applications were used
to provide different kinds of e-services to various 5
clients.
BITS Pilani
 Key aspects of SLA
Typically, the service-level objectives (SLOs) for
these applications were response time and
throughput of the application end-user requests.

The capacity buildup was to cater to the estimated

peak load experienced by the application. The
activity of determining the number of servers and
their capacity that could satisfactorily serve the
application end-user requests at peak loads is
called capacity planning

BITS Pilani
 Key aspects of SLA
Due to the increasing complexity of managing the hugh
Data centres, enterprises started outsourcing the
application hosting to the infrastructure providers. They
would procure the hardware and make it available for
application hosting.
It necessitated the enterprises to enter into a legal
agreement with the infrastructure service providers to
guarantee a minimum quality of service (QoS).
Typically, the QoS parameters are related to the availability
of the system CPU, data storage, and network for
efficient execution of the application at peak loads.
This legal agreement is known as the service-level
agreement (SLA)
BITS Pilani
 Key aspects of SLA
For example, one SLA may state that the application’s
server machine will be available for 99.9% of the key
business hours of the application’s end users, also called
core time, and 85% of the non-core time.
Another SLA may state that the service provider would
respond to a reported issue in less than 10 minutes
during the core time, but would respond in one hour
during non-core time.
These SLAs are known as the infrastructure SLAs, and the
infrastructure service providers are known as
Application Service Providers (ASPs)

BITS Pilani
 Key aspects of SLA
The dedicated hosting practice resulted in massive
redundancies within the ASP’s data centers due to the
underutilization of many of their servers. This is
because the applications were not fully utilizing their
servers’ capacity at nonpeak loads.
To reduce the redundancies and increase the server
utilization in data centers, ASPs started co-hosting
applications with complementary workload patterns.
Co-hosting of applications means deploying more than
one application on a single server. This led to further
cost advantage for both the ASPs and enterprises.

BITS Pilani
 Key aspects of SLA
However, newer challenges such as application
performance isolation and security guarantees emerged
and needed to be addressed. Performance isolation
implies that one application should not steal the
resources being utilized by other co-located applications.
Hence, appropriate measures are needed to guarantee
security and performance isolation. These challenges
prevented ASPs from fully realizing the benefits of co-
hosting.
Virtualization technologies have been proposed to
overcome the above challenges. The ASPs could exploit
the containerization features of virtualization technologies
to provide performance isolation and guarantee data
security to different co-hosted applications
BITS Pilani
 Key aspects of SLA
Adoption of virtualization technologies required ASPs to get
more detailed insight into the application runtime
characteristics with high accuracy. Based on these
characteristics, ASPs can allocate system resources
more efficiently to these applications on-demand, so
that application-level metrics can be monitored and met
effectively.
These metrics are request rates and response times.
Therefore, different SLAs than the infrastructure SLAs are
required. These SLAs are called application SLAs. These
service providers are known as Managed Service
Providers (MSP) because the service providers were
responsible for managing the application availability too.
BITS Pilani
 Key aspects of SLA
To fulfill the SLOs mentioned in the application SLA and
also make their IT infrastructure elastic, an in-depth
understanding of the application’s behavior is required for
the MSPs. Elasticity implies progressively scaling up the
IT infrastructure to take the increasing load of an
application. The customer is billed based on their
application usage of infrastructure resources for a given
period only. The infrastructure can be augmented by
procuring resources dynamically from multiple sources,
including other MSPs, if resources are scarce at their
data centers. This kind of new hosting infrastructure is
called cloud platform.

BITS Pilani
 Key aspects of SLA
Traditionally, load balancing techniques and admission
control mechanisms have been used to provide
guaranteed quality of service (QoS) for hosted web
applications. These mechanisms can be viewed as the
first attempt towards managing the SLOs.
Now it is also possible for a customer and the service
provider to mutually agree upon a set of SLAs with
different performance and cost structure rather than a
single SLA. The customer has the flexibility to choose
any of the agreed SLAs from the available offerings. At
runtime, the customer can switch between the different
SLAs.

BITS Pilani
 Key aspects of SLA
Key Components of a Service-Level Agreement
Service Level Parameter
Describes an observable property of a service whose value is
measurable. Metrics These are definitions of values of service
properties that are measured from a service providing system
or computed from other metrics and constants.
Metrics are the key instrument to describe exactly what SLA
parameters mean by specifying how to measure or compute
the parameter values.
Function A function specifies how to compute a metric’s value
from the values of other metrics and constants. Functions are
central to describing exactly how SLA parameters are
computed from resource metrics.
Measurement directives These specify how to measure a metric.

BITS Pilani
 Key aspects of SLA
Key Contractual Elements of an Infrastructural SLA Hardware
availability 99% uptime in a calendar month Power availability
99.99% of the time in a calendar month
Data center network availability 99.99% of the time in a
calendar month
Backbone network availability 99.999% of the time in a
calendar month
Service credit for unavailability Refund of service credit
prorated on downtime period
Outage notification guarantee Notification of customer within 1
hr of complete downtime
Internet latency guarantee When latency is measured at 5 min
intervals to an upstream provider, the average doesn’t exceed
60 msec
Packet loss guarantee Shall not exceed 1% in a calendar
month BITS Pilani
 Key aspects of SLA
Key contractual components of an application SLA
Service level parameter metric Web site response time (e.g.,
max of 3.5 sec per user request)
Latency of web server (WS) (e.g., max of 0.2 sec per request)
Latency of DB (e.g., max of 0.5 sec per query)Function
Average latency of WS (latency of web server 1+latency of web
server 2 ) /2
Web site response time Average latency of web server+ latency
of database Measurement directive
DB latency available via http://mgmtserver/em/latency WS
latency available via http://mgmtserver/ws/instanceno/ latency
Service level objective Service assurance
web site latency , 1 sec when concurrent connection , 1000
Penalty 1000 USD for every minute while the SLO was breached
BITS Pilani
 Key aspects of SLA
Each SLA goes through a sequence of steps starting from
identification of terms and conditions, activation and
monitoring of the stated terms and conditions, and
eventual termination of contract once the hosting
relationship ceases to exist.
Such a sequence of steps is called SLA life cycle and
consists of the following five phases:
1. Contract definition
2. Publishing and discovery
3. Negotiation
4. Operationalization
5. De-commissioning
BITS Pilani
 Key aspects of SLA
.Some of the parameters
The SLA class (Platinum, Gold, Silver, etc.) to which the
application belongs to.
The amount of penalty associated with SLA breach.
Whether the application is at the threshold of breaching
the SLA.
Whether the application has already breached the SLA.
The number of applications belonging to the same
customer that has breached SLA.
The number of applications belonging to the same
customer about to breach SLA
. The type of action to be performed to rectify the situation.
BITS Pilani
Multitenancy – What is it?

BITS Pilani
Pros and Cons
House Apartment
Effective use of land - +
Privacy + -
Infrastructure sharing - +
Maintenance cost sharing - +
Freedom + -

House: Privacy and freedom

Apartment: Cost efficiency

BITS Pilani
Traditional Deployment Model

BITS Pilani
 Multitenancy – Introduction
• Multi-tenancy is an architecture in which a single instance of a software
application serves multiple customers. Each customer is called a
tenant. Tenants may be given the ability to customize some parts of the
application, such as color of the user interface (UI) or business rules,
but they cannot customize the application's code.
• A software-as-a-service (SaaS) provider, for example, can run one
instance of its application on one instance of a database and provide
web access to multiple customers. In such a scenario, each tenant's
data is isolated and remains invisible to other tenants.

BITS Pilani
Multitenancy – Introduction

 Multi-tenancy is an architectural pattern

 A single instance of the software is run on the service
provider’s infrastructure
 Multiple tenants access the same instance.
 In contrast to the multi-user model, multi-tenancy
requires customizing the single instance according to the
multi-faceted requirements of many tenants.

BITS Pilani
 Multitenancy – key aspects

A Multi-tenants application lets customers (tenants) share the

same hardware resources, by offering them one shared
application and database instance ,while allowing them to
configure the application to fit there needs as if it runs on
dedicated environment.

These definition focus on what we believe to be the key aspects

of multi tenancy:
1.The ability of the application to share hardware resources.
2.The offering of a high degree of configurability of the software.
3.The architectural approach in which the tenants make use of a
single application and database instance.
7

BITS Pilani
 Multi-tenants Deployment Modes
for Application Server
Fully isolated Application server Application Server
Each tenant accesses an application Tenant A
server running on a dedicated servers.
Application server
Tenant B
Virtualized Application Server Application server Virtual
Each tenant accesses a dedicated Tenant A machine
application running on a separate virtual Virtual
machine. Tenant B machine
Application server
Shared Virtual Server
Each tenant accesses a dedicated Tenant A
Virtual
application server running on a shared machine
virtual machine.
Tenant B

Shared Application Server

Session thread
The tenant shared the application server Tenant A
and access application resources Application Server 8
through separate session or threads.
Tenant B Session Thread
BITS Pilani
 Multi-tenants Deployment Modes in Data Centers
Fully isolated data center
The tenants do not share any Tenant A
data center resources
Tenant B
Virtualized servers Virtual
The tenants share the same Tenant A Machine Database

host but access different

databases running on separate Virtual Database
virtual machines Machine
Tenant B
Shared Server
The tenants share the same
server (Hostname or IP) but Tenant A
access different databases
Tenant B
Shared Database
The tenants share the same
server and database (shared or Tenant A
different ports) but access
different schema(tables)
Tenant B
Shared Schema
The tenants share the same
Tenant A
server, database and schema
(tables). The irrespective data
is segregated by key and rows. Tenant B BITS Pilani
 Conceptual framework of Software as a Service
Presentation Menu and Navigation User Controls Display and Rendering Reporting

Security Application Operation

Engine
Identity and federation Monitoring and
User Profile Workflow
Altering
Authentication and Notification and Execution Backup and Restore
Single Sign on Subscription Handling

Authorization and Metadata Provisioning

Orchestration
Role-based Access Execution
Control Engine Configuration and
Customization
Metadata
Entitlement Performance and
Services
Data Availability
Encryption Messaging Synchronization
Metering and Indicators

Regularity Controls

Infrastructure
Networking and 10
Database Storage Computer
Communications
BITS Pilani
Introduction to cloud security

If cloud computing is so great, why isn’t everyone doing it?

• The cloud acts as a big black box, nothing inside the cloud is
visible to the clients
• Clients have no idea or control over what happens inside a
cloud
• Even if the cloud provider is honest, it can have malicious
system admins who can tamper with the VMs and violate
confidentiality and integrity
• Clouds are still subject to traditional data confidentiality,
integrity, availability, and privacy issues, plus some
additional attacks 2

BITS Pilani
Companies are still afraid to use clouds

BITS Pilani
 Cloud Security Issues
• Most security problems stem from:
– Loss of Control
• Take back control
– Data and apps may still need to be on the cloud
– But can they be managed in some way by the consumer?
– Lack of trust
• Increase trust (mechanisms)
– Technology
– Policy, regulation
– Contracts (incentives): topic of a future talk
– Multi-tenancy
• Private cloud
– Takes away the reasons to use a cloud in the first place
• VPC: its still not a separate system
• Strong separation
• These problems exist mainly in 3rd party management models
– Self-managed clouds still have security issues, but not related to
above 4

BITS Pilani
Loss of Control in the Cloud

Consumer’s loss of control

– Data, applications, resources are located with
provider
– User identity management is handled by the cloud
– User access control rules, security policies and
enforcement are managed by the cloud provider
– Consumer relies on provider to ensure
• Data security and privacy
• Resource availability
• Monitoring and repairing of services/resources

BITS Pilani
Multi-tenancy Issues in the
Cloud
• Conflict between tenants’ opposing goals
– Tenants share a pool of resources and have opposing goals
• How does multi-tenancy deal with conflict of interest?
– Can tenants get along together and ‘play nicely’ ?
– If they can’t, can we isolate them?
• How to provide separation between tenants?
• Cloud Computing brings new threats
Multiple independent users share the same physical infrastructure
Thus an attacker can legitimately be in the same physical machine
as the target

BITS Pilani
Taxonomy of Fear

• Confidentiality
– Fear of loss of control over data
• Will the sensitive data stored on a cloud remain
confidential?
• Will cloud compromises leak confidential client data
– Will the cloud provider itself be honest and won’t peek
into the data?
• Integrity
– How do I know that the cloud provider is doing the
computations correctly?
– How do I ensure that the cloud provider really stored
my data without tampering with it?
7

BITS Pilani
Taxonomy of Fear

Availability
• Will critical systems go down at the client, if the
provider is attacked in a Denial of Service attack?
• What happens if cloud provider goes out of
business?
• Would cloud scale well-enough?
• Often-voiced concern
– Although cloud providers argue their downtime
compares well with cloud user’s own data centers

BITS Pilani
Taxonomy of Fear

• Privacy issues raised via massive data mining

• Cloud now stores data from a lot of clients, and can
run data mining algorithms to get large amounts of
information on clients
• Increased attack surface
• Entity outside the organization now stores and
computes data, and so
• Attackers can now target the communication link
between cloud provider and client
• Cloud provider employees can be phished

BITS Pilani
Taxonomy of Fear

• Audit-ability and forensics (out of control of data)

• Difficult to audit data held outside organisation in a
cloud
• Forensics also made difficult since now clients don’t
maintain data locally
• Legal quagmire and transitive trust issues
• Who is responsible for complying with regulations?
• e.g., SOX, HIPAA, GLBA ?
• If cloud provider subcontracts to third party clouds,
will the data still be secure?

10

BITS Pilani
Threat Model

•A threat model helps in analysing a security problem,

design mitigation strategies, and evaluate solutions
•Steps:
– Identify attackers, assets, threats and other
components
– Rank the threats
– Choose mitigation strategies
– Build solutions based on the strategies

11

BITS Pilani
Threat Model

• Basic components
– Attacker modelling
• Choose what attacker to consider
–insider vs. outsider?
–single vs. collaborator?
• Attacker motivation and capabilities
– Attacker goals
– Vulnerabilities / threats

12

BITS Pilani
Thank you

13

BITS Pilani
Hadoop

Name of the elephant!!

BITS Pilani
 Or have u thought of the
following?
• How do “big bazaar/more/D’Mart” target promotions
guaranteed to make you buy?

• How can Airtel(4G) increase Ad-campaign efficiency?

• What’s in your search? How is Google able to make such good

predictions about your search?

• I have huge amount of data(nw sites) data(twitter) data(blog)

data(feeds) data(forums) ? What do I do with it?

BITS Pilani
Wow, that’s so much of DATA
to process!!!

: exactly, and that what we call as ”BIG Data”

• Hadoop is one of the best-known cloud platforms for big
data today
• It solves a specific class of data-crunching problems that
frequently comes up in the domain of Internet computing
and high-performance computing.
• Managing lots of information (growing by the day and
doubling by year)
• Working with many new types of data (totally unstructured)

BITS Pilani
 One of the research in the year 2012, Hadoop held the
world record for the fastest system to sort large data (500 GB
of data in59 sec and 100 terabytes of data in 68 seconds)
Designed to answer the question: “How to process big
data with reasonable cost and time?”

Super, so tell me more about Hadoop,

the data cruncher
Okay, okay…. Sit tight,

BITS Pilani
Hadoop Features

• Hadoop is optimized for batch-processing

applications, and scales to the number of CPUs
available in the cluster
• Provides Framework for Massive parallel
processing
• Programmer can focus on their program, and the
framework takes care of the details of
parallelization, fault-tolerance, locality
optimization, load balancing
• Paradigm shift: In MapReduce programming
model, computation goes to data rather than data
coming to program. Processing takes place
where data is.

BITS Pilani
Hadoop Framework Tools

BITS Pilani
Hadoop

Name of the elephant!!

BITS Pilani
Hadoop common Component

MapReduce – offline computing engine (Data

Processing Framework)
HDFS – Hadoop Distributed file system (Data Storage
Framework

Frameworks like Hbase, Pig and Hive have been built

on top of Hadoop.

• Pig is a dataflow language and execution environment

over Hadoop.
• Hbase is a distributed key-value store which supports
SQL-like queries similar to Google’s BigTable
• Hive is a distributed data warehouse to manage data
stored in the Hadoop File System.

BITS Pilani
MapReduce (Data Processing
Framework)

MapReduce
Software Processes Using large
In a reliable
Framework large clusters Nodes of
and fault-
for easily amount of having commodity
tolerant
running data in thousands hardware
manner
applications parallel of nodes

BITS Pilani
MapReduce Processing flow

Logical flow:

Two phases

Example:

BITS Pilani
Architecture Overview

blog.raremile.com
BITS Pilani
Architecture (cont.)
NameNode/ Master Node:

• Stores metadata for the files, like the directory structure of a typical FS.

• The server holding the NameNode instance is quite crucial, as there is only
one.

• Transaction log for file deletes/adds, etc.

• Handles creation of more replica blocks when necessary after a DataNode

failure
DataNode/ Slave Node:

• Stores the actual data in HDFS

• Can run on any underlying filesystem (ext3/4, NTFS, etc)

• Notifies NameNode of what blocks it has

• NameNode replicates blocks 2x in local rack, 1x elsewhere

BITS Pilani
Architecture (cont.)

The Job Tracker:

• Central authority for the complete MapReduce cluster
and responsible for scheduling and monitoring
MapReduce jobs
• Responds to client request for job submission and status
The Task Tracker:
• Workers that accepts map and reduce tasks from job
tracker, launches them and keeps track of their progress,
reports the same to job tracker.
• Keeps track of resource usage of tasks and kills the tasks
that overshoots their memory limits

BITS Pilani
Distributed Word Count

MapReduce Programming Model

BITS Pilani
Example: Word Count

def mapper(line):
foreach word in line.split():
output(word, 1)

def reducer(key, values):

output(key, sum(values))

BITS Pilani
Word Count Execution

BITS Pilani
An Optimization: The
Combiner
• Local reduce function for repeated keys produced by
same map
• Decreases amount of intermediate data

• Example: local counting for Word Count:

An Optimization: The Combiner
def combiner(key, values):
output(key, sum(values))

BITS Pilani
Word Count with Combiner

BITS Pilani
Snapshot of MarketRatings
example and Program demo

BITS Pilani
MapReduce Execution Details

Mappers preferentially scheduled on same node or same

rack as their input block
– Minimize network use to improve performance

Mappers save outputs to local disk before serving to

reducers
– Allows recovery if a reducer crashes
– Allows running more reducers than # of nodes

BITS Pilani
Fault Tolerance in MapReduce

1. If a task crashes:
– Retry on another node
• OK for a map because it had no dependencies
• OK for reduce because map outputs are on disk
– If the same task repeatedly fails, fail the job or ignore that
input block
2. If a node crashes:
– Relaunch its current tasks on other nodes
– Relaunch any maps the node previously ran
• Necessary because their output files were lost along
with the crashed node

BITS Pilani
Fault Tolerance in MapReduce

3. If a task is going slowly (straggler):

– Launch second copy of task on another node
– Take the output of whichever copy finishes first, and kill
the other one
• Critical for performance in large clusters

BITS Pilani
Challenges of Cloud
Environment
Cheap nodes fail, especially when you have many
– Mean time between failures for 1 node = 3 years
– MTBF for 1000 nodes = 1 day
– Solution: Build fault tolerance into system

Commodity network = low bandwidth

– Solution: Push computation to the data

Programming distributed systems is hard

– Solution: Restricted programming model: users write
data-parallel “map” and “reduce” functions and system
handles work distribution and failures
BITS Pilani
Hadoop

Thanks Hadoop

BITS Pilani