Fundamental Cloud Architecture

Vellore-632014, Tamil Nadu, India

 Outline

Fundamental Cloud Architectures

CSE3035 Principles of Cloud Computing; School of C.S.E, Vellore Institute of Technology (VIT)-Vellore; 12:31 PM Slide. 2
 Fundamental Cloud Architectures
❑ Fundamental Cloud
Architectures 1. Workload Distribution Architecture
2. Resource Pooling Architecture
3. Dynamic Scalability Architecture
4. Elastic Resource Capacity Architecture
5. Service Load Balancing Architecture
6. Cloud Bursting Architecture
7. Elastic Disk Provisioning Architecture
8. Redundant Storage Architecture
• These architectures represent key principles and design patterns that
are commonly used in cloud computing to build scalable, flexible,
and reliable systems
CSE3035 Principles of Cloud Computing; School of C.S.E, Vellore Institute of Technology (VIT)-Vellore; 12:31 PM Slide. 3
 Workload Distribution Architecture
❑ Workload Distribution
Architecture • Involves distributing computing tasks or workloads
across multiple servers or resources
• Balances the load to ensure optimal resource
utilization and performance
• Reduces the risk of bottlenecks and improves
overall system efficiency
– Example: Content Delivery Networks (CDN)
– Description
• CDNs distribute website content (such as images, videos, and
scripts) across multiple servers worldwide
• When a user requests content, it is served from the nearest
CDN server, reducing latency and distributing the workload
CSE3035 Principles of Cloud Computing; School of C.S.E, Vellore Institute of Technology (VIT)-Vellore; 12:31 PM Slide. 4
 Workload Distribution Architecture

load
balancer Virtual
Cloud
Server
Server A
A

Virtual
Server
B
Cloud
Server A
Cloud service
consumers

A redundant copy of Cloud Service A is implemented on Virtual Server B. The load balancer intercepts cloud service consumer
requests and directs them to both Virtual Servers A and B to ensure even workload distribution.

CSE3035 Principles of Cloud Computing; School of C.S.E, Vellore Institute of Technology (VIT)-Vellore; 12:31 PM Slide. 5
 Resource Pooling Architecture
❑ Workload Distribution
Architecture • Aggregates and virtualizes resources from multiple
❑ Resource Pooling physical hardware systems into a shared pool
Architecture
• Resources can be dynamically allocated and
reassigned based on demand
• Enables efficient utilization and optimization of
resources
– Example: Virtualized Data Centers
– Description
• In virtualized data centers, physical servers are pooled
together, and virtual machines (VMs) are created as needed
• Resources such as CPU, memory, and storage are dynamically
allocated from the resource pool to VMs based on demand
CSE3035 Principles of Cloud Computing; School of C.S.E, Vellore Institute of Technology (VIT)-Vellore; 12:31 PM Slide. 6
 Resource Pooling Architecture
Pool A

Virtual server pool CPU pool Memory pool Network pool

Pool B

Virtual server pool CPU pool Network pool

Pool C

Virtual server pool CPU pool Memory pool

Pools B and C are sibling pools that are taken from the larger Pool A, which has been allocated to a cloud consumer. This is an
alternative to taking the IT resources for Pool B and Pool C from a general reserve of IT resources that is shared throughout the
cloud

CSE3035 Principles of Cloud Computing; School of C.S.E, Vellore Institute of Technology (VIT)-Vellore; 12:31 PM Slide. 7
 Dynamic Scalability Architecture
❑ Workload Distribution
Architecture • Allows the system to adapt to varying workloads by
❑ Resource Pooling adjusting the number of resources allocated
Architecture
❑ Dynamic Scalability • Resources can be scaled up during peak demand and
Architecture scaled down during periods of low demand
• Ensures cost-effectiveness and responsiveness to
changing requirements
– Example: Auto-scaling in Cloud Platforms
– Description
• Cloud platforms like AWS and Azure offer auto-scaling features
• For example, an application hosted on Amazon EC2 instances can
automatically add or remove instances based on traffic demand,
ensuring optimal performance during peak times and cost savings
during low traffic periods
CSE3035 Principles of Cloud Computing; School of C.S.E, Vellore Institute of Technology (VIT)-Vellore; 12:31 PM Slide. 8
 Dynamic Scalability Architecture
Cloud service
instances

1 2

Virtual
Server
host
Cloud service automated scaling
consumers listener

Cloud service consumers are sending requests to a cloud service (1). The automated scaling listener monitors the cloud service to
determine if predefined capacity thresholds are being exceeded
CSE3035 Principles of Cloud Computing; School of C.S.E, Vellore Institute of Technology (VIT)-Vellore; 12:31 PM Slide. 9
 Dynamic Scalability Architecture
The number of requests coming from automated scaling Cloud service
cloud service consumers increases (3). listener instances

Cloud service
consumers

X Virtual
4 X Server
5
host

automated scaling listener notices cloud service

Instances overloaded with requests
The workload exceeds the performance thresholds. If the cloud service implementation is deemed eligible for
The automated scaling listener determines the next additional scaling, the automated scaling listener initiates
course of action based on a predefined scaling policy the scaling process (5).
(4).
CSE3035 Principles of Cloud Computing; School of C.S.E, Vellore Institute of Technology (VIT)-Vellore; 12:31 PM Slide. 10
 Dynamic Scalability Architecture
Cloud service Virtual Server
8 instances host

Cloud service automated scaling Resource

consumers listener 6 7
replication

The automated scaling listener sends a signal to the resource replication mechanism (6), which creates more instances of the
cloud service (7). Now that the increased workload has been accommodated, the automated scaling listener resumes monitoring
and detracting and adding IT resources, as required (8).

CSE3035 Principles of Cloud Computing; School of C.S.E, Vellore Institute of Technology (VIT)-Vellore; 12:31 PM Slide. 11
 Elastic Resource Capacity Architecture
❑ Workload Distribution
Architecture • Similar to dynamic scalability, it involves the ability to
❑ Resource Pooling rapidly provision and de-provision resources based on
Architecture demand
❑ Dynamic Scalability
Architecture • "Elasticity" refers to the system's ability to scale
❑ Elastic Resource resources up or down dynamically
Capacity Architecture
• Supports automatic adjustment of resources without
manual intervention
– Example: Amazon Elastic Compute Cloud (EC2)
– Description
• Amazon EC2 allows users to easily scale compute capacity up
or down based on demand
• Users can launch additional instances when traffic increases
and terminate them when demand decreases
CSE3035 Principles of Cloud Computing; School of C.S.E, Vellore Institute of Technology (VIT)-Vellore; 12:31 PM Slide. 12
 Elastic Resource Capacity Architecture
2 resource pool with memory and
1 Cloud CPU sub-pools
service

Automated Memory pool 4

Cloud service scaling
consumers listener

CPU pool
hypervisor

intelligent automation engine

Cloud service consumers are actively sending requests to a cloud service (1), which are monitored by an automated scaling listener
(2). An intelligent automation engine script is deployed with workflow logic (3) that is capable of notifying the resource pool using
allocation requests (4).

CSE3035 Principles of Cloud Computing; School of C.S.E, Vellore Institute of Technology (VIT)-Vellore; 12:31 PM Slide. 13
 Elastic Resource Capacity Architecture
Cloud service consumers
resource pool
5 with memory
Cloud
service and
CPU sub-
pools
Memory pool
Automated scaling
Cloud service consumer requests increase (5), listener
causing the automated scaling listener to
signal the intelligent automation engine to 8
execute the script (6). hypervisor CPU pool
6 7
The script runs the workflow logic that
signals the hypervisor to allocate more IT
resources from the resource pools (7).

The hypervisor allocates additional CPU and

RAM to the virtual server, enabling the
increased workload to be handled (8).
intelligent automation engine

CSE3035 Principles of Cloud Computing; School of C.S.E, Vellore Institute of Technology (VIT)-Vellore; 12:31 PM Slide. 14
 Service Load Balancing Architecture
❑ Workload Distribution
Architecture • Distributes incoming network traffic across multiple
❑ Resource Pooling servers or resources
Architecture
❑ Dynamic Scalability • Ensures even distribution of workload and prevents
Architecture
overloading of individual servers
❑ Elastic Resource
Capacity Architecture • Enhances availability and reliability by spreading
❑ Service Load Balancing
Architecture the load
– Example: Application Load Balancers (ALB)
– Description
• ALBs distribute incoming web traffic across multiple EC2
instances to ensure that no single server is overwhelmed
• It balances the load, enhances fault tolerance, and improves the
availability of the application
CSE3035 Principles of Cloud Computing; School of C.S.E, Vellore Institute of Technology (VIT)-Vellore; 12:31 PM Slide. 15
 Service Load Balancing Architecture
Cloud service consumers

The load balancer intercepts

messages sent by cloud service
consumers (1) and forwards them
to the virtual servers so that the
workload processing is horizontally 1
scaled (2). load balancer

Cloud
Cloud service
service Cloud A
A service
A 2
2 Virtual server
Virtual server 2
Virtual server C
A B
CSE3035 Principles of Cloud Computing; School of C.S.E, Vellore Institute of Technology (VIT)-Vellore; 12:31 PM Slide. 16
 Service Load Balancing Architecture
Cloud service consumers

Cloud service consumer requests

are sent to Cloud Service A on
Virtual Server A (1). The cloud
service implementation includes load balancer
built-in load balancing logic that
is capable of distributing requests
to the neighboring Cloud Service 1
A implementations on Virtual Cloud
Cloud Cloud
Servers B and C (2) service Virtual
service service
A server C
A A
2 Virtual
Virtual
server A server B
2

CSE3035 Principles of Cloud Computing; School of C.S.E, Vellore Institute of Technology (VIT)-Vellore; 12:31 PM Slide. 17
 Cloud Bursting Architecture
❑ Workload Distribution
Architecture • Involves the dynamic allocation of additional resources
❑ Resource Pooling from a public cloud to handle peak workloads
Architecture
❑ Dynamic Scalability • Organizations can use their private cloud resources for
Architecture normal operations and "burst" into the public cloud
❑ Elastic Resource when demand spikes
Capacity Architecture
❑ Service Load Balancing • Provides a cost-effective solution for handling
Architecture
occasional high demand
❑ Cloud Bursting
Architecture – Example: Netflix
– Description
• Netflix uses cloud bursting to handle peak demand. During
high traffic times, such as when a popular show is released,
Netflix can burst into public cloud resources to scale its
infrastructure and meet increased demand.
CSE3035 Principles of Cloud Computing; School of C.S.E, Vellore Institute of Technology (VIT)-Vellore; 12:31 PM Slide. 18
 Cloud Bursting Architecture
2

An automated scaling listener Service

monitors the usage of on- A
premise Service A, and Cloud
redirects Service Consumer Service
C’s request to Service A’s automated A
redundant implementation in scaling listener
the cloud (Cloud Service A) 1
once Service A’s usage
threshold has been exceeded
(1). A resource replication
system is used to keep state Service Service Service
management databases Consumer Consumer Consumer
synchronized (2). A B C

CSE3035 Principles of Cloud Computing; School of C.S.E, Vellore Institute of Technology (VIT)-Vellore; 12:31 PM Slide. 19
 Elastic Disk Provisioning Architecture
❑ Workload Distribution
Architecture • Allows for the dynamic allocation and scaling of
❑ Resource Pooling storage resources based on application requirements
Architecture
❑ Dynamic Scalability • Ensures that applications have sufficient storage
Architecture
capacity to handle data growth
❑ Elastic Resource
Capacity Architecture • Supports efficient use of storage resources without
❑ Service Load Balancing
Architecture over-provisioning
❑ Cloud Bursting – Example: Amazon Elastic Block Store (EBS)
Architecture
❑ Elastic Disk Provisioning – Description
Architecture • EBS in AWS allows users to dynamically provision and resize
storage volumes for EC2 instances. As data storage needs
change, users can adjust the size of EBS volumes without
disrupting the running instances
CSE3035 Principles of Cloud Computing; School of C.S.E, Vellore Institute of Technology (VIT)-Vellore; 12:31 PM Slide. 20
 Elastic Disk Provisioning Architecture
The cloud consumer requests a virtual
server with three hard disks, each with a
capacity of 150 GB (1) 3

The virtual server is provisioned

according to the elastic disk provisioning 150 GB
architecture, with a total of 450 GB of
disk space (2)
150 GB
1
The 450 GB is allocated to the virtual
150 GB
server by the cloud provider (3)

The cloud consumer has not installed any Provisioned: 450GB 2

software yet, meaning the actual used used: 0 GB 4
space is currently 0 GB (4)
allocated: 450GB 5
Because the 450 GB are already allocated
and reserved for the cloud consumer, it
will be charged for 450 GB of disk usage
as of the point of allocation (5)

CSE3035 Principles of Cloud Computing; School of C.S.E, Vellore Institute of Technology (VIT)-Vellore; 12:31 PM Slide. 21
 Elastic Disk Provisioning Architecture
The cloud consumer requests a virtual
server with three hard disks, each with a
capacity of 150 GB (1) 3
The virtual server is provisioned by this
architecture with a total of 450 GB of disk 150 GB
space (2)

The 450 GB are set as the maximum disk 150 GB

1
usage that is allowed for this virtual
server, although no physical disk space 150 GB
has been reserved or allocated yet (3)
Provisioned: 450GB 2
The cloud consumer has not installed any used: 0 GB 4
software, meaning the actual used space is
currently at 0 GB (4) allocated: 0 GB 5
Because the allocated disk space is equal
to the actual used space (which is
currently at zero), the cloud consumer is
not charged for any disk space usage (5)

CSE3035 Principles of Cloud Computing; School of C.S.E, Vellore Institute of Technology (VIT)-Vellore; 12:31 PM Slide. 22
 Elastic Disk Provisioning Architecture
A request is received from a cloud
consumer, and the provisioning of a
new virtual server instance begins (1) thin provisioning
program
As part of the provisioning process, the
hard disks are chosen as dynamic or
thin-provisioned disks (2) 4
The hypervisor calls a dynamic disk
allocation component to create thin
disks for the virtual server (3) 3 hypervisor 2

Virtual server disks are created via the

thin-provisioning program and saved in 1
a folder of near-zero size. The size of
this folder and its files grow as
operating applications are installed and
additional files are copied onto the 5 pay-per-use
virtual server (4) Cloud consumer
monitor
The pay-per-use monitor tracks the
actual dynamically allocated storage
for billing purposes (5)

CSE3035 Principles of Cloud Computing; School of C.S.E, Vellore Institute of Technology (VIT)-Vellore; 12:31 PM Slide. 23
 Redundant Storage Architecture
❑ Workload Distribution
Architecture • Involves replicating data across multiple storage
❑ Resource Pooling devices or locations to ensure data availability and
Architecture
reliability
❑ Dynamic Scalability
Architecture • Protects against hardware failures or data corruption
❑ Elastic Resource
Capacity Architecture • Commonly used in distributed and fault-tolerant
❑ Service Load Balancing systems
Architecture
– Example: RAID (Redundant Array of Independent Disks)
❑ Cloud Bursting
Architecture – Description
❑ Elastic Disk Provisioning • RAID is a storage technology that combines multiple disks into
Architecture
a single logical unit to provide redundancy and fault tolerance.
❑ Redundant Storage
Data is mirrored or parity information is stored across disks,
Architecture
ensuring that if one disk fails, data is still accessible from other
disks
CSE3035 Principles of Cloud Computing; School of C.S.E, Vellore Institute of Technology (VIT)-Vellore; 12:31 PM Slide. 24
 Redundant Storage Architecture
2

Storage replication

primary storage device secondary storage device

Storage replication is used to keep the redundant storage device

synchronized with the primary storage device.
CSE3035 Principles of Cloud Computing; School of C.S.E, Vellore Institute of Technology (VIT)-Vellore; 12:31 PM Slide. 25
CSE3035 Principles of Cloud Computing; School of C.S.E, Vellore Institute of Technology (VIT)-Vellore; 12:31 PM Slide. 26