Professional Documents
Culture Documents
Drmahmoud - 2195 - 3382 - 1 - Assignment MT Cloud Computing
Drmahmoud - 2195 - 3382 - 1 - Assignment MT Cloud Computing
Submission
Course Name: Introduction to Cloud Computing Deadline: 26th April 2020
Instructor’s Name: Dr. Mahmoud Aljawarneh Total Marks: 15
IMPORTANT instructions:
Read all the questions carefully and then solve in this file.
Marks of each question are mentioned at the end of each question.
Solved Assignment must be mailed to the respective faculty by the due date and in Word Format
Document. (No scanned or Picture Format File will be acceptable).
Non-submissions of the assignments will be considered as absent for the Midterm and will be
marked “0” on ZABDESK Recap Sheet.
While checking of the Assignment, SZABIST Plagiarism policy is to be followed.
Group assignments are not allowed.
_____________________________________________________________________________
Q.No.1: Define cloud computing? How many types of deployment models are used in cloud?
Shortly discuss and differentiate each of those model types? [5 Marks]
Q.No.2: Research Amazon EC2, Google Cloud Platform and write a detailed comparison
using your own wording and give references where appropriate. Focus on the following
comparative points:
1. Business model (IaaS, PaaS etc)
2. VM instance types offered (such as micro, small, medium, large etc.) along with
their number of vCPUs and memory sizes
3. Storage
4. Development environment offered
5. OS environments offered
6. Performance and scalability
7. Reliability and fault tolerance
8. Pricing model
9. Auto-Scaling/Elasticity
10. Monitoring tools/service provided
[10 Marks]
This Amazon Web Services is took as subordinate of Amazon.com, it started to provide coud
computing facilities in 2006 and launched, they provide different type of solutions platforms and
services has absolutely concreted the technique for cloud computing. But the Amazon EC2 was
launched by amazon in previous 2006 August, which was a public beta, as they’re having 50 plus
more services like few example are defined below such as:
Google cloud platform was an essential and made different kind of services available and also the
solutions were provided, which allows you to being on the same infrastructure such they merged
hardware and software infrastructure while using their Google services such as Gmail, Youtube.
They started to provide their service in 2008 and launched, they’re also having different kind of
product and services which are also 50 to 60 plus and more.
Amazon EC2 Business Model: Amazon EC2 or Elastic Computing Service is uses
Infrastructure-as-a-service business delivery model, provides the usages related to intra-cloud
network service, storage’s, processor provisioning, virtual instances and other resources. Basically
IaaS model gives users more controls and flexible functionality.
Google Cloud Platform Business Model: Whereas Google Cloud Platform provides all
types of delivery models, either its IaaS, SaaS, or PaaS. Such google google has its service Google
Compute Engine which provide Infrastructure-as-a-service facilitating consumers to host their
workloads, another one service is Google App Engine, which provide Platform-as-a-service giving
functionalities to software developers and Google cloud storage somehow provide Software-as-a-
Service which allow us to store large unset data.
Google Cloud Platform VM Instances and their size of CPUs and memories:
There are different kind of VM instances delivered by Google cloud platform. Basically those
Google Compute Machine are defined as Google Compute Instances.
Types of General purpose machines offering the great price performance for a different kind of
workloads of consumers:
Type of N1 Machine offers up-to 96 vCPUS and memory per vCPU is 6.5 MB and those
are obtainable on Haswell, Broadwell, Skylake CPUs Platform, Ivy Bridge and Intel Sandy
Bridge. Where N1 is f1-mico and G1-small pooled types of core’s machines having up-to
one vCPU made available for short interval of bursting it.
E2 machines typically are cost personalized Virtual Machines that offering up-to 16 vCPUs
and 8Gb of memory used per Central processing Unit. E2 machines having predetermined
platforms of CPU it would be either running on the 2nd generation of AMD EPYC Rome or
Intel Processor. Where E2 is having E2-micro, e2-small and e2-medium it also has pooled
of core’s machines which having several type of CPUs available for short internal of
bursting.
N2 machines providing up-to 80 CPUs and memory of 8Gb per CPU and obtainable on the
Intel cascade Lake CPU platforms. Whereas N2 could have n2-small, medium, micro.
Depend on the memory workload and pricings.
N2D machines offering up-to 224 vCPUs and memory of 8 Gb per CPU are accessible on
platform such as second generation of AMD EPYC Rome.
Machines CPUs Memory Machines CPUs Memory Machines CPUs Memory
N1- 1 3.75 N2- 2 8 N2d- 2 8
standard-1 standard-2 standard-2
N1- 2 7.50 N2- 4 16 N2d- 4 16
standard-2 standard-4 standard-4
N1- 4 15 N2- 8 32 N2d- 8 32
standard-4 standard-8 standard-8
N1- 8 30 N2- 16 64 N2d- 16 64
standard-8 standard- standard-
N1- 16 60 16 16
standard- N2- 32 128 N2d- 32 128
16 standard- standard-
N1- 32 120 32 32
standard- N2- 48 192 N2d- 48 192
32 standard- standard-
N1- 62 240 48 48
standard- N2- 64 256 N2d- 64 256
64 standard- standard-
N1- 128 360 64 64
standard- N2- 80 320 N2d- 80 320
96 standard- standard-
80 80
N1 High Memory Machines Types N2 High Memory Machine Types N2D High Memory Machine Type
Machines CPUs Memory Machines CPUs Memory Machines CPUs Memory
N1- 2 13 N2- 2 16 N2d- 2 16
highmem- highmem- highmem-
2 2 2
N1- 4 26 N2- 4 32 N2d- 4 32
highmem- highmem- highmem-
4 4 4
N1- 8 52 N2- 8 64 N2d- 8 64
highmem- highmem- highmem-
8 8 8
N1- 16 104 N2- 16 128 N2d- 16 128
highmem- highmem- highmem-
16 16 16
N1- 32 208 N2- 32 256 N2d- 32 256
highmem- highmem- highmem-
32 32 32
N1- 64 416 N2- 48 384 N2d- 48 384
highmem- highmem- highmem-
64 48 48
N1- 96 624 N2- 64 512 N2d- 64 512
highmem- highmem- highmem-
96 64 64
N2- 80 640 N2d- 80 786
highmem- highmem-
80 80
E2- highmem-4 4 32
E2- highmem-8 8 64
Amazon T3 Instance:
Instances of T3 are next level generation all-purpose instance categories which are burstable that
delivers standard level of performance of Central Processing Unit. Where this instance offers a
stability of memory, compute and resources of networks especially intended for use of apps with
moderate Central processing unit practice that practice impermanent points of procedure.
Amazon T2 Instances:
Where TS instance of amazon provide very great performance of bursting that offer a standard
performance of central processing unit with the aptitude to erupt overhead route. T2 instances
ability and performance would have unlimited which can tolerates a very great performance of
CPU for as extended as capacity of working requirements. T2 instances are good choice if you
need variety of general purposes in one specific instance to perform all workloads.
Amazon M5 Instances:
Amazon EC2 instances of M5 are bring as latest version generated, which used for general
purposes and motorized by Platinum 8175M processors of Intel Xeon. This instance also delivers
the balancing environment of memory, compute and resources of networks.
Amazon M4 instances:
EC2 M6g Amazon instances are motorized by 2nd generation of Intel Xeon E5-2676 v3 Scalable
processors. It also delivers a balancing memory, compute and resources of networks for a wide-
ranging regular set of work capacities.
Less Frequency Access Lowest Frequency Low Frequency High Frequency Access
Access Access
Near-line Archive Cold-line Standard
Their Class of Near-line Archive storage class Their Class of Cold- Mostly the general projects
is highly durable and so specifically storage line is highly durable get started with storage
fast storage for designed for cost and so fast storage for which holding standard
accessing data reduced operative and for long accessing data reduced classes which is especially
amount of than monthly term maintenance for amount of than quarter heightened for higher
which once. accessing data reduced a once. access related to frequency
amount of than yearly and performance.
which once
Amazon Elastic File System (EFS): Simplified EFS system, scalability for balancing capabilities
of workloads of Linux-based system while using Amazon web services and resources are on
premises.
Amazon Elastic Block Store (EBS): this storage is locally designed for Amazon EC2 and for
NoSQL Database, and relational database management system, Big Data Processing in which data
mining comes, enterprises/industries applications or for the setup of recovery and back up file
system.
Amazon Simple Storage Services (Amazon S3): This one is the most rleaible for accessing like
from anywhere having any internet location you can access the data file. Its counted in swerverless
computing, and for the content of which is generated by user.
AWS Backup: In this features the backups are fully managed by Amazon that makes easier for
user to access and get back up whenever the consumer wants to get even on across Amazon web
services or on premises using the gateway of AWS storage.
Amazon S3 Glacier: Extremely reasonable for extensive term classes of storage that may
interchange compliances regulatory and archives.
Amazon Storage Gateway: A cloud of fusion which is hybrid storage enhancing situation of on-
premises with the use of Amazon Cloud Services (ACS) for transferring and bursting.
Flexible Environment:
Instances of application executed within container of Dockers which utilize computer engine
virtual machine, there are some following listed languages which are supported by flexible
environment:
Python
Ruby
Java
NodeJS
Go
PHP
.NET
Those application which receives stable traffic and faces regular fluctuations of traffic and uses the
constraints of feature scaling, like sometimes it would be up or down progressively. Customs and
relies on frameworks that having code which is native one. While utilizing this flexible
environment, retrieving the services which are provided by Google cloud development that exists
in the Google compute engine.
Standard Environment:
Instances of application executed using sandbox. And in this standard environment scaling
performs its rapid actions like it quickly measure and move either Up or to Down. Some specific
versions of programming language are given below which are supported by standard environment:
Java 11 & 8
Node.js 10 & 8
PHP 7.3, 7.2 & 5.5
Ruby Beta Version 25
Python 2.7 &3.7
Go 1.11, 1.12 & 1.13 (All are beta)
Those applications envisioned to execute at actual low or free of cost. There’s environment given
by them that is only for what you need and when you needed.
Cloud9 which is provided by Amazon web services, makes flexible to execute your any kind of
environment related to development which is coped on instance of Amazon EC2. This
environment having many beneficial features and functions such as it makes you able to get hint of
code and completion of code throughout the debugging process. The service of Cloud9 gives
flexibility of experiencing shell which is based on browser that’s so reliable to work on, further
software is needed to be install, put it though git-hub or hit commands of whatever language you
want.
This development environment makes you able to execute or build your applications that based on
server-less, Those configurations are needed to make server-less application so it take all
predefined configurations with Libraries, SDKs and plugins. So it save much more time, consumer
don’t faces much difficulties while using this platform.
There are some following Operating Systems are available by maintainers to Guests.
But Google don’t do any compromise in experience of consumer and performance. While the time
of providing high extreme resources to user is when the traffic of users instance goes on peak. So
when you needed a required resources, it must provide by Google Cloud Platform.
GKE which is Google Kubernetes Engine and GCE which is Google Compute Engine, in both
these, collections are combined with auto-scalar and based on metrics this processes see the
condition like in growing and shrinking the resources to consumer.
Server-less platform of Google Cloud offered already coped storages, databases, compute system
and other kind of facilities. So the performance is take its place rapidly act to scaling from 0 to
volume requests which volume even could go high, whatever resources consumed by users,
metrics are auto-saved. Based on these metrics user pays.
Metrics crosswise over infrastructure and Applications provided by Google cloud platform, which
help to take decision when you make data determined scaling. And Cloud-spanner, Big-table such
type of these databases can increase performance and constantly provide theirs best even on huge
size of data.
Depends on user needs when the instance needs to be start or stop, just those actions performed by
users, so when you begin to use the instance, the agent makes metrics according to usage of
instance while Amazon EC2 uses billing system, this gives ability to reduce cost, just you pay for
what you’ve used.
Amazon EC2 also provides the optimization on cost reduction and computational performance.
Such we’ve fleet of Amazon EC2, using a single API you can access the fleets of EC2, and
provisioning the capacity of compute crossways to various instances of EC2. Even you can state as
per requirements which instance has to be scale up or down. And all resources are provided on-
Demand of user. On the other side, the process of hibernation is seen which is used to hibernate
your instance like when the user wants to stop instance or resume, when the instance is hibernated
you can’t even perform function on instance, it’s also gives capability to reduce cost and enhance
its performance.
Reliability and Fault Tolerance of Google Cloud platform: Google cloud platforms
services are meant to be reliable for any kind of industry organization even paid or non-paid such
as used for institutions of governments, any kind of business, schools and colleges. Google having
powerful infrastructure which is used by 1 billion of users and Google cloud platform uses and
support eight applications, variety of services provided by Google cloud platform to use which as
App Engine offering Infrastructure as a service and another one is Compute Engine which using
model of Platform-as-a-service, also provide best offers via Gmail to utilize it. Where for the fault
tolerant for GCP is defined as, the architecture of Microsoft active directory deployed on Google
cloud which has Internet information Services IIS and SQL server services.
Reliability and Fault Tolerance of Amazon EC2: The Amazon EC2 is so influential
and great instance. It’s also called a virtual machine as it as numerous advantages like offering
capacity of computing which is quiet greatly scalable and even easy to use just like start the
machine and stop. They offering several amongst operating systems to choose, the need of straight
investments’ are eradicated, fastest services provided by Amazon, you can easily make your own
application and deploy in the fastest moment and on Amazon EC2 gives elasticity of several
servers to be launch when needed like as per your requirements. As fault tolerance is managed on
Amazon EC2, when the users get allowance to access those offered resources, so the mechanism of
EC2 works as user couldn’t face any error from initializing to configuring. It also comprises
regions which took as geographical that known as for fault-tolerance is zone of availability. So
when the user accessing the instance so Amazon machine image (AMI) must be selected and user
doesn’t face any delay.
There are other Customer Friendly Pricing accessible which offered by Google cloud pricing
models, Pay as you go; in which when you’ll be needing service so you’ve to pay. 2nd one is
Termination fees; in this model you don’t have to pay in each second when turning it off. And the
last one is No up-front costs: there’s no any cost required by google, some greats services are
provided by Google.
On-Demand: The resources provided by Amazon EC2 is depending on consumer, if it gets less
subscription is the beginning, and after some time, the consumer needed more resources, so it
could be offer at On-Demand manner. There’s no any long term commitment done between
provider and consumers, as users uses application on demands and pay for them.
Spot Instances: Which allows you to get capacity of computing while requesting and discounts of
90% provided on getting on Demand offer.
Saving Plans: This one is the most flexible pricing on which you get one time subscription for
even one month constant or one year.
Reserved Instances: Discounts provided in Reserved Instance are quite good and up to 75%
given. When you get subscription of Instance it give additional features to use, so you can use
those reserved instance whenever you want.
Dedicated Hosts: Dedicated hosting only specified for use and it’s took as EC2 server physically.
And provides hourly based subscription plan on-Demand.
There are some essentials concepts and services are used to auto-scaling in Google Cloud platform
which are:
Managed Groups of Instances: There are both concepts are seen like unmanaged and managed
groups, so the auto-scaling only to be perform with managed group of instances in Google cloud
platform, whether you’re adding or deleting instances which made through template of instance.
Policy of Auto-Scaling and Utilization Targeted: In this procedure you makes some policy and
creates auto-scalar when to choose the level of target utilization, there are some following
standards defined below through which you can define policies:
- Utilization of CPU which is averaged.
- Allocation to capacity of HYPER TEXT TRANSFER PROTOCOL load assessment.
- Cloud metrics of monitoring.
There are two types of periods are observed and jumble-sale by Google Cloud Platform:
Period of Cool down Period of Stabilization
When instance to be initialize, sometimes usage On the edge of last 10 minutes of period,
statuses aren’t reflected usually, so these kind of the groups of endorsed targeted size is
statistics of usage wouldn’t be steadfast for the calculated by auto-scalar. These preceding
auto-scalar to comprehend. When you’ll be 10 minutes are stated as the period of
omitting the corresponding data, meanwhile, in stabilization. But it could make delay to
this period of time, you take an cool down process scaling down while usage of these last 10
which permits instances to complete initializing of minutes. But this feature is considered as
instance afore the gathering usage information auto-scaling
process starts by the auto-scalar and the cool-down
time amount is 60 sec.
Where the procedure of Scaling at Dynamic works as, it’s always to make changes on demand, on
the other side, Predictive scaling process works as; those demands which are predicted are made
available as EC2 instances automatically. Both of these mechanism could be merge together to
make rapid scaling process.
Especially those applications which powered by cloud and their performances, robustness and
overall healthiness shown while getting monitor services. Metadata, events and metrics are
collected From GCP, which comes in full-stack monitoring.
Those tools are listed below:
Google Cloud Storage Monitoring tool: Which is used to monitor storage’s containers in real-
time and analyze it.
Google Cloud File-store Monitoring tool: It manages the services of file-storage which requires
for application, and the file-system interface and data’s shared file-system is necessary for
applications. This platform allows you to make sure the readiness and robustness of instances,
rapidly trouble-shooting of errors.
Some of other tools like for resources, billing, storage disks, and networks, on the other side, it’s
confidentiality, integrity and availability are also monitored by Google Cloud Platform.
There are two kind of monitoring tools comes in Amazon EC2: One is Manual Monitoring Tools
and other one is Automatic Monitoring Tools.
Manual Monitoring Tools are must be enable by users to check performance for specific resource
which is being utilized by consumer.
Automatic Monitoring Tools, which provides auto-check in performances of services which are
used by consumer.
Manual Monitoring Tools: This part of manual monitoring allows users to monitor those
provided services by himself such like check of status, scripting monitors and alarming Cloud-
Watch aren’t enclosed with it. Some of the dashboards are displayed to user to monitor all of these
things according to the user’s requirements.
References
https://kinsta.com/blog/google-cloud-vs-aws/
https://cloud.google.com/compute/docs/machine-types#predefined_machine_types
https://aws.amazon.com/ec2/instance-types/
https://aws.amazon.com/products/storage/